diff --git "a/trainer_state.json" "b/trainer_state.json" new file mode 100644--- /dev/null +++ "b/trainer_state.json" @@ -0,0 +1,17376 @@ +{ + "best_global_step": 10720, + "best_metric": 0.6943244934082031, + "best_model_checkpoint": "saves_multiple/ia3/llama-3-8b-instruct/train_conala_101112_1760638010/checkpoint-10720", + "epoch": 20.0, + "eval_steps": 536, + "global_step": 10720, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.009328358208955223, + "grad_norm": 1.1089756488800049, + "learning_rate": 1.8656716417910447e-07, + "loss": 3.1108, + "num_input_tokens_seen": 1280, + "step": 5 + }, + { + "epoch": 0.018656716417910446, + "grad_norm": 1.2169495820999146, + "learning_rate": 4.197761194029851e-07, + "loss": 2.5596, + "num_input_tokens_seen": 2592, + "step": 10 + }, + { + "epoch": 0.027985074626865673, + "grad_norm": 1.3528355360031128, + "learning_rate": 6.529850746268657e-07, + "loss": 2.5889, + "num_input_tokens_seen": 4128, + "step": 15 + }, + { + "epoch": 0.03731343283582089, + "grad_norm": 1.6517956256866455, + "learning_rate": 8.861940298507463e-07, + "loss": 2.837, + "num_input_tokens_seen": 5472, + "step": 20 + }, + { + "epoch": 0.04664179104477612, + "grad_norm": 1.193107008934021, + "learning_rate": 1.119402985074627e-06, + "loss": 3.0598, + "num_input_tokens_seen": 6816, + "step": 25 + }, + { + "epoch": 0.055970149253731345, + "grad_norm": 0.9920525550842285, + "learning_rate": 1.3526119402985074e-06, + "loss": 2.745, + "num_input_tokens_seen": 8352, + "step": 30 + }, + { + "epoch": 0.06529850746268656, + "grad_norm": 0.8318398594856262, + "learning_rate": 1.585820895522388e-06, + "loss": 2.468, + "num_input_tokens_seen": 9824, + "step": 35 + }, + { + "epoch": 0.07462686567164178, + "grad_norm": 0.995405912399292, + "learning_rate": 1.8190298507462688e-06, + "loss": 3.0296, + "num_input_tokens_seen": 11200, + "step": 40 + }, + { + "epoch": 0.08395522388059702, + "grad_norm": 1.457022786140442, + "learning_rate": 2.0522388059701493e-06, + "loss": 2.6096, + "num_input_tokens_seen": 12608, + "step": 45 + }, + { + "epoch": 0.09328358208955224, + "grad_norm": 1.0550984144210815, + "learning_rate": 2.28544776119403e-06, + "loss": 2.3249, + "num_input_tokens_seen": 14272, + "step": 50 + }, + { + "epoch": 0.10261194029850747, + "grad_norm": 1.383970022201538, + "learning_rate": 2.5186567164179106e-06, + "loss": 2.8131, + "num_input_tokens_seen": 15552, + "step": 55 + }, + { + "epoch": 0.11194029850746269, + "grad_norm": 1.1210530996322632, + "learning_rate": 2.7518656716417913e-06, + "loss": 2.4388, + "num_input_tokens_seen": 17152, + "step": 60 + }, + { + "epoch": 0.12126865671641791, + "grad_norm": 1.851106882095337, + "learning_rate": 2.9850746268656716e-06, + "loss": 3.4585, + "num_input_tokens_seen": 18432, + "step": 65 + }, + { + "epoch": 0.13059701492537312, + "grad_norm": 0.9680728316307068, + "learning_rate": 3.2182835820895527e-06, + "loss": 3.1369, + "num_input_tokens_seen": 19648, + "step": 70 + }, + { + "epoch": 0.13992537313432835, + "grad_norm": 1.2384532690048218, + "learning_rate": 3.4514925373134334e-06, + "loss": 2.4447, + "num_input_tokens_seen": 21120, + "step": 75 + }, + { + "epoch": 0.14925373134328357, + "grad_norm": 1.6880192756652832, + "learning_rate": 3.684701492537314e-06, + "loss": 3.1773, + "num_input_tokens_seen": 22400, + "step": 80 + }, + { + "epoch": 0.15858208955223882, + "grad_norm": 1.1903709173202515, + "learning_rate": 3.917910447761194e-06, + "loss": 2.616, + "num_input_tokens_seen": 23776, + "step": 85 + }, + { + "epoch": 0.16791044776119404, + "grad_norm": 1.5434073209762573, + "learning_rate": 4.151119402985075e-06, + "loss": 3.0347, + "num_input_tokens_seen": 25184, + "step": 90 + }, + { + "epoch": 0.17723880597014927, + "grad_norm": 1.1992548704147339, + "learning_rate": 4.384328358208956e-06, + "loss": 2.4383, + "num_input_tokens_seen": 26752, + "step": 95 + }, + { + "epoch": 0.1865671641791045, + "grad_norm": 0.8413965702056885, + "learning_rate": 4.617537313432836e-06, + "loss": 2.6202, + "num_input_tokens_seen": 28096, + "step": 100 + }, + { + "epoch": 0.1958955223880597, + "grad_norm": 1.0322061777114868, + "learning_rate": 4.850746268656717e-06, + "loss": 2.7872, + "num_input_tokens_seen": 29472, + "step": 105 + }, + { + "epoch": 0.20522388059701493, + "grad_norm": 0.8668414950370789, + "learning_rate": 5.083955223880597e-06, + "loss": 2.5983, + "num_input_tokens_seen": 30976, + "step": 110 + }, + { + "epoch": 0.21455223880597016, + "grad_norm": 2.050095796585083, + "learning_rate": 5.3171641791044776e-06, + "loss": 2.556, + "num_input_tokens_seen": 32320, + "step": 115 + }, + { + "epoch": 0.22388059701492538, + "grad_norm": 2.0419604778289795, + "learning_rate": 5.550373134328359e-06, + "loss": 3.1618, + "num_input_tokens_seen": 33632, + "step": 120 + }, + { + "epoch": 0.2332089552238806, + "grad_norm": 1.1782195568084717, + "learning_rate": 5.783582089552239e-06, + "loss": 3.0985, + "num_input_tokens_seen": 35008, + "step": 125 + }, + { + "epoch": 0.24253731343283583, + "grad_norm": 1.0821821689605713, + "learning_rate": 6.01679104477612e-06, + "loss": 2.2475, + "num_input_tokens_seen": 36832, + "step": 130 + }, + { + "epoch": 0.251865671641791, + "grad_norm": 0.7770797610282898, + "learning_rate": 6.25e-06, + "loss": 2.8589, + "num_input_tokens_seen": 38208, + "step": 135 + }, + { + "epoch": 0.26119402985074625, + "grad_norm": 1.1001685857772827, + "learning_rate": 6.4832089552238806e-06, + "loss": 2.6812, + "num_input_tokens_seen": 39712, + "step": 140 + }, + { + "epoch": 0.27052238805970147, + "grad_norm": 1.5636498928070068, + "learning_rate": 6.716417910447762e-06, + "loss": 3.1454, + "num_input_tokens_seen": 41056, + "step": 145 + }, + { + "epoch": 0.2798507462686567, + "grad_norm": 1.4796608686447144, + "learning_rate": 6.949626865671642e-06, + "loss": 3.1741, + "num_input_tokens_seen": 42368, + "step": 150 + }, + { + "epoch": 0.2891791044776119, + "grad_norm": 1.0013152360916138, + "learning_rate": 7.182835820895523e-06, + "loss": 2.591, + "num_input_tokens_seen": 43968, + "step": 155 + }, + { + "epoch": 0.29850746268656714, + "grad_norm": 0.8377255797386169, + "learning_rate": 7.416044776119403e-06, + "loss": 2.4014, + "num_input_tokens_seen": 45664, + "step": 160 + }, + { + "epoch": 0.30783582089552236, + "grad_norm": 1.4879392385482788, + "learning_rate": 7.649253731343284e-06, + "loss": 3.1059, + "num_input_tokens_seen": 47008, + "step": 165 + }, + { + "epoch": 0.31716417910447764, + "grad_norm": 1.5627535581588745, + "learning_rate": 7.882462686567164e-06, + "loss": 2.8706, + "num_input_tokens_seen": 48448, + "step": 170 + }, + { + "epoch": 0.32649253731343286, + "grad_norm": 1.084007740020752, + "learning_rate": 8.115671641791045e-06, + "loss": 2.7406, + "num_input_tokens_seen": 49728, + "step": 175 + }, + { + "epoch": 0.3358208955223881, + "grad_norm": 0.7953682541847229, + "learning_rate": 8.348880597014926e-06, + "loss": 2.4427, + "num_input_tokens_seen": 51200, + "step": 180 + }, + { + "epoch": 0.3451492537313433, + "grad_norm": 0.9184035658836365, + "learning_rate": 8.582089552238805e-06, + "loss": 2.9168, + "num_input_tokens_seen": 52544, + "step": 185 + }, + { + "epoch": 0.35447761194029853, + "grad_norm": 1.67295503616333, + "learning_rate": 8.815298507462687e-06, + "loss": 2.8236, + "num_input_tokens_seen": 53984, + "step": 190 + }, + { + "epoch": 0.36380597014925375, + "grad_norm": 0.7877234220504761, + "learning_rate": 9.048507462686568e-06, + "loss": 2.8397, + "num_input_tokens_seen": 55424, + "step": 195 + }, + { + "epoch": 0.373134328358209, + "grad_norm": 1.2638061046600342, + "learning_rate": 9.281716417910449e-06, + "loss": 2.7188, + "num_input_tokens_seen": 56864, + "step": 200 + }, + { + "epoch": 0.3824626865671642, + "grad_norm": 1.1728886365890503, + "learning_rate": 9.514925373134328e-06, + "loss": 2.6282, + "num_input_tokens_seen": 58528, + "step": 205 + }, + { + "epoch": 0.3917910447761194, + "grad_norm": 1.6651438474655151, + "learning_rate": 9.74813432835821e-06, + "loss": 3.4402, + "num_input_tokens_seen": 59968, + "step": 210 + }, + { + "epoch": 0.40111940298507465, + "grad_norm": 1.2047502994537354, + "learning_rate": 9.98134328358209e-06, + "loss": 2.86, + "num_input_tokens_seen": 61344, + "step": 215 + }, + { + "epoch": 0.41044776119402987, + "grad_norm": 1.1095367670059204, + "learning_rate": 1.021455223880597e-05, + "loss": 2.828, + "num_input_tokens_seen": 62816, + "step": 220 + }, + { + "epoch": 0.4197761194029851, + "grad_norm": 1.2994441986083984, + "learning_rate": 1.0447761194029851e-05, + "loss": 2.7174, + "num_input_tokens_seen": 64224, + "step": 225 + }, + { + "epoch": 0.4291044776119403, + "grad_norm": 1.2255555391311646, + "learning_rate": 1.0680970149253732e-05, + "loss": 2.8342, + "num_input_tokens_seen": 65536, + "step": 230 + }, + { + "epoch": 0.43843283582089554, + "grad_norm": 1.0212032794952393, + "learning_rate": 1.0914179104477611e-05, + "loss": 2.3114, + "num_input_tokens_seen": 66944, + "step": 235 + }, + { + "epoch": 0.44776119402985076, + "grad_norm": 1.2241199016571045, + "learning_rate": 1.1147388059701493e-05, + "loss": 2.8167, + "num_input_tokens_seen": 68256, + "step": 240 + }, + { + "epoch": 0.457089552238806, + "grad_norm": 1.316506028175354, + "learning_rate": 1.1380597014925374e-05, + "loss": 2.9998, + "num_input_tokens_seen": 69536, + "step": 245 + }, + { + "epoch": 0.4664179104477612, + "grad_norm": 0.9730677604675293, + "learning_rate": 1.1613805970149253e-05, + "loss": 2.737, + "num_input_tokens_seen": 71008, + "step": 250 + }, + { + "epoch": 0.47574626865671643, + "grad_norm": 1.0576303005218506, + "learning_rate": 1.1847014925373134e-05, + "loss": 2.56, + "num_input_tokens_seen": 72352, + "step": 255 + }, + { + "epoch": 0.48507462686567165, + "grad_norm": 0.8291972279548645, + "learning_rate": 1.2080223880597015e-05, + "loss": 2.8403, + "num_input_tokens_seen": 73824, + "step": 260 + }, + { + "epoch": 0.4944029850746269, + "grad_norm": 1.1251211166381836, + "learning_rate": 1.2313432835820896e-05, + "loss": 2.7121, + "num_input_tokens_seen": 75392, + "step": 265 + }, + { + "epoch": 0.503731343283582, + "grad_norm": 0.8069990873336792, + "learning_rate": 1.2546641791044777e-05, + "loss": 2.6208, + "num_input_tokens_seen": 76864, + "step": 270 + }, + { + "epoch": 0.5130597014925373, + "grad_norm": 0.8595970273017883, + "learning_rate": 1.2779850746268657e-05, + "loss": 2.4393, + "num_input_tokens_seen": 78304, + "step": 275 + }, + { + "epoch": 0.5223880597014925, + "grad_norm": 1.1785235404968262, + "learning_rate": 1.3013059701492538e-05, + "loss": 2.8368, + "num_input_tokens_seen": 79776, + "step": 280 + }, + { + "epoch": 0.5317164179104478, + "grad_norm": 1.7950292825698853, + "learning_rate": 1.3246268656716417e-05, + "loss": 2.9442, + "num_input_tokens_seen": 81216, + "step": 285 + }, + { + "epoch": 0.5410447761194029, + "grad_norm": 2.2627832889556885, + "learning_rate": 1.34794776119403e-05, + "loss": 3.444, + "num_input_tokens_seen": 82496, + "step": 290 + }, + { + "epoch": 0.5503731343283582, + "grad_norm": 0.8742973804473877, + "learning_rate": 1.371268656716418e-05, + "loss": 2.4853, + "num_input_tokens_seen": 84128, + "step": 295 + }, + { + "epoch": 0.5597014925373134, + "grad_norm": 1.4135011434555054, + "learning_rate": 1.394589552238806e-05, + "loss": 3.0828, + "num_input_tokens_seen": 85472, + "step": 300 + }, + { + "epoch": 0.5690298507462687, + "grad_norm": 1.198190450668335, + "learning_rate": 1.417910447761194e-05, + "loss": 2.893, + "num_input_tokens_seen": 86944, + "step": 305 + }, + { + "epoch": 0.5783582089552238, + "grad_norm": 1.0254361629486084, + "learning_rate": 1.4412313432835823e-05, + "loss": 2.4969, + "num_input_tokens_seen": 88512, + "step": 310 + }, + { + "epoch": 0.5876865671641791, + "grad_norm": 1.1274100542068481, + "learning_rate": 1.46455223880597e-05, + "loss": 2.5741, + "num_input_tokens_seen": 89920, + "step": 315 + }, + { + "epoch": 0.5970149253731343, + "grad_norm": 1.3756201267242432, + "learning_rate": 1.4878731343283583e-05, + "loss": 3.1601, + "num_input_tokens_seen": 91360, + "step": 320 + }, + { + "epoch": 0.6063432835820896, + "grad_norm": 0.8133345246315002, + "learning_rate": 1.5111940298507463e-05, + "loss": 2.593, + "num_input_tokens_seen": 92800, + "step": 325 + }, + { + "epoch": 0.6156716417910447, + "grad_norm": 1.632451057434082, + "learning_rate": 1.5345149253731346e-05, + "loss": 2.8385, + "num_input_tokens_seen": 94208, + "step": 330 + }, + { + "epoch": 0.625, + "grad_norm": 1.023456335067749, + "learning_rate": 1.5578358208955223e-05, + "loss": 2.8064, + "num_input_tokens_seen": 95680, + "step": 335 + }, + { + "epoch": 0.6343283582089553, + "grad_norm": 0.9471065998077393, + "learning_rate": 1.5811567164179105e-05, + "loss": 2.5349, + "num_input_tokens_seen": 97216, + "step": 340 + }, + { + "epoch": 0.6436567164179104, + "grad_norm": 1.0985511541366577, + "learning_rate": 1.6044776119402986e-05, + "loss": 2.6984, + "num_input_tokens_seen": 98720, + "step": 345 + }, + { + "epoch": 0.6529850746268657, + "grad_norm": 1.4537957906723022, + "learning_rate": 1.6277985074626867e-05, + "loss": 2.8225, + "num_input_tokens_seen": 100032, + "step": 350 + }, + { + "epoch": 0.6623134328358209, + "grad_norm": 0.9889291524887085, + "learning_rate": 1.6511194029850744e-05, + "loss": 2.9027, + "num_input_tokens_seen": 101344, + "step": 355 + }, + { + "epoch": 0.6716417910447762, + "grad_norm": 2.062244176864624, + "learning_rate": 1.674440298507463e-05, + "loss": 3.4309, + "num_input_tokens_seen": 102528, + "step": 360 + }, + { + "epoch": 0.6809701492537313, + "grad_norm": 1.3980239629745483, + "learning_rate": 1.6977611940298507e-05, + "loss": 2.8665, + "num_input_tokens_seen": 104032, + "step": 365 + }, + { + "epoch": 0.6902985074626866, + "grad_norm": 1.3209247589111328, + "learning_rate": 1.7210820895522388e-05, + "loss": 2.519, + "num_input_tokens_seen": 105504, + "step": 370 + }, + { + "epoch": 0.6996268656716418, + "grad_norm": 1.0633630752563477, + "learning_rate": 1.744402985074627e-05, + "loss": 2.6755, + "num_input_tokens_seen": 106976, + "step": 375 + }, + { + "epoch": 0.7089552238805971, + "grad_norm": 1.101513147354126, + "learning_rate": 1.767723880597015e-05, + "loss": 3.0792, + "num_input_tokens_seen": 108352, + "step": 380 + }, + { + "epoch": 0.7182835820895522, + "grad_norm": 1.3991308212280273, + "learning_rate": 1.791044776119403e-05, + "loss": 2.7999, + "num_input_tokens_seen": 109824, + "step": 385 + }, + { + "epoch": 0.7276119402985075, + "grad_norm": 2.640174150466919, + "learning_rate": 1.8143656716417912e-05, + "loss": 3.5447, + "num_input_tokens_seen": 111040, + "step": 390 + }, + { + "epoch": 0.7369402985074627, + "grad_norm": 1.7576851844787598, + "learning_rate": 1.837686567164179e-05, + "loss": 2.5767, + "num_input_tokens_seen": 112512, + "step": 395 + }, + { + "epoch": 0.746268656716418, + "grad_norm": 1.8274682760238647, + "learning_rate": 1.8610074626865674e-05, + "loss": 2.4848, + "num_input_tokens_seen": 114080, + "step": 400 + }, + { + "epoch": 0.7555970149253731, + "grad_norm": 1.6939542293548584, + "learning_rate": 1.8843283582089552e-05, + "loss": 2.3819, + "num_input_tokens_seen": 115552, + "step": 405 + }, + { + "epoch": 0.7649253731343284, + "grad_norm": 1.0596818923950195, + "learning_rate": 1.9076492537313433e-05, + "loss": 2.3608, + "num_input_tokens_seen": 117024, + "step": 410 + }, + { + "epoch": 0.7742537313432836, + "grad_norm": 0.9721699357032776, + "learning_rate": 1.9309701492537314e-05, + "loss": 2.3925, + "num_input_tokens_seen": 118496, + "step": 415 + }, + { + "epoch": 0.7835820895522388, + "grad_norm": 1.1635557413101196, + "learning_rate": 1.9542910447761195e-05, + "loss": 2.5635, + "num_input_tokens_seen": 119872, + "step": 420 + }, + { + "epoch": 0.792910447761194, + "grad_norm": 1.1643205881118774, + "learning_rate": 1.9776119402985073e-05, + "loss": 2.3643, + "num_input_tokens_seen": 121376, + "step": 425 + }, + { + "epoch": 0.8022388059701493, + "grad_norm": 1.6637731790542603, + "learning_rate": 2.0009328358208958e-05, + "loss": 2.6457, + "num_input_tokens_seen": 122784, + "step": 430 + }, + { + "epoch": 0.8115671641791045, + "grad_norm": 1.2357561588287354, + "learning_rate": 2.0242537313432835e-05, + "loss": 2.3509, + "num_input_tokens_seen": 124352, + "step": 435 + }, + { + "epoch": 0.8208955223880597, + "grad_norm": 0.9775716662406921, + "learning_rate": 2.0475746268656717e-05, + "loss": 1.9666, + "num_input_tokens_seen": 126016, + "step": 440 + }, + { + "epoch": 0.8302238805970149, + "grad_norm": 0.7204633951187134, + "learning_rate": 2.0708955223880598e-05, + "loss": 2.0463, + "num_input_tokens_seen": 127712, + "step": 445 + }, + { + "epoch": 0.8395522388059702, + "grad_norm": 1.7701642513275146, + "learning_rate": 2.094216417910448e-05, + "loss": 2.6431, + "num_input_tokens_seen": 129152, + "step": 450 + }, + { + "epoch": 0.8488805970149254, + "grad_norm": 1.1431199312210083, + "learning_rate": 2.1175373134328356e-05, + "loss": 2.573, + "num_input_tokens_seen": 130624, + "step": 455 + }, + { + "epoch": 0.8582089552238806, + "grad_norm": 1.1034901142120361, + "learning_rate": 2.140858208955224e-05, + "loss": 2.3312, + "num_input_tokens_seen": 131872, + "step": 460 + }, + { + "epoch": 0.8675373134328358, + "grad_norm": 0.8847706913948059, + "learning_rate": 2.164179104477612e-05, + "loss": 2.5823, + "num_input_tokens_seen": 133344, + "step": 465 + }, + { + "epoch": 0.8768656716417911, + "grad_norm": 1.3927074670791626, + "learning_rate": 2.1875e-05, + "loss": 2.5795, + "num_input_tokens_seen": 134656, + "step": 470 + }, + { + "epoch": 0.8861940298507462, + "grad_norm": 1.2494655847549438, + "learning_rate": 2.2108208955223884e-05, + "loss": 2.4169, + "num_input_tokens_seen": 136032, + "step": 475 + }, + { + "epoch": 0.8955223880597015, + "grad_norm": 1.0177122354507446, + "learning_rate": 2.2341417910447762e-05, + "loss": 2.1405, + "num_input_tokens_seen": 137568, + "step": 480 + }, + { + "epoch": 0.9048507462686567, + "grad_norm": 1.3683202266693115, + "learning_rate": 2.2574626865671643e-05, + "loss": 2.5997, + "num_input_tokens_seen": 139072, + "step": 485 + }, + { + "epoch": 0.914179104477612, + "grad_norm": 0.9738945364952087, + "learning_rate": 2.2807835820895524e-05, + "loss": 2.5676, + "num_input_tokens_seen": 140480, + "step": 490 + }, + { + "epoch": 0.9235074626865671, + "grad_norm": 1.3408888578414917, + "learning_rate": 2.3041044776119405e-05, + "loss": 2.3508, + "num_input_tokens_seen": 141824, + "step": 495 + }, + { + "epoch": 0.9328358208955224, + "grad_norm": 1.2888084650039673, + "learning_rate": 2.3274253731343283e-05, + "loss": 1.9532, + "num_input_tokens_seen": 143520, + "step": 500 + }, + { + "epoch": 0.9421641791044776, + "grad_norm": 1.1598825454711914, + "learning_rate": 2.3507462686567168e-05, + "loss": 2.4983, + "num_input_tokens_seen": 144896, + "step": 505 + }, + { + "epoch": 0.9514925373134329, + "grad_norm": 1.7070703506469727, + "learning_rate": 2.3740671641791045e-05, + "loss": 2.4613, + "num_input_tokens_seen": 146240, + "step": 510 + }, + { + "epoch": 0.960820895522388, + "grad_norm": 1.037986159324646, + "learning_rate": 2.3973880597014926e-05, + "loss": 1.8677, + "num_input_tokens_seen": 147744, + "step": 515 + }, + { + "epoch": 0.9701492537313433, + "grad_norm": 1.7092000246047974, + "learning_rate": 2.4207089552238807e-05, + "loss": 2.5524, + "num_input_tokens_seen": 149024, + "step": 520 + }, + { + "epoch": 0.9794776119402985, + "grad_norm": 0.9678240418434143, + "learning_rate": 2.444029850746269e-05, + "loss": 2.0176, + "num_input_tokens_seen": 150528, + "step": 525 + }, + { + "epoch": 0.9888059701492538, + "grad_norm": 1.1144541501998901, + "learning_rate": 2.467350746268657e-05, + "loss": 2.9873, + "num_input_tokens_seen": 151840, + "step": 530 + }, + { + "epoch": 0.9981343283582089, + "grad_norm": 1.0049551725387573, + "learning_rate": 2.490671641791045e-05, + "loss": 2.3421, + "num_input_tokens_seen": 153280, + "step": 535 + }, + { + "epoch": 1.0, + "eval_loss": 2.2859151363372803, + "eval_runtime": 2.4567, + "eval_samples_per_second": 96.876, + "eval_steps_per_second": 24.423, + "num_input_tokens_seen": 153344, + "step": 536 + }, + { + "epoch": 1.007462686567164, + "grad_norm": 1.4303237199783325, + "learning_rate": 2.5139925373134332e-05, + "loss": 2.0261, + "num_input_tokens_seen": 154432, + "step": 540 + }, + { + "epoch": 1.0167910447761195, + "grad_norm": 1.2744429111480713, + "learning_rate": 2.537313432835821e-05, + "loss": 1.8061, + "num_input_tokens_seen": 156032, + "step": 545 + }, + { + "epoch": 1.0261194029850746, + "grad_norm": 1.466960072517395, + "learning_rate": 2.560634328358209e-05, + "loss": 2.3013, + "num_input_tokens_seen": 157312, + "step": 550 + }, + { + "epoch": 1.0354477611940298, + "grad_norm": 1.6853851079940796, + "learning_rate": 2.583955223880597e-05, + "loss": 2.1532, + "num_input_tokens_seen": 158624, + "step": 555 + }, + { + "epoch": 1.044776119402985, + "grad_norm": 1.0950875282287598, + "learning_rate": 2.6072761194029853e-05, + "loss": 2.1603, + "num_input_tokens_seen": 160128, + "step": 560 + }, + { + "epoch": 1.0541044776119404, + "grad_norm": 1.1786143779754639, + "learning_rate": 2.6305970149253734e-05, + "loss": 2.4548, + "num_input_tokens_seen": 161344, + "step": 565 + }, + { + "epoch": 1.0634328358208955, + "grad_norm": 1.3120322227478027, + "learning_rate": 2.6539179104477612e-05, + "loss": 2.1779, + "num_input_tokens_seen": 162656, + "step": 570 + }, + { + "epoch": 1.0727611940298507, + "grad_norm": 1.3216437101364136, + "learning_rate": 2.6772388059701493e-05, + "loss": 2.1769, + "num_input_tokens_seen": 164128, + "step": 575 + }, + { + "epoch": 1.0820895522388059, + "grad_norm": 1.2234454154968262, + "learning_rate": 2.7005597014925377e-05, + "loss": 2.2049, + "num_input_tokens_seen": 165504, + "step": 580 + }, + { + "epoch": 1.0914179104477613, + "grad_norm": 2.3886806964874268, + "learning_rate": 2.7238805970149255e-05, + "loss": 2.0877, + "num_input_tokens_seen": 166880, + "step": 585 + }, + { + "epoch": 1.1007462686567164, + "grad_norm": 1.1248254776000977, + "learning_rate": 2.7472014925373136e-05, + "loss": 1.9818, + "num_input_tokens_seen": 168352, + "step": 590 + }, + { + "epoch": 1.1100746268656716, + "grad_norm": 0.5814436078071594, + "learning_rate": 2.7705223880597014e-05, + "loss": 2.1113, + "num_input_tokens_seen": 169888, + "step": 595 + }, + { + "epoch": 1.1194029850746268, + "grad_norm": 1.8702235221862793, + "learning_rate": 2.79384328358209e-05, + "loss": 1.8746, + "num_input_tokens_seen": 171296, + "step": 600 + }, + { + "epoch": 1.1287313432835822, + "grad_norm": 1.3895623683929443, + "learning_rate": 2.817164179104478e-05, + "loss": 2.0862, + "num_input_tokens_seen": 172736, + "step": 605 + }, + { + "epoch": 1.1380597014925373, + "grad_norm": 1.4130311012268066, + "learning_rate": 2.8404850746268657e-05, + "loss": 2.4122, + "num_input_tokens_seen": 174144, + "step": 610 + }, + { + "epoch": 1.1473880597014925, + "grad_norm": 1.3028775453567505, + "learning_rate": 2.863805970149254e-05, + "loss": 1.8381, + "num_input_tokens_seen": 175680, + "step": 615 + }, + { + "epoch": 1.1567164179104479, + "grad_norm": 1.7709122896194458, + "learning_rate": 2.8871268656716423e-05, + "loss": 2.2243, + "num_input_tokens_seen": 177184, + "step": 620 + }, + { + "epoch": 1.166044776119403, + "grad_norm": 1.5977100133895874, + "learning_rate": 2.91044776119403e-05, + "loss": 2.5059, + "num_input_tokens_seen": 178496, + "step": 625 + }, + { + "epoch": 1.1753731343283582, + "grad_norm": 1.028340220451355, + "learning_rate": 2.9337686567164178e-05, + "loss": 1.9563, + "num_input_tokens_seen": 179936, + "step": 630 + }, + { + "epoch": 1.1847014925373134, + "grad_norm": 0.9285154938697815, + "learning_rate": 2.957089552238806e-05, + "loss": 2.0002, + "num_input_tokens_seen": 181408, + "step": 635 + }, + { + "epoch": 1.1940298507462686, + "grad_norm": 1.252426028251648, + "learning_rate": 2.9804104477611944e-05, + "loss": 1.5745, + "num_input_tokens_seen": 183104, + "step": 640 + }, + { + "epoch": 1.203358208955224, + "grad_norm": 1.2990620136260986, + "learning_rate": 3.003731343283582e-05, + "loss": 1.7921, + "num_input_tokens_seen": 184960, + "step": 645 + }, + { + "epoch": 1.212686567164179, + "grad_norm": 1.5973103046417236, + "learning_rate": 3.0270522388059703e-05, + "loss": 1.7672, + "num_input_tokens_seen": 186400, + "step": 650 + }, + { + "epoch": 1.2220149253731343, + "grad_norm": 1.4033995866775513, + "learning_rate": 3.050373134328358e-05, + "loss": 2.1697, + "num_input_tokens_seen": 187872, + "step": 655 + }, + { + "epoch": 1.2313432835820897, + "grad_norm": 1.6351929903030396, + "learning_rate": 3.0736940298507465e-05, + "loss": 1.755, + "num_input_tokens_seen": 189216, + "step": 660 + }, + { + "epoch": 1.2406716417910448, + "grad_norm": 1.09102463722229, + "learning_rate": 3.0970149253731346e-05, + "loss": 1.9752, + "num_input_tokens_seen": 190752, + "step": 665 + }, + { + "epoch": 1.25, + "grad_norm": 1.0789620876312256, + "learning_rate": 3.120335820895523e-05, + "loss": 1.7131, + "num_input_tokens_seen": 192256, + "step": 670 + }, + { + "epoch": 1.2593283582089552, + "grad_norm": 1.1025460958480835, + "learning_rate": 3.14365671641791e-05, + "loss": 1.7189, + "num_input_tokens_seen": 193760, + "step": 675 + }, + { + "epoch": 1.2686567164179103, + "grad_norm": 1.4812594652175903, + "learning_rate": 3.166977611940299e-05, + "loss": 2.1145, + "num_input_tokens_seen": 195008, + "step": 680 + }, + { + "epoch": 1.2779850746268657, + "grad_norm": 1.1191529035568237, + "learning_rate": 3.190298507462687e-05, + "loss": 1.4613, + "num_input_tokens_seen": 196512, + "step": 685 + }, + { + "epoch": 1.287313432835821, + "grad_norm": 1.1792001724243164, + "learning_rate": 3.2136194029850745e-05, + "loss": 1.8445, + "num_input_tokens_seen": 197984, + "step": 690 + }, + { + "epoch": 1.296641791044776, + "grad_norm": 1.0382626056671143, + "learning_rate": 3.2369402985074626e-05, + "loss": 1.6369, + "num_input_tokens_seen": 199392, + "step": 695 + }, + { + "epoch": 1.3059701492537314, + "grad_norm": 1.0897375345230103, + "learning_rate": 3.2602611940298514e-05, + "loss": 2.1958, + "num_input_tokens_seen": 200640, + "step": 700 + }, + { + "epoch": 1.3152985074626866, + "grad_norm": 1.2580362558364868, + "learning_rate": 3.283582089552239e-05, + "loss": 1.6014, + "num_input_tokens_seen": 202144, + "step": 705 + }, + { + "epoch": 1.3246268656716418, + "grad_norm": 1.1206879615783691, + "learning_rate": 3.306902985074627e-05, + "loss": 1.6375, + "num_input_tokens_seen": 203552, + "step": 710 + }, + { + "epoch": 1.333955223880597, + "grad_norm": 1.322446346282959, + "learning_rate": 3.330223880597015e-05, + "loss": 2.0027, + "num_input_tokens_seen": 204864, + "step": 715 + }, + { + "epoch": 1.3432835820895521, + "grad_norm": 1.0179561376571655, + "learning_rate": 3.353544776119403e-05, + "loss": 2.2298, + "num_input_tokens_seen": 206176, + "step": 720 + }, + { + "epoch": 1.3526119402985075, + "grad_norm": 1.2582423686981201, + "learning_rate": 3.376865671641791e-05, + "loss": 1.6419, + "num_input_tokens_seen": 207584, + "step": 725 + }, + { + "epoch": 1.3619402985074627, + "grad_norm": 1.3614541292190552, + "learning_rate": 3.4001865671641794e-05, + "loss": 1.7515, + "num_input_tokens_seen": 208928, + "step": 730 + }, + { + "epoch": 1.3712686567164178, + "grad_norm": 0.9724900126457214, + "learning_rate": 3.423507462686567e-05, + "loss": 2.0114, + "num_input_tokens_seen": 210368, + "step": 735 + }, + { + "epoch": 1.3805970149253732, + "grad_norm": 1.055847406387329, + "learning_rate": 3.4468283582089556e-05, + "loss": 1.9961, + "num_input_tokens_seen": 211680, + "step": 740 + }, + { + "epoch": 1.3899253731343284, + "grad_norm": 1.2472032308578491, + "learning_rate": 3.470149253731344e-05, + "loss": 1.5429, + "num_input_tokens_seen": 213120, + "step": 745 + }, + { + "epoch": 1.3992537313432836, + "grad_norm": 1.0937882661819458, + "learning_rate": 3.493470149253731e-05, + "loss": 1.913, + "num_input_tokens_seen": 214592, + "step": 750 + }, + { + "epoch": 1.4085820895522387, + "grad_norm": 0.8679872155189514, + "learning_rate": 3.516791044776119e-05, + "loss": 1.5539, + "num_input_tokens_seen": 215936, + "step": 755 + }, + { + "epoch": 1.417910447761194, + "grad_norm": 0.9714317321777344, + "learning_rate": 3.540111940298508e-05, + "loss": 1.802, + "num_input_tokens_seen": 217312, + "step": 760 + }, + { + "epoch": 1.4272388059701493, + "grad_norm": 0.9607062935829163, + "learning_rate": 3.5634328358208955e-05, + "loss": 1.0814, + "num_input_tokens_seen": 218912, + "step": 765 + }, + { + "epoch": 1.4365671641791045, + "grad_norm": 1.6796518564224243, + "learning_rate": 3.5867537313432836e-05, + "loss": 1.699, + "num_input_tokens_seen": 220352, + "step": 770 + }, + { + "epoch": 1.4458955223880596, + "grad_norm": 1.723616600036621, + "learning_rate": 3.610074626865672e-05, + "loss": 1.7402, + "num_input_tokens_seen": 221696, + "step": 775 + }, + { + "epoch": 1.455223880597015, + "grad_norm": 0.8586432337760925, + "learning_rate": 3.63339552238806e-05, + "loss": 1.8986, + "num_input_tokens_seen": 223072, + "step": 780 + }, + { + "epoch": 1.4645522388059702, + "grad_norm": 0.9879969358444214, + "learning_rate": 3.656716417910448e-05, + "loss": 1.6987, + "num_input_tokens_seen": 224384, + "step": 785 + }, + { + "epoch": 1.4738805970149254, + "grad_norm": 1.419543981552124, + "learning_rate": 3.680037313432836e-05, + "loss": 1.6977, + "num_input_tokens_seen": 225824, + "step": 790 + }, + { + "epoch": 1.4832089552238805, + "grad_norm": 1.1860175132751465, + "learning_rate": 3.7033582089552234e-05, + "loss": 1.6987, + "num_input_tokens_seen": 227328, + "step": 795 + }, + { + "epoch": 1.4925373134328357, + "grad_norm": 1.3971459865570068, + "learning_rate": 3.726679104477612e-05, + "loss": 1.593, + "num_input_tokens_seen": 228704, + "step": 800 + }, + { + "epoch": 1.501865671641791, + "grad_norm": 1.1254042387008667, + "learning_rate": 3.7500000000000003e-05, + "loss": 1.4536, + "num_input_tokens_seen": 230304, + "step": 805 + }, + { + "epoch": 1.5111940298507462, + "grad_norm": 0.8769680261611938, + "learning_rate": 3.773320895522388e-05, + "loss": 1.3063, + "num_input_tokens_seen": 232000, + "step": 810 + }, + { + "epoch": 1.5205223880597014, + "grad_norm": 0.8746810555458069, + "learning_rate": 3.7966417910447766e-05, + "loss": 1.4181, + "num_input_tokens_seen": 233440, + "step": 815 + }, + { + "epoch": 1.5298507462686568, + "grad_norm": 0.9232519865036011, + "learning_rate": 3.819962686567165e-05, + "loss": 1.5245, + "num_input_tokens_seen": 234720, + "step": 820 + }, + { + "epoch": 1.539179104477612, + "grad_norm": 1.5356154441833496, + "learning_rate": 3.843283582089552e-05, + "loss": 1.2293, + "num_input_tokens_seen": 235968, + "step": 825 + }, + { + "epoch": 1.5485074626865671, + "grad_norm": 1.111283302307129, + "learning_rate": 3.86660447761194e-05, + "loss": 1.3806, + "num_input_tokens_seen": 237344, + "step": 830 + }, + { + "epoch": 1.5578358208955225, + "grad_norm": 0.9892191290855408, + "learning_rate": 3.889925373134329e-05, + "loss": 1.1069, + "num_input_tokens_seen": 238752, + "step": 835 + }, + { + "epoch": 1.5671641791044775, + "grad_norm": 1.3349765539169312, + "learning_rate": 3.9132462686567164e-05, + "loss": 1.6438, + "num_input_tokens_seen": 240160, + "step": 840 + }, + { + "epoch": 1.5764925373134329, + "grad_norm": 0.7101888656616211, + "learning_rate": 3.9365671641791046e-05, + "loss": 1.6477, + "num_input_tokens_seen": 241504, + "step": 845 + }, + { + "epoch": 1.585820895522388, + "grad_norm": 1.1528596878051758, + "learning_rate": 3.959888059701493e-05, + "loss": 1.4325, + "num_input_tokens_seen": 242816, + "step": 850 + }, + { + "epoch": 1.5951492537313432, + "grad_norm": 0.9072254300117493, + "learning_rate": 3.983208955223881e-05, + "loss": 1.1335, + "num_input_tokens_seen": 244416, + "step": 855 + }, + { + "epoch": 1.6044776119402986, + "grad_norm": 1.0796308517456055, + "learning_rate": 4.006529850746269e-05, + "loss": 1.8614, + "num_input_tokens_seen": 245856, + "step": 860 + }, + { + "epoch": 1.6138059701492538, + "grad_norm": 0.7929595708847046, + "learning_rate": 4.029850746268657e-05, + "loss": 1.435, + "num_input_tokens_seen": 247296, + "step": 865 + }, + { + "epoch": 1.623134328358209, + "grad_norm": 1.1046881675720215, + "learning_rate": 4.053171641791045e-05, + "loss": 1.4087, + "num_input_tokens_seen": 248704, + "step": 870 + }, + { + "epoch": 1.6324626865671643, + "grad_norm": 0.5989983081817627, + "learning_rate": 4.076492537313433e-05, + "loss": 1.2666, + "num_input_tokens_seen": 250176, + "step": 875 + }, + { + "epoch": 1.6417910447761193, + "grad_norm": 1.1555819511413574, + "learning_rate": 4.099813432835821e-05, + "loss": 1.0849, + "num_input_tokens_seen": 251808, + "step": 880 + }, + { + "epoch": 1.6511194029850746, + "grad_norm": 0.9130282402038574, + "learning_rate": 4.1231343283582094e-05, + "loss": 1.3016, + "num_input_tokens_seen": 253248, + "step": 885 + }, + { + "epoch": 1.6604477611940298, + "grad_norm": 0.8903413414955139, + "learning_rate": 4.146455223880597e-05, + "loss": 1.5082, + "num_input_tokens_seen": 254880, + "step": 890 + }, + { + "epoch": 1.669776119402985, + "grad_norm": 1.1448264122009277, + "learning_rate": 4.169776119402986e-05, + "loss": 1.5326, + "num_input_tokens_seen": 256160, + "step": 895 + }, + { + "epoch": 1.6791044776119404, + "grad_norm": 0.8502844572067261, + "learning_rate": 4.193097014925373e-05, + "loss": 1.267, + "num_input_tokens_seen": 257632, + "step": 900 + }, + { + "epoch": 1.6884328358208955, + "grad_norm": 1.16700279712677, + "learning_rate": 4.216417910447761e-05, + "loss": 1.1433, + "num_input_tokens_seen": 259072, + "step": 905 + }, + { + "epoch": 1.6977611940298507, + "grad_norm": 1.2369780540466309, + "learning_rate": 4.239738805970149e-05, + "loss": 1.3541, + "num_input_tokens_seen": 260384, + "step": 910 + }, + { + "epoch": 1.707089552238806, + "grad_norm": 0.8746596574783325, + "learning_rate": 4.2630597014925374e-05, + "loss": 1.1894, + "num_input_tokens_seen": 261952, + "step": 915 + }, + { + "epoch": 1.716417910447761, + "grad_norm": 0.6375977993011475, + "learning_rate": 4.2863805970149255e-05, + "loss": 1.1554, + "num_input_tokens_seen": 263456, + "step": 920 + }, + { + "epoch": 1.7257462686567164, + "grad_norm": 0.9331329464912415, + "learning_rate": 4.3097014925373137e-05, + "loss": 1.2327, + "num_input_tokens_seen": 265056, + "step": 925 + }, + { + "epoch": 1.7350746268656716, + "grad_norm": 0.8953114748001099, + "learning_rate": 4.333022388059702e-05, + "loss": 1.3655, + "num_input_tokens_seen": 266624, + "step": 930 + }, + { + "epoch": 1.7444029850746268, + "grad_norm": 0.7951314449310303, + "learning_rate": 4.35634328358209e-05, + "loss": 1.3618, + "num_input_tokens_seen": 267936, + "step": 935 + }, + { + "epoch": 1.7537313432835822, + "grad_norm": 0.8884233236312866, + "learning_rate": 4.379664179104478e-05, + "loss": 1.1496, + "num_input_tokens_seen": 269184, + "step": 940 + }, + { + "epoch": 1.7630597014925373, + "grad_norm": 0.7594141364097595, + "learning_rate": 4.402985074626866e-05, + "loss": 1.3473, + "num_input_tokens_seen": 270496, + "step": 945 + }, + { + "epoch": 1.7723880597014925, + "grad_norm": 0.598574161529541, + "learning_rate": 4.4263059701492535e-05, + "loss": 0.6974, + "num_input_tokens_seen": 271936, + "step": 950 + }, + { + "epoch": 1.7817164179104479, + "grad_norm": 1.2228045463562012, + "learning_rate": 4.449626865671642e-05, + "loss": 1.195, + "num_input_tokens_seen": 273216, + "step": 955 + }, + { + "epoch": 1.7910447761194028, + "grad_norm": 0.7876660823822021, + "learning_rate": 4.4729477611940304e-05, + "loss": 1.3448, + "num_input_tokens_seen": 274624, + "step": 960 + }, + { + "epoch": 1.8003731343283582, + "grad_norm": 1.1846826076507568, + "learning_rate": 4.496268656716418e-05, + "loss": 1.3283, + "num_input_tokens_seen": 276032, + "step": 965 + }, + { + "epoch": 1.8097014925373134, + "grad_norm": 1.0630214214324951, + "learning_rate": 4.519589552238806e-05, + "loss": 1.4041, + "num_input_tokens_seen": 277280, + "step": 970 + }, + { + "epoch": 1.8190298507462686, + "grad_norm": 1.0395996570587158, + "learning_rate": 4.542910447761195e-05, + "loss": 1.1267, + "num_input_tokens_seen": 278784, + "step": 975 + }, + { + "epoch": 1.828358208955224, + "grad_norm": 1.4839059114456177, + "learning_rate": 4.566231343283582e-05, + "loss": 1.3096, + "num_input_tokens_seen": 280256, + "step": 980 + }, + { + "epoch": 1.837686567164179, + "grad_norm": 1.142371654510498, + "learning_rate": 4.58955223880597e-05, + "loss": 0.9236, + "num_input_tokens_seen": 281696, + "step": 985 + }, + { + "epoch": 1.8470149253731343, + "grad_norm": 0.9519606828689575, + "learning_rate": 4.6128731343283584e-05, + "loss": 1.5925, + "num_input_tokens_seen": 283072, + "step": 990 + }, + { + "epoch": 1.8563432835820897, + "grad_norm": 1.0071145296096802, + "learning_rate": 4.6361940298507465e-05, + "loss": 0.9443, + "num_input_tokens_seen": 284640, + "step": 995 + }, + { + "epoch": 1.8656716417910446, + "grad_norm": 1.6559454202651978, + "learning_rate": 4.6595149253731346e-05, + "loss": 1.396, + "num_input_tokens_seen": 286048, + "step": 1000 + }, + { + "epoch": 1.875, + "grad_norm": 1.0475866794586182, + "learning_rate": 4.682835820895523e-05, + "loss": 1.2397, + "num_input_tokens_seen": 287360, + "step": 1005 + }, + { + "epoch": 1.8843283582089554, + "grad_norm": 0.8573509454727173, + "learning_rate": 4.70615671641791e-05, + "loss": 1.162, + "num_input_tokens_seen": 288864, + "step": 1010 + }, + { + "epoch": 1.8936567164179103, + "grad_norm": 0.9946274757385254, + "learning_rate": 4.729477611940299e-05, + "loss": 1.3201, + "num_input_tokens_seen": 290240, + "step": 1015 + }, + { + "epoch": 1.9029850746268657, + "grad_norm": 0.6469167470932007, + "learning_rate": 4.752798507462687e-05, + "loss": 1.4253, + "num_input_tokens_seen": 291616, + "step": 1020 + }, + { + "epoch": 1.912313432835821, + "grad_norm": 0.9352241158485413, + "learning_rate": 4.7761194029850745e-05, + "loss": 0.8883, + "num_input_tokens_seen": 293152, + "step": 1025 + }, + { + "epoch": 1.921641791044776, + "grad_norm": 0.9758053421974182, + "learning_rate": 4.7994402985074626e-05, + "loss": 1.2574, + "num_input_tokens_seen": 294560, + "step": 1030 + }, + { + "epoch": 1.9309701492537314, + "grad_norm": 1.0760853290557861, + "learning_rate": 4.8227611940298514e-05, + "loss": 1.2659, + "num_input_tokens_seen": 295776, + "step": 1035 + }, + { + "epoch": 1.9402985074626866, + "grad_norm": 0.7091391086578369, + "learning_rate": 4.846082089552239e-05, + "loss": 1.2977, + "num_input_tokens_seen": 297280, + "step": 1040 + }, + { + "epoch": 1.9496268656716418, + "grad_norm": 0.7541832327842712, + "learning_rate": 4.869402985074627e-05, + "loss": 0.8454, + "num_input_tokens_seen": 298912, + "step": 1045 + }, + { + "epoch": 1.9589552238805972, + "grad_norm": 0.7487569451332092, + "learning_rate": 4.892723880597015e-05, + "loss": 1.173, + "num_input_tokens_seen": 300512, + "step": 1050 + }, + { + "epoch": 1.9682835820895521, + "grad_norm": 0.960292637348175, + "learning_rate": 4.916044776119403e-05, + "loss": 1.6719, + "num_input_tokens_seen": 301920, + "step": 1055 + }, + { + "epoch": 1.9776119402985075, + "grad_norm": 0.588718831539154, + "learning_rate": 4.939365671641791e-05, + "loss": 0.9126, + "num_input_tokens_seen": 303424, + "step": 1060 + }, + { + "epoch": 1.9869402985074627, + "grad_norm": 0.9057972431182861, + "learning_rate": 4.9626865671641794e-05, + "loss": 1.1876, + "num_input_tokens_seen": 305056, + "step": 1065 + }, + { + "epoch": 1.9962686567164178, + "grad_norm": 1.2616747617721558, + "learning_rate": 4.986007462686567e-05, + "loss": 1.758, + "num_input_tokens_seen": 306336, + "step": 1070 + }, + { + "epoch": 2.0, + "eval_loss": 1.1337827444076538, + "eval_runtime": 2.461, + "eval_samples_per_second": 96.707, + "eval_steps_per_second": 24.38, + "num_input_tokens_seen": 306640, + "step": 1072 + }, + { + "epoch": 2.0055970149253732, + "grad_norm": 0.9143099784851074, + "learning_rate": 4.999999469854421e-05, + "loss": 1.3074, + "num_input_tokens_seen": 307472, + "step": 1075 + }, + { + "epoch": 2.014925373134328, + "grad_norm": 0.8238868713378906, + "learning_rate": 4.999993505719229e-05, + "loss": 0.9381, + "num_input_tokens_seen": 308816, + "step": 1080 + }, + { + "epoch": 2.0242537313432836, + "grad_norm": 0.8195098042488098, + "learning_rate": 4.9999809147827334e-05, + "loss": 1.0955, + "num_input_tokens_seen": 310128, + "step": 1085 + }, + { + "epoch": 2.033582089552239, + "grad_norm": 1.0817739963531494, + "learning_rate": 4.9999616970783104e-05, + "loss": 0.8561, + "num_input_tokens_seen": 311472, + "step": 1090 + }, + { + "epoch": 2.042910447761194, + "grad_norm": 1.2672911882400513, + "learning_rate": 4.999935852656898e-05, + "loss": 1.5905, + "num_input_tokens_seen": 312784, + "step": 1095 + }, + { + "epoch": 2.0522388059701493, + "grad_norm": 0.9316788911819458, + "learning_rate": 4.9999033815870047e-05, + "loss": 1.117, + "num_input_tokens_seen": 314192, + "step": 1100 + }, + { + "epoch": 2.0615671641791047, + "grad_norm": 0.6610453724861145, + "learning_rate": 4.999864283954701e-05, + "loss": 1.2185, + "num_input_tokens_seen": 315536, + "step": 1105 + }, + { + "epoch": 2.0708955223880596, + "grad_norm": 0.6827360391616821, + "learning_rate": 4.999818559863626e-05, + "loss": 1.13, + "num_input_tokens_seen": 316912, + "step": 1110 + }, + { + "epoch": 2.080223880597015, + "grad_norm": 0.5703538060188293, + "learning_rate": 4.99976620943498e-05, + "loss": 1.1319, + "num_input_tokens_seen": 318544, + "step": 1115 + }, + { + "epoch": 2.08955223880597, + "grad_norm": 0.909595787525177, + "learning_rate": 4.9997072328075315e-05, + "loss": 1.0894, + "num_input_tokens_seen": 319856, + "step": 1120 + }, + { + "epoch": 2.0988805970149254, + "grad_norm": 0.6943990588188171, + "learning_rate": 4.999641630137609e-05, + "loss": 0.847, + "num_input_tokens_seen": 321616, + "step": 1125 + }, + { + "epoch": 2.1082089552238807, + "grad_norm": 0.7988004088401794, + "learning_rate": 4.9995694015991105e-05, + "loss": 1.0244, + "num_input_tokens_seen": 322896, + "step": 1130 + }, + { + "epoch": 2.1175373134328357, + "grad_norm": 0.6587883830070496, + "learning_rate": 4.9994905473834905e-05, + "loss": 0.987, + "num_input_tokens_seen": 324368, + "step": 1135 + }, + { + "epoch": 2.126865671641791, + "grad_norm": 0.5682168006896973, + "learning_rate": 4.999405067699773e-05, + "loss": 1.209, + "num_input_tokens_seen": 325744, + "step": 1140 + }, + { + "epoch": 2.1361940298507465, + "grad_norm": 0.6682947874069214, + "learning_rate": 4.99931296277454e-05, + "loss": 1.0224, + "num_input_tokens_seen": 327344, + "step": 1145 + }, + { + "epoch": 2.1455223880597014, + "grad_norm": 0.9913126230239868, + "learning_rate": 4.999214232851937e-05, + "loss": 0.878, + "num_input_tokens_seen": 328656, + "step": 1150 + }, + { + "epoch": 2.154850746268657, + "grad_norm": 0.7120636701583862, + "learning_rate": 4.99910887819367e-05, + "loss": 1.439, + "num_input_tokens_seen": 330128, + "step": 1155 + }, + { + "epoch": 2.1641791044776117, + "grad_norm": 1.0704147815704346, + "learning_rate": 4.998996899079005e-05, + "loss": 1.0681, + "num_input_tokens_seen": 331472, + "step": 1160 + }, + { + "epoch": 2.173507462686567, + "grad_norm": 0.9248518347740173, + "learning_rate": 4.998878295804768e-05, + "loss": 0.8161, + "num_input_tokens_seen": 332912, + "step": 1165 + }, + { + "epoch": 2.1828358208955225, + "grad_norm": 0.7334563136100769, + "learning_rate": 4.998753068685346e-05, + "loss": 0.706, + "num_input_tokens_seen": 334064, + "step": 1170 + }, + { + "epoch": 2.1921641791044775, + "grad_norm": 0.8477208614349365, + "learning_rate": 4.998621218052679e-05, + "loss": 1.5075, + "num_input_tokens_seen": 335344, + "step": 1175 + }, + { + "epoch": 2.201492537313433, + "grad_norm": 0.6208241581916809, + "learning_rate": 4.99848274425627e-05, + "loss": 1.022, + "num_input_tokens_seen": 336880, + "step": 1180 + }, + { + "epoch": 2.2108208955223883, + "grad_norm": 0.9742389917373657, + "learning_rate": 4.998337647663173e-05, + "loss": 1.1986, + "num_input_tokens_seen": 338032, + "step": 1185 + }, + { + "epoch": 2.220149253731343, + "grad_norm": 0.4950648546218872, + "learning_rate": 4.998185928658e-05, + "loss": 0.9244, + "num_input_tokens_seen": 339632, + "step": 1190 + }, + { + "epoch": 2.2294776119402986, + "grad_norm": 0.3771772086620331, + "learning_rate": 4.998027587642917e-05, + "loss": 0.9206, + "num_input_tokens_seen": 341104, + "step": 1195 + }, + { + "epoch": 2.2388059701492535, + "grad_norm": 0.9318837523460388, + "learning_rate": 4.9978626250376435e-05, + "loss": 1.3159, + "num_input_tokens_seen": 342448, + "step": 1200 + }, + { + "epoch": 2.248134328358209, + "grad_norm": 0.7884561419487, + "learning_rate": 4.99769104127945e-05, + "loss": 1.1778, + "num_input_tokens_seen": 343888, + "step": 1205 + }, + { + "epoch": 2.2574626865671643, + "grad_norm": 0.827698826789856, + "learning_rate": 4.9975128368231574e-05, + "loss": 1.0434, + "num_input_tokens_seen": 345392, + "step": 1210 + }, + { + "epoch": 2.2667910447761193, + "grad_norm": 0.6649134755134583, + "learning_rate": 4.997328012141138e-05, + "loss": 0.9794, + "num_input_tokens_seen": 346704, + "step": 1215 + }, + { + "epoch": 2.2761194029850746, + "grad_norm": 0.5411231517791748, + "learning_rate": 4.997136567723311e-05, + "loss": 1.6011, + "num_input_tokens_seen": 348080, + "step": 1220 + }, + { + "epoch": 2.28544776119403, + "grad_norm": 0.9752087593078613, + "learning_rate": 4.9969385040771445e-05, + "loss": 1.1857, + "num_input_tokens_seen": 349616, + "step": 1225 + }, + { + "epoch": 2.294776119402985, + "grad_norm": 0.7940465211868286, + "learning_rate": 4.99673382172765e-05, + "loss": 0.861, + "num_input_tokens_seen": 351152, + "step": 1230 + }, + { + "epoch": 2.3041044776119404, + "grad_norm": 0.5355964303016663, + "learning_rate": 4.996522521217386e-05, + "loss": 1.2815, + "num_input_tokens_seen": 352784, + "step": 1235 + }, + { + "epoch": 2.3134328358208958, + "grad_norm": 0.21570368111133575, + "learning_rate": 4.996304603106451e-05, + "loss": 1.2083, + "num_input_tokens_seen": 354032, + "step": 1240 + }, + { + "epoch": 2.3227611940298507, + "grad_norm": 1.1430596113204956, + "learning_rate": 4.996080067972487e-05, + "loss": 0.921, + "num_input_tokens_seen": 355632, + "step": 1245 + }, + { + "epoch": 2.332089552238806, + "grad_norm": 0.7686033844947815, + "learning_rate": 4.995848916410677e-05, + "loss": 1.1927, + "num_input_tokens_seen": 357072, + "step": 1250 + }, + { + "epoch": 2.341417910447761, + "grad_norm": 0.5642296075820923, + "learning_rate": 4.995611149033739e-05, + "loss": 0.9771, + "num_input_tokens_seen": 358416, + "step": 1255 + }, + { + "epoch": 2.3507462686567164, + "grad_norm": 0.47573304176330566, + "learning_rate": 4.995366766471929e-05, + "loss": 0.7563, + "num_input_tokens_seen": 359920, + "step": 1260 + }, + { + "epoch": 2.360074626865672, + "grad_norm": 1.1514606475830078, + "learning_rate": 4.99511576937304e-05, + "loss": 1.173, + "num_input_tokens_seen": 361328, + "step": 1265 + }, + { + "epoch": 2.3694029850746268, + "grad_norm": 0.927814781665802, + "learning_rate": 4.9948581584023965e-05, + "loss": 0.881, + "num_input_tokens_seen": 362800, + "step": 1270 + }, + { + "epoch": 2.378731343283582, + "grad_norm": 0.9004287719726562, + "learning_rate": 4.994593934242855e-05, + "loss": 1.3498, + "num_input_tokens_seen": 364112, + "step": 1275 + }, + { + "epoch": 2.388059701492537, + "grad_norm": 0.7968723773956299, + "learning_rate": 4.9943230975948016e-05, + "loss": 0.9902, + "num_input_tokens_seen": 365488, + "step": 1280 + }, + { + "epoch": 2.3973880597014925, + "grad_norm": 0.6904419660568237, + "learning_rate": 4.99404564917615e-05, + "loss": 0.9548, + "num_input_tokens_seen": 367120, + "step": 1285 + }, + { + "epoch": 2.406716417910448, + "grad_norm": 0.9925227165222168, + "learning_rate": 4.993761589722341e-05, + "loss": 1.2151, + "num_input_tokens_seen": 368368, + "step": 1290 + }, + { + "epoch": 2.416044776119403, + "grad_norm": 0.5743345618247986, + "learning_rate": 4.9934709199863386e-05, + "loss": 0.759, + "num_input_tokens_seen": 369776, + "step": 1295 + }, + { + "epoch": 2.425373134328358, + "grad_norm": 0.8110805749893188, + "learning_rate": 4.993173640738629e-05, + "loss": 1.4178, + "num_input_tokens_seen": 371120, + "step": 1300 + }, + { + "epoch": 2.4347014925373136, + "grad_norm": 0.8734095096588135, + "learning_rate": 4.992869752767218e-05, + "loss": 1.1693, + "num_input_tokens_seen": 372464, + "step": 1305 + }, + { + "epoch": 2.4440298507462686, + "grad_norm": 0.7220936417579651, + "learning_rate": 4.99255925687763e-05, + "loss": 1.0819, + "num_input_tokens_seen": 374096, + "step": 1310 + }, + { + "epoch": 2.453358208955224, + "grad_norm": 1.0692979097366333, + "learning_rate": 4.992242153892906e-05, + "loss": 0.9744, + "num_input_tokens_seen": 375632, + "step": 1315 + }, + { + "epoch": 2.4626865671641793, + "grad_norm": 0.8209245204925537, + "learning_rate": 4.991918444653598e-05, + "loss": 0.8906, + "num_input_tokens_seen": 377296, + "step": 1320 + }, + { + "epoch": 2.4720149253731343, + "grad_norm": 0.8805601596832275, + "learning_rate": 4.9915881300177725e-05, + "loss": 0.9994, + "num_input_tokens_seen": 378640, + "step": 1325 + }, + { + "epoch": 2.4813432835820897, + "grad_norm": 0.6118848919868469, + "learning_rate": 4.991251210861002e-05, + "loss": 0.819, + "num_input_tokens_seen": 380048, + "step": 1330 + }, + { + "epoch": 2.4906716417910446, + "grad_norm": 0.669678807258606, + "learning_rate": 4.9909076880763684e-05, + "loss": 1.0021, + "num_input_tokens_seen": 381648, + "step": 1335 + }, + { + "epoch": 2.5, + "grad_norm": 0.6483420729637146, + "learning_rate": 4.9905575625744564e-05, + "loss": 0.8689, + "num_input_tokens_seen": 383184, + "step": 1340 + }, + { + "epoch": 2.5093283582089554, + "grad_norm": 1.2593048810958862, + "learning_rate": 4.990200835283353e-05, + "loss": 1.104, + "num_input_tokens_seen": 384688, + "step": 1345 + }, + { + "epoch": 2.5186567164179103, + "grad_norm": 0.829369843006134, + "learning_rate": 4.9898375071486465e-05, + "loss": 1.1996, + "num_input_tokens_seen": 386032, + "step": 1350 + }, + { + "epoch": 2.5279850746268657, + "grad_norm": 0.5561212301254272, + "learning_rate": 4.989467579133419e-05, + "loss": 0.9894, + "num_input_tokens_seen": 387312, + "step": 1355 + }, + { + "epoch": 2.5373134328358207, + "grad_norm": 0.7098934054374695, + "learning_rate": 4.98909105221825e-05, + "loss": 0.8448, + "num_input_tokens_seen": 388688, + "step": 1360 + }, + { + "epoch": 2.546641791044776, + "grad_norm": 0.4232780933380127, + "learning_rate": 4.98870792740121e-05, + "loss": 0.9069, + "num_input_tokens_seen": 390352, + "step": 1365 + }, + { + "epoch": 2.5559701492537314, + "grad_norm": 0.7883185744285583, + "learning_rate": 4.988318205697856e-05, + "loss": 0.9542, + "num_input_tokens_seen": 391824, + "step": 1370 + }, + { + "epoch": 2.5652985074626864, + "grad_norm": 1.0433207750320435, + "learning_rate": 4.9879218881412366e-05, + "loss": 1.0168, + "num_input_tokens_seen": 393264, + "step": 1375 + }, + { + "epoch": 2.574626865671642, + "grad_norm": 0.5009554624557495, + "learning_rate": 4.9875189757818805e-05, + "loss": 1.1053, + "num_input_tokens_seen": 394736, + "step": 1380 + }, + { + "epoch": 2.583955223880597, + "grad_norm": 0.9394410848617554, + "learning_rate": 4.9871094696877995e-05, + "loss": 0.9154, + "num_input_tokens_seen": 396176, + "step": 1385 + }, + { + "epoch": 2.593283582089552, + "grad_norm": 0.4841640889644623, + "learning_rate": 4.986693370944481e-05, + "loss": 0.8661, + "num_input_tokens_seen": 397488, + "step": 1390 + }, + { + "epoch": 2.6026119402985075, + "grad_norm": 0.808097779750824, + "learning_rate": 4.986270680654891e-05, + "loss": 0.9676, + "num_input_tokens_seen": 398864, + "step": 1395 + }, + { + "epoch": 2.611940298507463, + "grad_norm": 0.5548291802406311, + "learning_rate": 4.985841399939465e-05, + "loss": 0.9366, + "num_input_tokens_seen": 400368, + "step": 1400 + }, + { + "epoch": 2.621268656716418, + "grad_norm": 0.45987874269485474, + "learning_rate": 4.98540552993611e-05, + "loss": 0.8253, + "num_input_tokens_seen": 401616, + "step": 1405 + }, + { + "epoch": 2.6305970149253732, + "grad_norm": 0.5872688293457031, + "learning_rate": 4.9849630718001986e-05, + "loss": 1.1314, + "num_input_tokens_seen": 403088, + "step": 1410 + }, + { + "epoch": 2.6399253731343286, + "grad_norm": 0.6833263635635376, + "learning_rate": 4.9845140267045654e-05, + "loss": 0.8905, + "num_input_tokens_seen": 404560, + "step": 1415 + }, + { + "epoch": 2.6492537313432836, + "grad_norm": 0.8479456901550293, + "learning_rate": 4.984058395839508e-05, + "loss": 1.0808, + "num_input_tokens_seen": 405968, + "step": 1420 + }, + { + "epoch": 2.658582089552239, + "grad_norm": 0.9290595650672913, + "learning_rate": 4.983596180412778e-05, + "loss": 0.6998, + "num_input_tokens_seen": 407408, + "step": 1425 + }, + { + "epoch": 2.667910447761194, + "grad_norm": 0.7048742771148682, + "learning_rate": 4.983127381649585e-05, + "loss": 0.8699, + "num_input_tokens_seen": 408944, + "step": 1430 + }, + { + "epoch": 2.6772388059701493, + "grad_norm": 0.6732962727546692, + "learning_rate": 4.9826520007925846e-05, + "loss": 0.9525, + "num_input_tokens_seen": 410384, + "step": 1435 + }, + { + "epoch": 2.6865671641791042, + "grad_norm": 0.9075595736503601, + "learning_rate": 4.982170039101883e-05, + "loss": 1.0679, + "num_input_tokens_seen": 411760, + "step": 1440 + }, + { + "epoch": 2.6958955223880596, + "grad_norm": 0.6941021680831909, + "learning_rate": 4.981681497855029e-05, + "loss": 0.7066, + "num_input_tokens_seen": 413104, + "step": 1445 + }, + { + "epoch": 2.705223880597015, + "grad_norm": 1.181061029434204, + "learning_rate": 4.981186378347011e-05, + "loss": 1.1281, + "num_input_tokens_seen": 414640, + "step": 1450 + }, + { + "epoch": 2.71455223880597, + "grad_norm": 0.687186598777771, + "learning_rate": 4.9806846818902575e-05, + "loss": 0.8417, + "num_input_tokens_seen": 416048, + "step": 1455 + }, + { + "epoch": 2.7238805970149254, + "grad_norm": 1.3342939615249634, + "learning_rate": 4.980176409814629e-05, + "loss": 1.4869, + "num_input_tokens_seen": 417520, + "step": 1460 + }, + { + "epoch": 2.7332089552238807, + "grad_norm": 0.7983704209327698, + "learning_rate": 4.9796615634674155e-05, + "loss": 0.9103, + "num_input_tokens_seen": 419120, + "step": 1465 + }, + { + "epoch": 2.7425373134328357, + "grad_norm": 0.7783291339874268, + "learning_rate": 4.9791401442133345e-05, + "loss": 0.8103, + "num_input_tokens_seen": 420592, + "step": 1470 + }, + { + "epoch": 2.751865671641791, + "grad_norm": 0.6954224705696106, + "learning_rate": 4.9786121534345265e-05, + "loss": 0.8192, + "num_input_tokens_seen": 421776, + "step": 1475 + }, + { + "epoch": 2.7611940298507465, + "grad_norm": 0.8065799474716187, + "learning_rate": 4.978077592530551e-05, + "loss": 0.8751, + "num_input_tokens_seen": 423376, + "step": 1480 + }, + { + "epoch": 2.7705223880597014, + "grad_norm": 0.7990866303443909, + "learning_rate": 4.977536462918383e-05, + "loss": 1.2996, + "num_input_tokens_seen": 424656, + "step": 1485 + }, + { + "epoch": 2.779850746268657, + "grad_norm": 0.8986721038818359, + "learning_rate": 4.9769887660324094e-05, + "loss": 0.8328, + "num_input_tokens_seen": 426224, + "step": 1490 + }, + { + "epoch": 2.789179104477612, + "grad_norm": 0.46251341700553894, + "learning_rate": 4.9764345033244265e-05, + "loss": 0.9605, + "num_input_tokens_seen": 427664, + "step": 1495 + }, + { + "epoch": 2.798507462686567, + "grad_norm": 0.5143011212348938, + "learning_rate": 4.9758736762636326e-05, + "loss": 0.6591, + "num_input_tokens_seen": 429072, + "step": 1500 + }, + { + "epoch": 2.8078358208955225, + "grad_norm": 0.8893252015113831, + "learning_rate": 4.9753062863366276e-05, + "loss": 1.0129, + "num_input_tokens_seen": 430512, + "step": 1505 + }, + { + "epoch": 2.8171641791044775, + "grad_norm": 1.4553245306015015, + "learning_rate": 4.974732335047408e-05, + "loss": 1.2147, + "num_input_tokens_seen": 431696, + "step": 1510 + }, + { + "epoch": 2.826492537313433, + "grad_norm": 0.5936267375946045, + "learning_rate": 4.974151823917361e-05, + "loss": 0.8709, + "num_input_tokens_seen": 433072, + "step": 1515 + }, + { + "epoch": 2.835820895522388, + "grad_norm": 0.6565489172935486, + "learning_rate": 4.973564754485265e-05, + "loss": 0.8886, + "num_input_tokens_seen": 434448, + "step": 1520 + }, + { + "epoch": 2.845149253731343, + "grad_norm": 0.6027125716209412, + "learning_rate": 4.97297112830728e-05, + "loss": 0.8748, + "num_input_tokens_seen": 435920, + "step": 1525 + }, + { + "epoch": 2.8544776119402986, + "grad_norm": 0.7925967574119568, + "learning_rate": 4.9723709469569476e-05, + "loss": 0.9201, + "num_input_tokens_seen": 437552, + "step": 1530 + }, + { + "epoch": 2.8638059701492535, + "grad_norm": 0.2911463975906372, + "learning_rate": 4.9717642120251846e-05, + "loss": 0.8428, + "num_input_tokens_seen": 439024, + "step": 1535 + }, + { + "epoch": 2.873134328358209, + "grad_norm": 1.0407238006591797, + "learning_rate": 4.97115092512028e-05, + "loss": 1.038, + "num_input_tokens_seen": 440336, + "step": 1540 + }, + { + "epoch": 2.8824626865671643, + "grad_norm": 0.818378210067749, + "learning_rate": 4.97053108786789e-05, + "loss": 1.2063, + "num_input_tokens_seen": 441648, + "step": 1545 + }, + { + "epoch": 2.8917910447761193, + "grad_norm": 0.6363953948020935, + "learning_rate": 4.9699047019110346e-05, + "loss": 0.8156, + "num_input_tokens_seen": 443152, + "step": 1550 + }, + { + "epoch": 2.9011194029850746, + "grad_norm": 0.6538462042808533, + "learning_rate": 4.969271768910093e-05, + "loss": 0.8972, + "num_input_tokens_seen": 444624, + "step": 1555 + }, + { + "epoch": 2.91044776119403, + "grad_norm": 0.8766194581985474, + "learning_rate": 4.968632290542796e-05, + "loss": 0.9455, + "num_input_tokens_seen": 446064, + "step": 1560 + }, + { + "epoch": 2.919776119402985, + "grad_norm": 0.4146958291530609, + "learning_rate": 4.9679862685042275e-05, + "loss": 0.8417, + "num_input_tokens_seen": 447600, + "step": 1565 + }, + { + "epoch": 2.9291044776119404, + "grad_norm": 0.5041606426239014, + "learning_rate": 4.967333704506817e-05, + "loss": 0.9226, + "num_input_tokens_seen": 449136, + "step": 1570 + }, + { + "epoch": 2.9384328358208958, + "grad_norm": 0.8427011966705322, + "learning_rate": 4.966674600280332e-05, + "loss": 1.0405, + "num_input_tokens_seen": 450416, + "step": 1575 + }, + { + "epoch": 2.9477611940298507, + "grad_norm": 0.8359346389770508, + "learning_rate": 4.9660089575718786e-05, + "loss": 1.1211, + "num_input_tokens_seen": 451792, + "step": 1580 + }, + { + "epoch": 2.957089552238806, + "grad_norm": 0.7467334866523743, + "learning_rate": 4.965336778145895e-05, + "loss": 1.3242, + "num_input_tokens_seen": 453264, + "step": 1585 + }, + { + "epoch": 2.966417910447761, + "grad_norm": 0.7228310108184814, + "learning_rate": 4.964658063784144e-05, + "loss": 1.1067, + "num_input_tokens_seen": 454736, + "step": 1590 + }, + { + "epoch": 2.9757462686567164, + "grad_norm": 0.5675063133239746, + "learning_rate": 4.963972816285715e-05, + "loss": 0.9799, + "num_input_tokens_seen": 456016, + "step": 1595 + }, + { + "epoch": 2.9850746268656714, + "grad_norm": 0.6983289122581482, + "learning_rate": 4.9632810374670094e-05, + "loss": 0.6799, + "num_input_tokens_seen": 457424, + "step": 1600 + }, + { + "epoch": 2.9944029850746268, + "grad_norm": 0.6364997029304504, + "learning_rate": 4.9625827291617454e-05, + "loss": 1.1588, + "num_input_tokens_seen": 458832, + "step": 1605 + }, + { + "epoch": 3.0, + "eval_loss": 0.9418517351150513, + "eval_runtime": 2.4621, + "eval_samples_per_second": 96.665, + "eval_steps_per_second": 24.369, + "num_input_tokens_seen": 459376, + "step": 1608 + }, + { + "epoch": 3.003731343283582, + "grad_norm": 0.8103802800178528, + "learning_rate": 4.961877893220949e-05, + "loss": 0.7492, + "num_input_tokens_seen": 459952, + "step": 1610 + }, + { + "epoch": 3.013059701492537, + "grad_norm": 0.4922623634338379, + "learning_rate": 4.961166531512947e-05, + "loss": 0.7371, + "num_input_tokens_seen": 461392, + "step": 1615 + }, + { + "epoch": 3.0223880597014925, + "grad_norm": 0.6636347770690918, + "learning_rate": 4.9604486459233655e-05, + "loss": 0.8566, + "num_input_tokens_seen": 462960, + "step": 1620 + }, + { + "epoch": 3.031716417910448, + "grad_norm": 0.6140598058700562, + "learning_rate": 4.959724238355123e-05, + "loss": 0.9432, + "num_input_tokens_seen": 464304, + "step": 1625 + }, + { + "epoch": 3.041044776119403, + "grad_norm": 0.5874170064926147, + "learning_rate": 4.9589933107284286e-05, + "loss": 0.9176, + "num_input_tokens_seen": 465648, + "step": 1630 + }, + { + "epoch": 3.050373134328358, + "grad_norm": 0.5059143304824829, + "learning_rate": 4.9582558649807696e-05, + "loss": 0.8083, + "num_input_tokens_seen": 467120, + "step": 1635 + }, + { + "epoch": 3.0597014925373136, + "grad_norm": 0.5132017135620117, + "learning_rate": 4.9575119030669157e-05, + "loss": 1.2625, + "num_input_tokens_seen": 468528, + "step": 1640 + }, + { + "epoch": 3.0690298507462686, + "grad_norm": 0.9038861393928528, + "learning_rate": 4.956761426958906e-05, + "loss": 1.0569, + "num_input_tokens_seen": 470000, + "step": 1645 + }, + { + "epoch": 3.078358208955224, + "grad_norm": 0.6921359300613403, + "learning_rate": 4.956004438646048e-05, + "loss": 1.0054, + "num_input_tokens_seen": 471376, + "step": 1650 + }, + { + "epoch": 3.0876865671641793, + "grad_norm": 0.7007588148117065, + "learning_rate": 4.955240940134912e-05, + "loss": 1.1785, + "num_input_tokens_seen": 472784, + "step": 1655 + }, + { + "epoch": 3.0970149253731343, + "grad_norm": 1.0070130825042725, + "learning_rate": 4.954470933449324e-05, + "loss": 0.8988, + "num_input_tokens_seen": 474096, + "step": 1660 + }, + { + "epoch": 3.1063432835820897, + "grad_norm": 0.8203631639480591, + "learning_rate": 4.953694420630361e-05, + "loss": 0.8576, + "num_input_tokens_seen": 475664, + "step": 1665 + }, + { + "epoch": 3.1156716417910446, + "grad_norm": 0.7540741562843323, + "learning_rate": 4.9529114037363496e-05, + "loss": 0.9056, + "num_input_tokens_seen": 477264, + "step": 1670 + }, + { + "epoch": 3.125, + "grad_norm": 0.7896616458892822, + "learning_rate": 4.952121884842851e-05, + "loss": 1.1048, + "num_input_tokens_seen": 478672, + "step": 1675 + }, + { + "epoch": 3.1343283582089554, + "grad_norm": 0.6675752401351929, + "learning_rate": 4.951325866042667e-05, + "loss": 0.8216, + "num_input_tokens_seen": 479952, + "step": 1680 + }, + { + "epoch": 3.1436567164179103, + "grad_norm": 0.5302525162696838, + "learning_rate": 4.950523349445824e-05, + "loss": 0.6157, + "num_input_tokens_seen": 481392, + "step": 1685 + }, + { + "epoch": 3.1529850746268657, + "grad_norm": 0.5145052671432495, + "learning_rate": 4.949714337179577e-05, + "loss": 1.0359, + "num_input_tokens_seen": 483120, + "step": 1690 + }, + { + "epoch": 3.1623134328358207, + "grad_norm": 0.717125654220581, + "learning_rate": 4.9488988313883956e-05, + "loss": 0.8349, + "num_input_tokens_seen": 484624, + "step": 1695 + }, + { + "epoch": 3.171641791044776, + "grad_norm": 0.5117978453636169, + "learning_rate": 4.9480768342339636e-05, + "loss": 0.7193, + "num_input_tokens_seen": 485968, + "step": 1700 + }, + { + "epoch": 3.1809701492537314, + "grad_norm": 0.38054159283638, + "learning_rate": 4.947248347895172e-05, + "loss": 0.8834, + "num_input_tokens_seen": 487472, + "step": 1705 + }, + { + "epoch": 3.1902985074626864, + "grad_norm": 0.37402284145355225, + "learning_rate": 4.946413374568112e-05, + "loss": 0.9463, + "num_input_tokens_seen": 488848, + "step": 1710 + }, + { + "epoch": 3.199626865671642, + "grad_norm": 0.7135728001594543, + "learning_rate": 4.94557191646607e-05, + "loss": 0.7664, + "num_input_tokens_seen": 490256, + "step": 1715 + }, + { + "epoch": 3.208955223880597, + "grad_norm": 0.9698038697242737, + "learning_rate": 4.944723975819522e-05, + "loss": 0.7456, + "num_input_tokens_seen": 491536, + "step": 1720 + }, + { + "epoch": 3.218283582089552, + "grad_norm": 1.1602495908737183, + "learning_rate": 4.943869554876127e-05, + "loss": 1.0839, + "num_input_tokens_seen": 493040, + "step": 1725 + }, + { + "epoch": 3.2276119402985075, + "grad_norm": 0.8880628347396851, + "learning_rate": 4.9430086559007235e-05, + "loss": 1.212, + "num_input_tokens_seen": 494352, + "step": 1730 + }, + { + "epoch": 3.236940298507463, + "grad_norm": 0.607707679271698, + "learning_rate": 4.94214128117532e-05, + "loss": 0.9598, + "num_input_tokens_seen": 495824, + "step": 1735 + }, + { + "epoch": 3.246268656716418, + "grad_norm": 0.7046701908111572, + "learning_rate": 4.941267432999088e-05, + "loss": 0.9935, + "num_input_tokens_seen": 497328, + "step": 1740 + }, + { + "epoch": 3.2555970149253732, + "grad_norm": 0.5249481797218323, + "learning_rate": 4.940387113688363e-05, + "loss": 0.8944, + "num_input_tokens_seen": 498608, + "step": 1745 + }, + { + "epoch": 3.264925373134328, + "grad_norm": 0.8343867063522339, + "learning_rate": 4.9395003255766313e-05, + "loss": 0.9051, + "num_input_tokens_seen": 500336, + "step": 1750 + }, + { + "epoch": 3.2742537313432836, + "grad_norm": 1.0522844791412354, + "learning_rate": 4.938607071014526e-05, + "loss": 1.2107, + "num_input_tokens_seen": 501808, + "step": 1755 + }, + { + "epoch": 3.283582089552239, + "grad_norm": 0.542831540107727, + "learning_rate": 4.937707352369822e-05, + "loss": 1.1221, + "num_input_tokens_seen": 503248, + "step": 1760 + }, + { + "epoch": 3.292910447761194, + "grad_norm": 0.6389322280883789, + "learning_rate": 4.936801172027428e-05, + "loss": 0.9097, + "num_input_tokens_seen": 504688, + "step": 1765 + }, + { + "epoch": 3.3022388059701493, + "grad_norm": 0.5740953683853149, + "learning_rate": 4.93588853238938e-05, + "loss": 0.9106, + "num_input_tokens_seen": 506192, + "step": 1770 + }, + { + "epoch": 3.3115671641791042, + "grad_norm": 0.5832641124725342, + "learning_rate": 4.934969435874838e-05, + "loss": 0.6408, + "num_input_tokens_seen": 507536, + "step": 1775 + }, + { + "epoch": 3.3208955223880596, + "grad_norm": 1.0188525915145874, + "learning_rate": 4.934043884920076e-05, + "loss": 1.0405, + "num_input_tokens_seen": 509008, + "step": 1780 + }, + { + "epoch": 3.330223880597015, + "grad_norm": 0.41951462626457214, + "learning_rate": 4.9331118819784773e-05, + "loss": 0.6662, + "num_input_tokens_seen": 510416, + "step": 1785 + }, + { + "epoch": 3.33955223880597, + "grad_norm": 0.5807256102561951, + "learning_rate": 4.932173429520528e-05, + "loss": 1.0887, + "num_input_tokens_seen": 511792, + "step": 1790 + }, + { + "epoch": 3.3488805970149254, + "grad_norm": 0.4270749092102051, + "learning_rate": 4.931228530033809e-05, + "loss": 0.8776, + "num_input_tokens_seen": 513328, + "step": 1795 + }, + { + "epoch": 3.3582089552238807, + "grad_norm": 0.9883014559745789, + "learning_rate": 4.930277186022992e-05, + "loss": 0.6688, + "num_input_tokens_seen": 514800, + "step": 1800 + }, + { + "epoch": 3.3675373134328357, + "grad_norm": 0.880587637424469, + "learning_rate": 4.929319400009831e-05, + "loss": 1.0507, + "num_input_tokens_seen": 516240, + "step": 1805 + }, + { + "epoch": 3.376865671641791, + "grad_norm": 0.9094780087471008, + "learning_rate": 4.9283551745331534e-05, + "loss": 1.0454, + "num_input_tokens_seen": 517552, + "step": 1810 + }, + { + "epoch": 3.3861940298507465, + "grad_norm": 0.3884633779525757, + "learning_rate": 4.927384512148861e-05, + "loss": 0.933, + "num_input_tokens_seen": 518928, + "step": 1815 + }, + { + "epoch": 3.3955223880597014, + "grad_norm": 0.7797597050666809, + "learning_rate": 4.926407415429913e-05, + "loss": 1.0549, + "num_input_tokens_seen": 520400, + "step": 1820 + }, + { + "epoch": 3.404850746268657, + "grad_norm": 0.750013530254364, + "learning_rate": 4.925423886966328e-05, + "loss": 0.651, + "num_input_tokens_seen": 521872, + "step": 1825 + }, + { + "epoch": 3.4141791044776117, + "grad_norm": 0.6904197335243225, + "learning_rate": 4.924433929365171e-05, + "loss": 0.8539, + "num_input_tokens_seen": 523280, + "step": 1830 + }, + { + "epoch": 3.423507462686567, + "grad_norm": 0.7021207213401794, + "learning_rate": 4.92343754525055e-05, + "loss": 0.9573, + "num_input_tokens_seen": 524816, + "step": 1835 + }, + { + "epoch": 3.4328358208955225, + "grad_norm": 0.3430628180503845, + "learning_rate": 4.922434737263607e-05, + "loss": 0.8023, + "num_input_tokens_seen": 526288, + "step": 1840 + }, + { + "epoch": 3.4421641791044775, + "grad_norm": 0.7586939930915833, + "learning_rate": 4.921425508062514e-05, + "loss": 0.9831, + "num_input_tokens_seen": 527632, + "step": 1845 + }, + { + "epoch": 3.451492537313433, + "grad_norm": 0.7476866841316223, + "learning_rate": 4.920409860322461e-05, + "loss": 0.9123, + "num_input_tokens_seen": 529104, + "step": 1850 + }, + { + "epoch": 3.4608208955223883, + "grad_norm": 0.7002041339874268, + "learning_rate": 4.919387796735655e-05, + "loss": 1.0619, + "num_input_tokens_seen": 530352, + "step": 1855 + }, + { + "epoch": 3.470149253731343, + "grad_norm": 0.45525532960891724, + "learning_rate": 4.918359320011306e-05, + "loss": 0.7285, + "num_input_tokens_seen": 531856, + "step": 1860 + }, + { + "epoch": 3.4794776119402986, + "grad_norm": 0.4835807681083679, + "learning_rate": 4.917324432875627e-05, + "loss": 0.7027, + "num_input_tokens_seen": 533360, + "step": 1865 + }, + { + "epoch": 3.4888059701492535, + "grad_norm": 0.9715596437454224, + "learning_rate": 4.916283138071821e-05, + "loss": 1.0218, + "num_input_tokens_seen": 535024, + "step": 1870 + }, + { + "epoch": 3.498134328358209, + "grad_norm": 0.7256936430931091, + "learning_rate": 4.9152354383600766e-05, + "loss": 1.0382, + "num_input_tokens_seen": 536464, + "step": 1875 + }, + { + "epoch": 3.5074626865671643, + "grad_norm": 0.7287065386772156, + "learning_rate": 4.91418133651756e-05, + "loss": 1.0702, + "num_input_tokens_seen": 537776, + "step": 1880 + }, + { + "epoch": 3.5167910447761193, + "grad_norm": 0.4963942766189575, + "learning_rate": 4.913120835338409e-05, + "loss": 0.7292, + "num_input_tokens_seen": 539504, + "step": 1885 + }, + { + "epoch": 3.5261194029850746, + "grad_norm": 0.8365123867988586, + "learning_rate": 4.912053937633722e-05, + "loss": 1.0227, + "num_input_tokens_seen": 540848, + "step": 1890 + }, + { + "epoch": 3.53544776119403, + "grad_norm": 0.7047020792961121, + "learning_rate": 4.910980646231554e-05, + "loss": 1.0341, + "num_input_tokens_seen": 542288, + "step": 1895 + }, + { + "epoch": 3.544776119402985, + "grad_norm": 0.36880356073379517, + "learning_rate": 4.9099009639769084e-05, + "loss": 0.9822, + "num_input_tokens_seen": 543632, + "step": 1900 + }, + { + "epoch": 3.5541044776119404, + "grad_norm": 0.38865232467651367, + "learning_rate": 4.908814893731728e-05, + "loss": 1.1924, + "num_input_tokens_seen": 545040, + "step": 1905 + }, + { + "epoch": 3.5634328358208958, + "grad_norm": 0.5657694935798645, + "learning_rate": 4.90772243837489e-05, + "loss": 0.6878, + "num_input_tokens_seen": 546512, + "step": 1910 + }, + { + "epoch": 3.5727611940298507, + "grad_norm": 0.9398228526115417, + "learning_rate": 4.906623600802195e-05, + "loss": 0.9327, + "num_input_tokens_seen": 547984, + "step": 1915 + }, + { + "epoch": 3.582089552238806, + "grad_norm": 0.65707927942276, + "learning_rate": 4.905518383926362e-05, + "loss": 1.24, + "num_input_tokens_seen": 549360, + "step": 1920 + }, + { + "epoch": 3.591417910447761, + "grad_norm": 0.3721582591533661, + "learning_rate": 4.90440679067702e-05, + "loss": 0.7977, + "num_input_tokens_seen": 550768, + "step": 1925 + }, + { + "epoch": 3.6007462686567164, + "grad_norm": 0.39796969294548035, + "learning_rate": 4.903288824000698e-05, + "loss": 0.7962, + "num_input_tokens_seen": 552112, + "step": 1930 + }, + { + "epoch": 3.6100746268656714, + "grad_norm": 0.4904060661792755, + "learning_rate": 4.902164486860823e-05, + "loss": 1.0244, + "num_input_tokens_seen": 553584, + "step": 1935 + }, + { + "epoch": 3.6194029850746268, + "grad_norm": 0.7589411735534668, + "learning_rate": 4.901033782237706e-05, + "loss": 0.7926, + "num_input_tokens_seen": 554992, + "step": 1940 + }, + { + "epoch": 3.628731343283582, + "grad_norm": 0.4022541046142578, + "learning_rate": 4.8998967131285356e-05, + "loss": 0.7408, + "num_input_tokens_seen": 556432, + "step": 1945 + }, + { + "epoch": 3.638059701492537, + "grad_norm": 0.5273334980010986, + "learning_rate": 4.898753282547373e-05, + "loss": 0.8595, + "num_input_tokens_seen": 557872, + "step": 1950 + }, + { + "epoch": 3.6473880597014925, + "grad_norm": 0.8417019844055176, + "learning_rate": 4.89760349352514e-05, + "loss": 1.2155, + "num_input_tokens_seen": 559248, + "step": 1955 + }, + { + "epoch": 3.656716417910448, + "grad_norm": 0.5158955454826355, + "learning_rate": 4.896447349109616e-05, + "loss": 0.8541, + "num_input_tokens_seen": 560656, + "step": 1960 + }, + { + "epoch": 3.666044776119403, + "grad_norm": 0.4387255609035492, + "learning_rate": 4.895284852365422e-05, + "loss": 0.7051, + "num_input_tokens_seen": 562032, + "step": 1965 + }, + { + "epoch": 3.675373134328358, + "grad_norm": 0.8183090686798096, + "learning_rate": 4.894116006374022e-05, + "loss": 0.8906, + "num_input_tokens_seen": 563440, + "step": 1970 + }, + { + "epoch": 3.6847014925373136, + "grad_norm": 0.5803379416465759, + "learning_rate": 4.8929408142337064e-05, + "loss": 0.7136, + "num_input_tokens_seen": 564976, + "step": 1975 + }, + { + "epoch": 3.6940298507462686, + "grad_norm": 0.5189712643623352, + "learning_rate": 4.891759279059591e-05, + "loss": 0.8475, + "num_input_tokens_seen": 566384, + "step": 1980 + }, + { + "epoch": 3.703358208955224, + "grad_norm": 0.5066812038421631, + "learning_rate": 4.890571403983603e-05, + "loss": 1.1771, + "num_input_tokens_seen": 567664, + "step": 1985 + }, + { + "epoch": 3.7126865671641793, + "grad_norm": 0.6285508275032043, + "learning_rate": 4.889377192154474e-05, + "loss": 0.6895, + "num_input_tokens_seen": 568976, + "step": 1990 + }, + { + "epoch": 3.7220149253731343, + "grad_norm": 0.7834022641181946, + "learning_rate": 4.888176646737735e-05, + "loss": 0.729, + "num_input_tokens_seen": 570608, + "step": 1995 + }, + { + "epoch": 3.7313432835820897, + "grad_norm": 0.5869835019111633, + "learning_rate": 4.8869697709157047e-05, + "loss": 0.9687, + "num_input_tokens_seen": 571888, + "step": 2000 + }, + { + "epoch": 3.7406716417910446, + "grad_norm": 0.47855448722839355, + "learning_rate": 4.8857565678874826e-05, + "loss": 0.7275, + "num_input_tokens_seen": 573360, + "step": 2005 + }, + { + "epoch": 3.75, + "grad_norm": 0.528303861618042, + "learning_rate": 4.884537040868938e-05, + "loss": 0.7372, + "num_input_tokens_seen": 574864, + "step": 2010 + }, + { + "epoch": 3.7593283582089554, + "grad_norm": 0.7065563797950745, + "learning_rate": 4.8833111930927056e-05, + "loss": 0.7643, + "num_input_tokens_seen": 576208, + "step": 2015 + }, + { + "epoch": 3.7686567164179103, + "grad_norm": 0.5224754214286804, + "learning_rate": 4.882079027808173e-05, + "loss": 0.7719, + "num_input_tokens_seen": 577584, + "step": 2020 + }, + { + "epoch": 3.7779850746268657, + "grad_norm": 0.6817386746406555, + "learning_rate": 4.880840548281475e-05, + "loss": 1.0075, + "num_input_tokens_seen": 578960, + "step": 2025 + }, + { + "epoch": 3.7873134328358207, + "grad_norm": 0.6030233502388, + "learning_rate": 4.879595757795482e-05, + "loss": 0.9526, + "num_input_tokens_seen": 580208, + "step": 2030 + }, + { + "epoch": 3.796641791044776, + "grad_norm": 0.5484159588813782, + "learning_rate": 4.878344659649796e-05, + "loss": 0.7969, + "num_input_tokens_seen": 581776, + "step": 2035 + }, + { + "epoch": 3.8059701492537314, + "grad_norm": 0.6507120728492737, + "learning_rate": 4.8770872571607365e-05, + "loss": 0.8364, + "num_input_tokens_seen": 583312, + "step": 2040 + }, + { + "epoch": 3.8152985074626864, + "grad_norm": 0.7006509900093079, + "learning_rate": 4.875823553661334e-05, + "loss": 0.9763, + "num_input_tokens_seen": 584816, + "step": 2045 + }, + { + "epoch": 3.824626865671642, + "grad_norm": 0.550528347492218, + "learning_rate": 4.874553552501323e-05, + "loss": 0.8319, + "num_input_tokens_seen": 586384, + "step": 2050 + }, + { + "epoch": 3.833955223880597, + "grad_norm": 0.5212632417678833, + "learning_rate": 4.8732772570471295e-05, + "loss": 0.8244, + "num_input_tokens_seen": 587696, + "step": 2055 + }, + { + "epoch": 3.843283582089552, + "grad_norm": 0.5709365606307983, + "learning_rate": 4.871994670681865e-05, + "loss": 0.822, + "num_input_tokens_seen": 589072, + "step": 2060 + }, + { + "epoch": 3.8526119402985075, + "grad_norm": 0.6631211638450623, + "learning_rate": 4.8707057968053175e-05, + "loss": 0.8181, + "num_input_tokens_seen": 590480, + "step": 2065 + }, + { + "epoch": 3.861940298507463, + "grad_norm": 0.629149317741394, + "learning_rate": 4.8694106388339393e-05, + "loss": 1.1227, + "num_input_tokens_seen": 591952, + "step": 2070 + }, + { + "epoch": 3.871268656716418, + "grad_norm": 0.8189889192581177, + "learning_rate": 4.868109200200841e-05, + "loss": 0.7863, + "num_input_tokens_seen": 593264, + "step": 2075 + }, + { + "epoch": 3.8805970149253732, + "grad_norm": 1.4549574851989746, + "learning_rate": 4.866801484355782e-05, + "loss": 0.8184, + "num_input_tokens_seen": 594768, + "step": 2080 + }, + { + "epoch": 3.8899253731343286, + "grad_norm": 0.703462541103363, + "learning_rate": 4.865487494765161e-05, + "loss": 0.7763, + "num_input_tokens_seen": 596112, + "step": 2085 + }, + { + "epoch": 3.8992537313432836, + "grad_norm": 0.49476179480552673, + "learning_rate": 4.864167234912005e-05, + "loss": 0.8525, + "num_input_tokens_seen": 597552, + "step": 2090 + }, + { + "epoch": 3.908582089552239, + "grad_norm": 0.8305684328079224, + "learning_rate": 4.8628407082959636e-05, + "loss": 1.0809, + "num_input_tokens_seen": 598864, + "step": 2095 + }, + { + "epoch": 3.917910447761194, + "grad_norm": 0.510263204574585, + "learning_rate": 4.861507918433298e-05, + "loss": 1.1617, + "num_input_tokens_seen": 600400, + "step": 2100 + }, + { + "epoch": 3.9272388059701493, + "grad_norm": 0.5133338570594788, + "learning_rate": 4.8601688688568695e-05, + "loss": 0.774, + "num_input_tokens_seen": 601872, + "step": 2105 + }, + { + "epoch": 3.9365671641791042, + "grad_norm": 0.4012562036514282, + "learning_rate": 4.858823563116135e-05, + "loss": 0.8495, + "num_input_tokens_seen": 603184, + "step": 2110 + }, + { + "epoch": 3.9458955223880596, + "grad_norm": 0.5179259777069092, + "learning_rate": 4.857472004777132e-05, + "loss": 0.9794, + "num_input_tokens_seen": 604400, + "step": 2115 + }, + { + "epoch": 3.955223880597015, + "grad_norm": 0.4794817566871643, + "learning_rate": 4.8561141974224726e-05, + "loss": 1.0841, + "num_input_tokens_seen": 605648, + "step": 2120 + }, + { + "epoch": 3.96455223880597, + "grad_norm": 0.623583972454071, + "learning_rate": 4.854750144651337e-05, + "loss": 0.6962, + "num_input_tokens_seen": 607024, + "step": 2125 + }, + { + "epoch": 3.9738805970149254, + "grad_norm": 0.861411452293396, + "learning_rate": 4.8533798500794536e-05, + "loss": 1.0734, + "num_input_tokens_seen": 608400, + "step": 2130 + }, + { + "epoch": 3.9832089552238807, + "grad_norm": 0.457359254360199, + "learning_rate": 4.852003317339102e-05, + "loss": 0.8193, + "num_input_tokens_seen": 609936, + "step": 2135 + }, + { + "epoch": 3.9925373134328357, + "grad_norm": 0.4516320824623108, + "learning_rate": 4.8506205500790944e-05, + "loss": 0.6676, + "num_input_tokens_seen": 611248, + "step": 2140 + }, + { + "epoch": 4.0, + "eval_loss": 0.853253960609436, + "eval_runtime": 2.4815, + "eval_samples_per_second": 95.912, + "eval_steps_per_second": 24.179, + "num_input_tokens_seen": 612008, + "step": 2144 + }, + { + "epoch": 4.001865671641791, + "grad_norm": 0.6006968021392822, + "learning_rate": 4.849231551964771e-05, + "loss": 0.9399, + "num_input_tokens_seen": 612264, + "step": 2145 + }, + { + "epoch": 4.0111940298507465, + "grad_norm": 0.2711598873138428, + "learning_rate": 4.847836326677986e-05, + "loss": 0.7277, + "num_input_tokens_seen": 613928, + "step": 2150 + }, + { + "epoch": 4.020522388059701, + "grad_norm": 0.40327250957489014, + "learning_rate": 4.846434877917102e-05, + "loss": 0.9723, + "num_input_tokens_seen": 615464, + "step": 2155 + }, + { + "epoch": 4.029850746268656, + "grad_norm": 0.4669245481491089, + "learning_rate": 4.845027209396977e-05, + "loss": 0.6076, + "num_input_tokens_seen": 617096, + "step": 2160 + }, + { + "epoch": 4.039179104477612, + "grad_norm": 0.6950826644897461, + "learning_rate": 4.8436133248489576e-05, + "loss": 0.893, + "num_input_tokens_seen": 618536, + "step": 2165 + }, + { + "epoch": 4.048507462686567, + "grad_norm": 0.523655354976654, + "learning_rate": 4.842193228020865e-05, + "loss": 0.8166, + "num_input_tokens_seen": 619880, + "step": 2170 + }, + { + "epoch": 4.057835820895522, + "grad_norm": 0.36943942308425903, + "learning_rate": 4.840766922676989e-05, + "loss": 0.7137, + "num_input_tokens_seen": 621384, + "step": 2175 + }, + { + "epoch": 4.067164179104478, + "grad_norm": 0.45686212182044983, + "learning_rate": 4.839334412598077e-05, + "loss": 1.1303, + "num_input_tokens_seen": 622920, + "step": 2180 + }, + { + "epoch": 4.076492537313433, + "grad_norm": 0.4610075354576111, + "learning_rate": 4.8378957015813225e-05, + "loss": 0.8777, + "num_input_tokens_seen": 624392, + "step": 2185 + }, + { + "epoch": 4.085820895522388, + "grad_norm": 0.6012514233589172, + "learning_rate": 4.8364507934403556e-05, + "loss": 0.7922, + "num_input_tokens_seen": 625800, + "step": 2190 + }, + { + "epoch": 4.095149253731344, + "grad_norm": 0.7339540719985962, + "learning_rate": 4.8349996920052356e-05, + "loss": 0.9388, + "num_input_tokens_seen": 627208, + "step": 2195 + }, + { + "epoch": 4.104477611940299, + "grad_norm": 0.494964063167572, + "learning_rate": 4.833542401122434e-05, + "loss": 0.962, + "num_input_tokens_seen": 628584, + "step": 2200 + }, + { + "epoch": 4.1138059701492535, + "grad_norm": 0.5948253273963928, + "learning_rate": 4.832078924654835e-05, + "loss": 0.9027, + "num_input_tokens_seen": 629992, + "step": 2205 + }, + { + "epoch": 4.123134328358209, + "grad_norm": 0.44517943263053894, + "learning_rate": 4.830609266481713e-05, + "loss": 0.9555, + "num_input_tokens_seen": 631368, + "step": 2210 + }, + { + "epoch": 4.132462686567164, + "grad_norm": 0.6131348609924316, + "learning_rate": 4.829133430498732e-05, + "loss": 1.1222, + "num_input_tokens_seen": 632712, + "step": 2215 + }, + { + "epoch": 4.141791044776119, + "grad_norm": 0.3560783565044403, + "learning_rate": 4.827651420617932e-05, + "loss": 0.4791, + "num_input_tokens_seen": 634056, + "step": 2220 + }, + { + "epoch": 4.151119402985074, + "grad_norm": 0.7231823801994324, + "learning_rate": 4.8261632407677174e-05, + "loss": 0.8956, + "num_input_tokens_seen": 635624, + "step": 2225 + }, + { + "epoch": 4.16044776119403, + "grad_norm": 0.31621918082237244, + "learning_rate": 4.824668894892844e-05, + "loss": 0.5216, + "num_input_tokens_seen": 636904, + "step": 2230 + }, + { + "epoch": 4.169776119402985, + "grad_norm": 0.39763179421424866, + "learning_rate": 4.82316838695442e-05, + "loss": 0.6496, + "num_input_tokens_seen": 638408, + "step": 2235 + }, + { + "epoch": 4.17910447761194, + "grad_norm": 0.517598569393158, + "learning_rate": 4.821661720929879e-05, + "loss": 0.74, + "num_input_tokens_seen": 639848, + "step": 2240 + }, + { + "epoch": 4.188432835820896, + "grad_norm": 0.9057689309120178, + "learning_rate": 4.820148900812984e-05, + "loss": 0.8148, + "num_input_tokens_seen": 641256, + "step": 2245 + }, + { + "epoch": 4.197761194029851, + "grad_norm": 0.5723500847816467, + "learning_rate": 4.8186299306138084e-05, + "loss": 0.9503, + "num_input_tokens_seen": 642664, + "step": 2250 + }, + { + "epoch": 4.207089552238806, + "grad_norm": 0.4558660686016083, + "learning_rate": 4.817104814358728e-05, + "loss": 0.7786, + "num_input_tokens_seen": 644136, + "step": 2255 + }, + { + "epoch": 4.2164179104477615, + "grad_norm": 0.45963600277900696, + "learning_rate": 4.8155735560904106e-05, + "loss": 0.5867, + "num_input_tokens_seen": 645640, + "step": 2260 + }, + { + "epoch": 4.225746268656716, + "grad_norm": 0.581231415271759, + "learning_rate": 4.814036159867803e-05, + "loss": 0.7584, + "num_input_tokens_seen": 647048, + "step": 2265 + }, + { + "epoch": 4.235074626865671, + "grad_norm": 0.2939874231815338, + "learning_rate": 4.812492629766126e-05, + "loss": 0.7185, + "num_input_tokens_seen": 648584, + "step": 2270 + }, + { + "epoch": 4.244402985074627, + "grad_norm": 0.5632947087287903, + "learning_rate": 4.810942969876855e-05, + "loss": 0.7459, + "num_input_tokens_seen": 649960, + "step": 2275 + }, + { + "epoch": 4.253731343283582, + "grad_norm": 0.7882364988327026, + "learning_rate": 4.8093871843077166e-05, + "loss": 1.0849, + "num_input_tokens_seen": 651336, + "step": 2280 + }, + { + "epoch": 4.263059701492537, + "grad_norm": 0.21400140225887299, + "learning_rate": 4.807825277182675e-05, + "loss": 0.8243, + "num_input_tokens_seen": 652712, + "step": 2285 + }, + { + "epoch": 4.272388059701493, + "grad_norm": 0.5372784733772278, + "learning_rate": 4.806257252641919e-05, + "loss": 0.759, + "num_input_tokens_seen": 654024, + "step": 2290 + }, + { + "epoch": 4.281716417910448, + "grad_norm": 0.5827494263648987, + "learning_rate": 4.804683114841855e-05, + "loss": 0.8494, + "num_input_tokens_seen": 655528, + "step": 2295 + }, + { + "epoch": 4.291044776119403, + "grad_norm": 0.4628760516643524, + "learning_rate": 4.803102867955093e-05, + "loss": 0.7361, + "num_input_tokens_seen": 656808, + "step": 2300 + }, + { + "epoch": 4.300373134328359, + "grad_norm": 0.4053623378276825, + "learning_rate": 4.8015165161704375e-05, + "loss": 0.6879, + "num_input_tokens_seen": 658088, + "step": 2305 + }, + { + "epoch": 4.309701492537314, + "grad_norm": 0.49805760383605957, + "learning_rate": 4.799924063692873e-05, + "loss": 0.9418, + "num_input_tokens_seen": 659560, + "step": 2310 + }, + { + "epoch": 4.3190298507462686, + "grad_norm": 0.8235376477241516, + "learning_rate": 4.798325514743558e-05, + "loss": 1.1064, + "num_input_tokens_seen": 660840, + "step": 2315 + }, + { + "epoch": 4.3283582089552235, + "grad_norm": 0.5123659372329712, + "learning_rate": 4.7967208735598105e-05, + "loss": 0.8163, + "num_input_tokens_seen": 662440, + "step": 2320 + }, + { + "epoch": 4.337686567164179, + "grad_norm": 0.7037407755851746, + "learning_rate": 4.795110144395096e-05, + "loss": 1.2618, + "num_input_tokens_seen": 663752, + "step": 2325 + }, + { + "epoch": 4.347014925373134, + "grad_norm": 0.374127596616745, + "learning_rate": 4.793493331519018e-05, + "loss": 0.7446, + "num_input_tokens_seen": 665128, + "step": 2330 + }, + { + "epoch": 4.356343283582089, + "grad_norm": 0.3240635395050049, + "learning_rate": 4.791870439217308e-05, + "loss": 0.6508, + "num_input_tokens_seen": 666568, + "step": 2335 + }, + { + "epoch": 4.365671641791045, + "grad_norm": 0.41617706418037415, + "learning_rate": 4.7902414717918105e-05, + "loss": 0.7572, + "num_input_tokens_seen": 668008, + "step": 2340 + }, + { + "epoch": 4.375, + "grad_norm": 0.45455697178840637, + "learning_rate": 4.788606433560474e-05, + "loss": 0.6974, + "num_input_tokens_seen": 669544, + "step": 2345 + }, + { + "epoch": 4.384328358208955, + "grad_norm": 0.48310038447380066, + "learning_rate": 4.786965328857339e-05, + "loss": 0.7164, + "num_input_tokens_seen": 671144, + "step": 2350 + }, + { + "epoch": 4.393656716417911, + "grad_norm": 0.5490239858627319, + "learning_rate": 4.785318162032527e-05, + "loss": 0.7397, + "num_input_tokens_seen": 672520, + "step": 2355 + }, + { + "epoch": 4.402985074626866, + "grad_norm": 0.6966809034347534, + "learning_rate": 4.783664937452228e-05, + "loss": 1.1397, + "num_input_tokens_seen": 673896, + "step": 2360 + }, + { + "epoch": 4.412313432835821, + "grad_norm": 0.43885648250579834, + "learning_rate": 4.78200565949869e-05, + "loss": 1.1128, + "num_input_tokens_seen": 675240, + "step": 2365 + }, + { + "epoch": 4.4216417910447765, + "grad_norm": 0.5218996405601501, + "learning_rate": 4.780340332570207e-05, + "loss": 0.8762, + "num_input_tokens_seen": 676680, + "step": 2370 + }, + { + "epoch": 4.4309701492537314, + "grad_norm": 0.5422924160957336, + "learning_rate": 4.7786689610811055e-05, + "loss": 0.5877, + "num_input_tokens_seen": 678152, + "step": 2375 + }, + { + "epoch": 4.440298507462686, + "grad_norm": 0.5626452565193176, + "learning_rate": 4.776991549461737e-05, + "loss": 0.9054, + "num_input_tokens_seen": 679752, + "step": 2380 + }, + { + "epoch": 4.449626865671641, + "grad_norm": 0.718069851398468, + "learning_rate": 4.775308102158461e-05, + "loss": 0.9045, + "num_input_tokens_seen": 681064, + "step": 2385 + }, + { + "epoch": 4.458955223880597, + "grad_norm": 0.6878509521484375, + "learning_rate": 4.773618623633639e-05, + "loss": 0.7619, + "num_input_tokens_seen": 682440, + "step": 2390 + }, + { + "epoch": 4.468283582089552, + "grad_norm": 0.40686583518981934, + "learning_rate": 4.771923118365617e-05, + "loss": 0.818, + "num_input_tokens_seen": 684040, + "step": 2395 + }, + { + "epoch": 4.477611940298507, + "grad_norm": 0.4362366199493408, + "learning_rate": 4.770221590848718e-05, + "loss": 0.7676, + "num_input_tokens_seen": 685448, + "step": 2400 + }, + { + "epoch": 4.486940298507463, + "grad_norm": 0.4988763630390167, + "learning_rate": 4.7685140455932267e-05, + "loss": 0.6888, + "num_input_tokens_seen": 686952, + "step": 2405 + }, + { + "epoch": 4.496268656716418, + "grad_norm": 0.3486083745956421, + "learning_rate": 4.76680048712538e-05, + "loss": 0.6207, + "num_input_tokens_seen": 688488, + "step": 2410 + }, + { + "epoch": 4.505597014925373, + "grad_norm": 0.3624007999897003, + "learning_rate": 4.765080919987356e-05, + "loss": 0.834, + "num_input_tokens_seen": 689928, + "step": 2415 + }, + { + "epoch": 4.514925373134329, + "grad_norm": 0.5246540307998657, + "learning_rate": 4.763355348737257e-05, + "loss": 0.8888, + "num_input_tokens_seen": 691272, + "step": 2420 + }, + { + "epoch": 4.524253731343284, + "grad_norm": 0.5089142322540283, + "learning_rate": 4.761623777949102e-05, + "loss": 0.8026, + "num_input_tokens_seen": 692584, + "step": 2425 + }, + { + "epoch": 4.5335820895522385, + "grad_norm": 0.4283466935157776, + "learning_rate": 4.759886212212814e-05, + "loss": 0.5876, + "num_input_tokens_seen": 694184, + "step": 2430 + }, + { + "epoch": 4.542910447761194, + "grad_norm": 0.6325791478157043, + "learning_rate": 4.758142656134205e-05, + "loss": 0.7506, + "num_input_tokens_seen": 695624, + "step": 2435 + }, + { + "epoch": 4.552238805970149, + "grad_norm": 0.513547956943512, + "learning_rate": 4.756393114334968e-05, + "loss": 0.8567, + "num_input_tokens_seen": 697096, + "step": 2440 + }, + { + "epoch": 4.561567164179104, + "grad_norm": 0.5923290848731995, + "learning_rate": 4.754637591452661e-05, + "loss": 0.8498, + "num_input_tokens_seen": 698472, + "step": 2445 + }, + { + "epoch": 4.57089552238806, + "grad_norm": 0.4583881199359894, + "learning_rate": 4.7528760921406955e-05, + "loss": 0.7261, + "num_input_tokens_seen": 699848, + "step": 2450 + }, + { + "epoch": 4.580223880597015, + "grad_norm": 0.3040519952774048, + "learning_rate": 4.751108621068328e-05, + "loss": 1.3091, + "num_input_tokens_seen": 701064, + "step": 2455 + }, + { + "epoch": 4.58955223880597, + "grad_norm": 0.5817825198173523, + "learning_rate": 4.7493351829206415e-05, + "loss": 0.8817, + "num_input_tokens_seen": 702536, + "step": 2460 + }, + { + "epoch": 4.598880597014926, + "grad_norm": 0.4939708709716797, + "learning_rate": 4.747555782398537e-05, + "loss": 0.9293, + "num_input_tokens_seen": 704040, + "step": 2465 + }, + { + "epoch": 4.608208955223881, + "grad_norm": 0.580519437789917, + "learning_rate": 4.74577042421872e-05, + "loss": 0.8995, + "num_input_tokens_seen": 705448, + "step": 2470 + }, + { + "epoch": 4.617537313432836, + "grad_norm": 0.6061758995056152, + "learning_rate": 4.743979113113689e-05, + "loss": 0.9684, + "num_input_tokens_seen": 706760, + "step": 2475 + }, + { + "epoch": 4.6268656716417915, + "grad_norm": 0.44554251432418823, + "learning_rate": 4.742181853831721e-05, + "loss": 0.6216, + "num_input_tokens_seen": 708200, + "step": 2480 + }, + { + "epoch": 4.6361940298507465, + "grad_norm": 0.5352923274040222, + "learning_rate": 4.740378651136861e-05, + "loss": 1.0056, + "num_input_tokens_seen": 709768, + "step": 2485 + }, + { + "epoch": 4.645522388059701, + "grad_norm": 0.45624014735221863, + "learning_rate": 4.738569509808907e-05, + "loss": 1.0827, + "num_input_tokens_seen": 711144, + "step": 2490 + }, + { + "epoch": 4.654850746268656, + "grad_norm": 0.4870242178440094, + "learning_rate": 4.7367544346434e-05, + "loss": 0.9834, + "num_input_tokens_seen": 712584, + "step": 2495 + }, + { + "epoch": 4.664179104477612, + "grad_norm": 0.45148831605911255, + "learning_rate": 4.7349334304516094e-05, + "loss": 0.8251, + "num_input_tokens_seen": 714152, + "step": 2500 + }, + { + "epoch": 4.673507462686567, + "grad_norm": 0.5345912575721741, + "learning_rate": 4.7331065020605204e-05, + "loss": 0.7264, + "num_input_tokens_seen": 715624, + "step": 2505 + }, + { + "epoch": 4.682835820895522, + "grad_norm": 0.5268016457557678, + "learning_rate": 4.7312736543128236e-05, + "loss": 0.8761, + "num_input_tokens_seen": 717032, + "step": 2510 + }, + { + "epoch": 4.692164179104478, + "grad_norm": 0.4194853901863098, + "learning_rate": 4.7294348920668974e-05, + "loss": 0.6835, + "num_input_tokens_seen": 718600, + "step": 2515 + }, + { + "epoch": 4.701492537313433, + "grad_norm": 0.4102332890033722, + "learning_rate": 4.727590220196799e-05, + "loss": 0.6783, + "num_input_tokens_seen": 720040, + "step": 2520 + }, + { + "epoch": 4.710820895522388, + "grad_norm": 0.4757780432701111, + "learning_rate": 4.7257396435922526e-05, + "loss": 0.8755, + "num_input_tokens_seen": 721288, + "step": 2525 + }, + { + "epoch": 4.720149253731344, + "grad_norm": 0.39596953988075256, + "learning_rate": 4.7238831671586305e-05, + "loss": 0.807, + "num_input_tokens_seen": 722760, + "step": 2530 + }, + { + "epoch": 4.729477611940299, + "grad_norm": 0.7331014275550842, + "learning_rate": 4.722020795816947e-05, + "loss": 0.835, + "num_input_tokens_seen": 724168, + "step": 2535 + }, + { + "epoch": 4.7388059701492535, + "grad_norm": 0.27252495288848877, + "learning_rate": 4.720152534503839e-05, + "loss": 0.9554, + "num_input_tokens_seen": 725736, + "step": 2540 + }, + { + "epoch": 4.7481343283582085, + "grad_norm": 0.7009444236755371, + "learning_rate": 4.71827838817156e-05, + "loss": 0.7535, + "num_input_tokens_seen": 727272, + "step": 2545 + }, + { + "epoch": 4.757462686567164, + "grad_norm": 0.4165675938129425, + "learning_rate": 4.716398361787959e-05, + "loss": 0.9351, + "num_input_tokens_seen": 728616, + "step": 2550 + }, + { + "epoch": 4.766791044776119, + "grad_norm": 0.5829318761825562, + "learning_rate": 4.714512460336475e-05, + "loss": 0.8948, + "num_input_tokens_seen": 729960, + "step": 2555 + }, + { + "epoch": 4.776119402985074, + "grad_norm": 0.795021653175354, + "learning_rate": 4.7126206888161194e-05, + "loss": 1.004, + "num_input_tokens_seen": 731272, + "step": 2560 + }, + { + "epoch": 4.78544776119403, + "grad_norm": 0.7469003200531006, + "learning_rate": 4.7107230522414615e-05, + "loss": 0.8848, + "num_input_tokens_seen": 732616, + "step": 2565 + }, + { + "epoch": 4.794776119402985, + "grad_norm": 0.6870582699775696, + "learning_rate": 4.7088195556426176e-05, + "loss": 0.8174, + "num_input_tokens_seen": 734120, + "step": 2570 + }, + { + "epoch": 4.80410447761194, + "grad_norm": 0.5851197838783264, + "learning_rate": 4.70691020406524e-05, + "loss": 0.8877, + "num_input_tokens_seen": 735592, + "step": 2575 + }, + { + "epoch": 4.813432835820896, + "grad_norm": 0.6174564361572266, + "learning_rate": 4.704995002570499e-05, + "loss": 0.9792, + "num_input_tokens_seen": 737000, + "step": 2580 + }, + { + "epoch": 4.822761194029851, + "grad_norm": 0.5155157446861267, + "learning_rate": 4.7030739562350713e-05, + "loss": 0.8597, + "num_input_tokens_seen": 738440, + "step": 2585 + }, + { + "epoch": 4.832089552238806, + "grad_norm": 0.42683500051498413, + "learning_rate": 4.701147070151127e-05, + "loss": 0.8309, + "num_input_tokens_seen": 739944, + "step": 2590 + }, + { + "epoch": 4.8414179104477615, + "grad_norm": 0.5841637253761292, + "learning_rate": 4.6992143494263164e-05, + "loss": 0.9175, + "num_input_tokens_seen": 741288, + "step": 2595 + }, + { + "epoch": 4.850746268656716, + "grad_norm": 0.5146892666816711, + "learning_rate": 4.697275799183755e-05, + "loss": 0.5322, + "num_input_tokens_seen": 742984, + "step": 2600 + }, + { + "epoch": 4.860074626865671, + "grad_norm": 0.6203369498252869, + "learning_rate": 4.695331424562011e-05, + "loss": 1.0274, + "num_input_tokens_seen": 744360, + "step": 2605 + }, + { + "epoch": 4.869402985074627, + "grad_norm": 0.45864784717559814, + "learning_rate": 4.693381230715091e-05, + "loss": 0.9907, + "num_input_tokens_seen": 745736, + "step": 2610 + }, + { + "epoch": 4.878731343283582, + "grad_norm": 0.8773572444915771, + "learning_rate": 4.6914252228124277e-05, + "loss": 0.912, + "num_input_tokens_seen": 747080, + "step": 2615 + }, + { + "epoch": 4.888059701492537, + "grad_norm": 0.32695674896240234, + "learning_rate": 4.689463406038866e-05, + "loss": 1.0409, + "num_input_tokens_seen": 748424, + "step": 2620 + }, + { + "epoch": 4.897388059701493, + "grad_norm": 0.4131785035133362, + "learning_rate": 4.6874957855946455e-05, + "loss": 0.7429, + "num_input_tokens_seen": 749896, + "step": 2625 + }, + { + "epoch": 4.906716417910448, + "grad_norm": 0.48450908064842224, + "learning_rate": 4.685522366695393e-05, + "loss": 0.6736, + "num_input_tokens_seen": 751208, + "step": 2630 + }, + { + "epoch": 4.916044776119403, + "grad_norm": 0.7251369953155518, + "learning_rate": 4.6835431545721034e-05, + "loss": 0.8011, + "num_input_tokens_seen": 752552, + "step": 2635 + }, + { + "epoch": 4.925373134328359, + "grad_norm": 0.7647229433059692, + "learning_rate": 4.6815581544711284e-05, + "loss": 0.9031, + "num_input_tokens_seen": 754024, + "step": 2640 + }, + { + "epoch": 4.934701492537314, + "grad_norm": 0.42396417260169983, + "learning_rate": 4.679567371654162e-05, + "loss": 0.6779, + "num_input_tokens_seen": 755336, + "step": 2645 + }, + { + "epoch": 4.9440298507462686, + "grad_norm": 0.4212580621242523, + "learning_rate": 4.677570811398228e-05, + "loss": 0.8517, + "num_input_tokens_seen": 756648, + "step": 2650 + }, + { + "epoch": 4.9533582089552235, + "grad_norm": 0.539574921131134, + "learning_rate": 4.675568478995662e-05, + "loss": 1.1393, + "num_input_tokens_seen": 757896, + "step": 2655 + }, + { + "epoch": 4.962686567164179, + "grad_norm": 0.5616459250450134, + "learning_rate": 4.673560379754103e-05, + "loss": 0.9301, + "num_input_tokens_seen": 759304, + "step": 2660 + }, + { + "epoch": 4.972014925373134, + "grad_norm": 0.42163729667663574, + "learning_rate": 4.6715465189964724e-05, + "loss": 0.8218, + "num_input_tokens_seen": 760840, + "step": 2665 + }, + { + "epoch": 4.981343283582089, + "grad_norm": 0.4331471025943756, + "learning_rate": 4.6695269020609676e-05, + "loss": 0.7897, + "num_input_tokens_seen": 762408, + "step": 2670 + }, + { + "epoch": 4.990671641791045, + "grad_norm": 0.4502560496330261, + "learning_rate": 4.667501534301043e-05, + "loss": 0.8075, + "num_input_tokens_seen": 763816, + "step": 2675 + }, + { + "epoch": 5.0, + "grad_norm": 0.7022036910057068, + "learning_rate": 4.665470421085395e-05, + "loss": 0.7588, + "num_input_tokens_seen": 764936, + "step": 2680 + }, + { + "epoch": 5.0, + "eval_loss": 0.811777651309967, + "eval_runtime": 2.4569, + "eval_samples_per_second": 96.872, + "eval_steps_per_second": 24.421, + "num_input_tokens_seen": 764936, + "step": 2680 + }, + { + "epoch": 5.009328358208955, + "grad_norm": 0.5874234437942505, + "learning_rate": 4.663433567797952e-05, + "loss": 0.7796, + "num_input_tokens_seen": 766312, + "step": 2685 + }, + { + "epoch": 5.018656716417911, + "grad_norm": 0.717956006526947, + "learning_rate": 4.661390979837858e-05, + "loss": 0.8541, + "num_input_tokens_seen": 767688, + "step": 2690 + }, + { + "epoch": 5.027985074626866, + "grad_norm": 0.5266153812408447, + "learning_rate": 4.659342662619454e-05, + "loss": 0.7842, + "num_input_tokens_seen": 769128, + "step": 2695 + }, + { + "epoch": 5.037313432835821, + "grad_norm": 0.39330294728279114, + "learning_rate": 4.657288621572273e-05, + "loss": 0.7526, + "num_input_tokens_seen": 770600, + "step": 2700 + }, + { + "epoch": 5.0466417910447765, + "grad_norm": 0.3557616174221039, + "learning_rate": 4.655228862141017e-05, + "loss": 0.8613, + "num_input_tokens_seen": 771880, + "step": 2705 + }, + { + "epoch": 5.0559701492537314, + "grad_norm": 0.41621243953704834, + "learning_rate": 4.653163389785546e-05, + "loss": 0.9352, + "num_input_tokens_seen": 773288, + "step": 2710 + }, + { + "epoch": 5.065298507462686, + "grad_norm": 0.5609784722328186, + "learning_rate": 4.651092209980866e-05, + "loss": 0.9306, + "num_input_tokens_seen": 774632, + "step": 2715 + }, + { + "epoch": 5.074626865671641, + "grad_norm": 0.5960310697555542, + "learning_rate": 4.649015328217108e-05, + "loss": 1.1506, + "num_input_tokens_seen": 776104, + "step": 2720 + }, + { + "epoch": 5.083955223880597, + "grad_norm": 0.45329827070236206, + "learning_rate": 4.6469327499995214e-05, + "loss": 0.835, + "num_input_tokens_seen": 777704, + "step": 2725 + }, + { + "epoch": 5.093283582089552, + "grad_norm": 0.5969468355178833, + "learning_rate": 4.644844480848452e-05, + "loss": 0.9937, + "num_input_tokens_seen": 779208, + "step": 2730 + }, + { + "epoch": 5.102611940298507, + "grad_norm": 0.6692522168159485, + "learning_rate": 4.642750526299333e-05, + "loss": 1.0487, + "num_input_tokens_seen": 780552, + "step": 2735 + }, + { + "epoch": 5.111940298507463, + "grad_norm": 0.7877621650695801, + "learning_rate": 4.640650891902666e-05, + "loss": 0.6848, + "num_input_tokens_seen": 781896, + "step": 2740 + }, + { + "epoch": 5.121268656716418, + "grad_norm": 0.3536301255226135, + "learning_rate": 4.638545583224011e-05, + "loss": 0.5179, + "num_input_tokens_seen": 783432, + "step": 2745 + }, + { + "epoch": 5.130597014925373, + "grad_norm": 0.41521406173706055, + "learning_rate": 4.6364346058439654e-05, + "loss": 0.6398, + "num_input_tokens_seen": 785256, + "step": 2750 + }, + { + "epoch": 5.139925373134329, + "grad_norm": 0.4111718237400055, + "learning_rate": 4.634317965358157e-05, + "loss": 0.7761, + "num_input_tokens_seen": 786984, + "step": 2755 + }, + { + "epoch": 5.149253731343284, + "grad_norm": 0.6352834701538086, + "learning_rate": 4.63219566737722e-05, + "loss": 1.012, + "num_input_tokens_seen": 788584, + "step": 2760 + }, + { + "epoch": 5.1585820895522385, + "grad_norm": 0.42443206906318665, + "learning_rate": 4.6300677175267914e-05, + "loss": 0.7009, + "num_input_tokens_seen": 789992, + "step": 2765 + }, + { + "epoch": 5.167910447761194, + "grad_norm": 0.4079339802265167, + "learning_rate": 4.6279341214474836e-05, + "loss": 0.806, + "num_input_tokens_seen": 791464, + "step": 2770 + }, + { + "epoch": 5.177238805970149, + "grad_norm": 0.4739495515823364, + "learning_rate": 4.625794884794879e-05, + "loss": 0.7271, + "num_input_tokens_seen": 792808, + "step": 2775 + }, + { + "epoch": 5.186567164179104, + "grad_norm": 0.4549640715122223, + "learning_rate": 4.62365001323951e-05, + "loss": 0.9343, + "num_input_tokens_seen": 794184, + "step": 2780 + }, + { + "epoch": 5.19589552238806, + "grad_norm": 0.409371018409729, + "learning_rate": 4.621499512466847e-05, + "loss": 0.7181, + "num_input_tokens_seen": 795560, + "step": 2785 + }, + { + "epoch": 5.205223880597015, + "grad_norm": 0.5138936042785645, + "learning_rate": 4.6193433881772825e-05, + "loss": 0.6005, + "num_input_tokens_seen": 796968, + "step": 2790 + }, + { + "epoch": 5.21455223880597, + "grad_norm": 0.39687439799308777, + "learning_rate": 4.617181646086112e-05, + "loss": 0.6697, + "num_input_tokens_seen": 798344, + "step": 2795 + }, + { + "epoch": 5.223880597014926, + "grad_norm": 0.39946186542510986, + "learning_rate": 4.615014291923527e-05, + "loss": 0.6044, + "num_input_tokens_seen": 799784, + "step": 2800 + }, + { + "epoch": 5.233208955223881, + "grad_norm": 0.7478322982788086, + "learning_rate": 4.61284133143459e-05, + "loss": 0.8586, + "num_input_tokens_seen": 801096, + "step": 2805 + }, + { + "epoch": 5.242537313432836, + "grad_norm": 0.46752989292144775, + "learning_rate": 4.6106627703792294e-05, + "loss": 0.9003, + "num_input_tokens_seen": 802440, + "step": 2810 + }, + { + "epoch": 5.251865671641791, + "grad_norm": 0.6429598331451416, + "learning_rate": 4.608478614532215e-05, + "loss": 0.7047, + "num_input_tokens_seen": 803912, + "step": 2815 + }, + { + "epoch": 5.2611940298507465, + "grad_norm": 0.7033519148826599, + "learning_rate": 4.6062888696831484e-05, + "loss": 0.7027, + "num_input_tokens_seen": 805224, + "step": 2820 + }, + { + "epoch": 5.270522388059701, + "grad_norm": 0.4604390263557434, + "learning_rate": 4.604093541636447e-05, + "loss": 0.6945, + "num_input_tokens_seen": 806568, + "step": 2825 + }, + { + "epoch": 5.279850746268656, + "grad_norm": 0.8423784971237183, + "learning_rate": 4.601892636211328e-05, + "loss": 0.7584, + "num_input_tokens_seen": 807880, + "step": 2830 + }, + { + "epoch": 5.289179104477612, + "grad_norm": 0.24393199384212494, + "learning_rate": 4.5996861592417906e-05, + "loss": 0.7116, + "num_input_tokens_seen": 809320, + "step": 2835 + }, + { + "epoch": 5.298507462686567, + "grad_norm": 0.5150722861289978, + "learning_rate": 4.597474116576603e-05, + "loss": 0.6896, + "num_input_tokens_seen": 810856, + "step": 2840 + }, + { + "epoch": 5.307835820895522, + "grad_norm": 0.7130671739578247, + "learning_rate": 4.595256514079289e-05, + "loss": 0.944, + "num_input_tokens_seen": 812168, + "step": 2845 + }, + { + "epoch": 5.317164179104478, + "grad_norm": 0.586578905582428, + "learning_rate": 4.593033357628107e-05, + "loss": 0.8741, + "num_input_tokens_seen": 813576, + "step": 2850 + }, + { + "epoch": 5.326492537313433, + "grad_norm": 0.3901928961277008, + "learning_rate": 4.5908046531160396e-05, + "loss": 0.7657, + "num_input_tokens_seen": 815080, + "step": 2855 + }, + { + "epoch": 5.335820895522388, + "grad_norm": 0.44602951407432556, + "learning_rate": 4.588570406450774e-05, + "loss": 0.795, + "num_input_tokens_seen": 816456, + "step": 2860 + }, + { + "epoch": 5.345149253731344, + "grad_norm": 0.4720853269100189, + "learning_rate": 4.586330623554691e-05, + "loss": 0.6834, + "num_input_tokens_seen": 817896, + "step": 2865 + }, + { + "epoch": 5.354477611940299, + "grad_norm": 0.7235953211784363, + "learning_rate": 4.5840853103648415e-05, + "loss": 0.7534, + "num_input_tokens_seen": 819400, + "step": 2870 + }, + { + "epoch": 5.3638059701492535, + "grad_norm": 0.6476424932479858, + "learning_rate": 4.581834472832942e-05, + "loss": 0.844, + "num_input_tokens_seen": 820712, + "step": 2875 + }, + { + "epoch": 5.373134328358209, + "grad_norm": 0.4089563488960266, + "learning_rate": 4.579578116925347e-05, + "loss": 0.679, + "num_input_tokens_seen": 822216, + "step": 2880 + }, + { + "epoch": 5.382462686567164, + "grad_norm": 0.47244200110435486, + "learning_rate": 4.577316248623041e-05, + "loss": 0.8439, + "num_input_tokens_seen": 823592, + "step": 2885 + }, + { + "epoch": 5.391791044776119, + "grad_norm": 0.6027782559394836, + "learning_rate": 4.575048873921621e-05, + "loss": 0.8282, + "num_input_tokens_seen": 824936, + "step": 2890 + }, + { + "epoch": 5.401119402985074, + "grad_norm": 0.3451048731803894, + "learning_rate": 4.57277599883128e-05, + "loss": 0.7081, + "num_input_tokens_seen": 826312, + "step": 2895 + }, + { + "epoch": 5.41044776119403, + "grad_norm": 0.39686471223831177, + "learning_rate": 4.570497629376789e-05, + "loss": 0.9558, + "num_input_tokens_seen": 827688, + "step": 2900 + }, + { + "epoch": 5.419776119402985, + "grad_norm": 0.5175461769104004, + "learning_rate": 4.5682137715974835e-05, + "loss": 0.6596, + "num_input_tokens_seen": 829160, + "step": 2905 + }, + { + "epoch": 5.42910447761194, + "grad_norm": 0.4657241404056549, + "learning_rate": 4.565924431547251e-05, + "loss": 1.0773, + "num_input_tokens_seen": 830376, + "step": 2910 + }, + { + "epoch": 5.438432835820896, + "grad_norm": 0.3800438642501831, + "learning_rate": 4.563629615294505e-05, + "loss": 0.7337, + "num_input_tokens_seen": 831848, + "step": 2915 + }, + { + "epoch": 5.447761194029851, + "grad_norm": 0.398224800825119, + "learning_rate": 4.561329328922179e-05, + "loss": 0.6372, + "num_input_tokens_seen": 833256, + "step": 2920 + }, + { + "epoch": 5.457089552238806, + "grad_norm": 0.6577410101890564, + "learning_rate": 4.559023578527706e-05, + "loss": 0.6335, + "num_input_tokens_seen": 834728, + "step": 2925 + }, + { + "epoch": 5.4664179104477615, + "grad_norm": 0.4930517077445984, + "learning_rate": 4.5567123702230004e-05, + "loss": 1.0827, + "num_input_tokens_seen": 836136, + "step": 2930 + }, + { + "epoch": 5.475746268656716, + "grad_norm": 0.6676226854324341, + "learning_rate": 4.554395710134446e-05, + "loss": 0.9178, + "num_input_tokens_seen": 837416, + "step": 2935 + }, + { + "epoch": 5.485074626865671, + "grad_norm": 0.558907151222229, + "learning_rate": 4.5520736044028764e-05, + "loss": 0.8492, + "num_input_tokens_seen": 838888, + "step": 2940 + }, + { + "epoch": 5.494402985074627, + "grad_norm": 0.3689069449901581, + "learning_rate": 4.5497460591835615e-05, + "loss": 0.7541, + "num_input_tokens_seen": 840392, + "step": 2945 + }, + { + "epoch": 5.503731343283582, + "grad_norm": 0.34908580780029297, + "learning_rate": 4.54741308064619e-05, + "loss": 0.7921, + "num_input_tokens_seen": 841928, + "step": 2950 + }, + { + "epoch": 5.513059701492537, + "grad_norm": 0.6024903059005737, + "learning_rate": 4.545074674974851e-05, + "loss": 0.7786, + "num_input_tokens_seen": 843336, + "step": 2955 + }, + { + "epoch": 5.522388059701493, + "grad_norm": 0.5780162811279297, + "learning_rate": 4.54273084836802e-05, + "loss": 0.7619, + "num_input_tokens_seen": 844808, + "step": 2960 + }, + { + "epoch": 5.531716417910448, + "grad_norm": 0.5162168145179749, + "learning_rate": 4.540381607038544e-05, + "loss": 0.7789, + "num_input_tokens_seen": 846248, + "step": 2965 + }, + { + "epoch": 5.541044776119403, + "grad_norm": 0.38391295075416565, + "learning_rate": 4.538026957213621e-05, + "loss": 0.8142, + "num_input_tokens_seen": 847720, + "step": 2970 + }, + { + "epoch": 5.550373134328359, + "grad_norm": 0.5930619239807129, + "learning_rate": 4.535666905134784e-05, + "loss": 0.8727, + "num_input_tokens_seen": 849480, + "step": 2975 + }, + { + "epoch": 5.559701492537314, + "grad_norm": 0.6126211285591125, + "learning_rate": 4.533301457057891e-05, + "loss": 0.79, + "num_input_tokens_seen": 850856, + "step": 2980 + }, + { + "epoch": 5.5690298507462686, + "grad_norm": 0.4824865758419037, + "learning_rate": 4.530930619253097e-05, + "loss": 0.7333, + "num_input_tokens_seen": 852264, + "step": 2985 + }, + { + "epoch": 5.5783582089552235, + "grad_norm": 0.44470852613449097, + "learning_rate": 4.5285543980048484e-05, + "loss": 0.9165, + "num_input_tokens_seen": 853672, + "step": 2990 + }, + { + "epoch": 5.587686567164179, + "grad_norm": 0.3760647773742676, + "learning_rate": 4.5261727996118594e-05, + "loss": 0.6637, + "num_input_tokens_seen": 855080, + "step": 2995 + }, + { + "epoch": 5.597014925373134, + "grad_norm": 0.46618416905403137, + "learning_rate": 4.5237858303870984e-05, + "loss": 0.7098, + "num_input_tokens_seen": 856648, + "step": 3000 + }, + { + "epoch": 5.606343283582089, + "grad_norm": 0.6944224834442139, + "learning_rate": 4.52139349665777e-05, + "loss": 1.1591, + "num_input_tokens_seen": 858024, + "step": 3005 + }, + { + "epoch": 5.615671641791045, + "grad_norm": 0.4327731132507324, + "learning_rate": 4.518995804765298e-05, + "loss": 0.7983, + "num_input_tokens_seen": 859368, + "step": 3010 + }, + { + "epoch": 5.625, + "grad_norm": 0.19002006947994232, + "learning_rate": 4.516592761065311e-05, + "loss": 0.5941, + "num_input_tokens_seen": 861000, + "step": 3015 + }, + { + "epoch": 5.634328358208955, + "grad_norm": 0.4216275215148926, + "learning_rate": 4.5141843719276235e-05, + "loss": 0.5455, + "num_input_tokens_seen": 862376, + "step": 3020 + }, + { + "epoch": 5.643656716417911, + "grad_norm": 0.46826615929603577, + "learning_rate": 4.5117706437362176e-05, + "loss": 0.6364, + "num_input_tokens_seen": 863720, + "step": 3025 + }, + { + "epoch": 5.652985074626866, + "grad_norm": 0.5293009281158447, + "learning_rate": 4.5093515828892285e-05, + "loss": 0.8533, + "num_input_tokens_seen": 865096, + "step": 3030 + }, + { + "epoch": 5.662313432835821, + "grad_norm": 0.49173009395599365, + "learning_rate": 4.5069271957989276e-05, + "loss": 0.7763, + "num_input_tokens_seen": 866440, + "step": 3035 + }, + { + "epoch": 5.6716417910447765, + "grad_norm": 0.650213360786438, + "learning_rate": 4.5044974888917035e-05, + "loss": 0.8123, + "num_input_tokens_seen": 867912, + "step": 3040 + }, + { + "epoch": 5.6809701492537314, + "grad_norm": 0.388984352350235, + "learning_rate": 4.5020624686080485e-05, + "loss": 0.6183, + "num_input_tokens_seen": 869288, + "step": 3045 + }, + { + "epoch": 5.690298507462686, + "grad_norm": 0.6041466593742371, + "learning_rate": 4.499622141402536e-05, + "loss": 0.7839, + "num_input_tokens_seen": 870600, + "step": 3050 + }, + { + "epoch": 5.699626865671641, + "grad_norm": 0.8006263971328735, + "learning_rate": 4.497176513743808e-05, + "loss": 0.9703, + "num_input_tokens_seen": 872040, + "step": 3055 + }, + { + "epoch": 5.708955223880597, + "grad_norm": 0.6720362305641174, + "learning_rate": 4.494725592114559e-05, + "loss": 0.6894, + "num_input_tokens_seen": 873480, + "step": 3060 + }, + { + "epoch": 5.718283582089552, + "grad_norm": 0.45492804050445557, + "learning_rate": 4.492269383011512e-05, + "loss": 0.9097, + "num_input_tokens_seen": 874888, + "step": 3065 + }, + { + "epoch": 5.727611940298507, + "grad_norm": 0.509631335735321, + "learning_rate": 4.489807892945409e-05, + "loss": 0.7447, + "num_input_tokens_seen": 876424, + "step": 3070 + }, + { + "epoch": 5.736940298507463, + "grad_norm": 0.546761691570282, + "learning_rate": 4.487341128440987e-05, + "loss": 0.9199, + "num_input_tokens_seen": 877864, + "step": 3075 + }, + { + "epoch": 5.746268656716418, + "grad_norm": 0.43162843585014343, + "learning_rate": 4.484869096036969e-05, + "loss": 0.6906, + "num_input_tokens_seen": 879208, + "step": 3080 + }, + { + "epoch": 5.755597014925373, + "grad_norm": 0.5381603240966797, + "learning_rate": 4.482391802286038e-05, + "loss": 0.8503, + "num_input_tokens_seen": 880584, + "step": 3085 + }, + { + "epoch": 5.764925373134329, + "grad_norm": 0.5543240308761597, + "learning_rate": 4.479909253754825e-05, + "loss": 0.6823, + "num_input_tokens_seen": 881928, + "step": 3090 + }, + { + "epoch": 5.774253731343284, + "grad_norm": 0.6778239011764526, + "learning_rate": 4.4774214570238884e-05, + "loss": 1.0828, + "num_input_tokens_seen": 883208, + "step": 3095 + }, + { + "epoch": 5.7835820895522385, + "grad_norm": 0.30579066276550293, + "learning_rate": 4.474928418687699e-05, + "loss": 1.024, + "num_input_tokens_seen": 884776, + "step": 3100 + }, + { + "epoch": 5.792910447761194, + "grad_norm": 0.5902828574180603, + "learning_rate": 4.472430145354622e-05, + "loss": 1.0466, + "num_input_tokens_seen": 886184, + "step": 3105 + }, + { + "epoch": 5.802238805970149, + "grad_norm": 0.47824278473854065, + "learning_rate": 4.469926643646901e-05, + "loss": 0.7681, + "num_input_tokens_seen": 887496, + "step": 3110 + }, + { + "epoch": 5.811567164179104, + "grad_norm": 0.5615352988243103, + "learning_rate": 4.467417920200635e-05, + "loss": 0.6348, + "num_input_tokens_seen": 888872, + "step": 3115 + }, + { + "epoch": 5.82089552238806, + "grad_norm": 0.7629348635673523, + "learning_rate": 4.4649039816657654e-05, + "loss": 0.692, + "num_input_tokens_seen": 890408, + "step": 3120 + }, + { + "epoch": 5.830223880597015, + "grad_norm": 0.560597836971283, + "learning_rate": 4.462384834706058e-05, + "loss": 0.8513, + "num_input_tokens_seen": 891816, + "step": 3125 + }, + { + "epoch": 5.83955223880597, + "grad_norm": 0.49337291717529297, + "learning_rate": 4.459860485999086e-05, + "loss": 0.7973, + "num_input_tokens_seen": 893096, + "step": 3130 + }, + { + "epoch": 5.848880597014926, + "grad_norm": 0.6568264365196228, + "learning_rate": 4.457330942236209e-05, + "loss": 0.9492, + "num_input_tokens_seen": 894600, + "step": 3135 + }, + { + "epoch": 5.858208955223881, + "grad_norm": 0.3848293125629425, + "learning_rate": 4.4547962101225584e-05, + "loss": 0.649, + "num_input_tokens_seen": 896168, + "step": 3140 + }, + { + "epoch": 5.867537313432836, + "grad_norm": 0.5074536800384521, + "learning_rate": 4.452256296377017e-05, + "loss": 0.9435, + "num_input_tokens_seen": 897608, + "step": 3145 + }, + { + "epoch": 5.8768656716417915, + "grad_norm": 0.5104394555091858, + "learning_rate": 4.4497112077322044e-05, + "loss": 1.0127, + "num_input_tokens_seen": 898760, + "step": 3150 + }, + { + "epoch": 5.8861940298507465, + "grad_norm": 0.5047051906585693, + "learning_rate": 4.4471609509344575e-05, + "loss": 0.7852, + "num_input_tokens_seen": 900232, + "step": 3155 + }, + { + "epoch": 5.895522388059701, + "grad_norm": 0.866063117980957, + "learning_rate": 4.444605532743811e-05, + "loss": 0.7958, + "num_input_tokens_seen": 901736, + "step": 3160 + }, + { + "epoch": 5.904850746268656, + "grad_norm": 0.5546243786811829, + "learning_rate": 4.442044959933982e-05, + "loss": 0.7499, + "num_input_tokens_seen": 903144, + "step": 3165 + }, + { + "epoch": 5.914179104477612, + "grad_norm": 0.6332898139953613, + "learning_rate": 4.439479239292351e-05, + "loss": 0.7573, + "num_input_tokens_seen": 904744, + "step": 3170 + }, + { + "epoch": 5.923507462686567, + "grad_norm": 0.5584033131599426, + "learning_rate": 4.4369083776199446e-05, + "loss": 0.908, + "num_input_tokens_seen": 906312, + "step": 3175 + }, + { + "epoch": 5.932835820895522, + "grad_norm": 0.43037375807762146, + "learning_rate": 4.434332381731416e-05, + "loss": 0.5965, + "num_input_tokens_seen": 907784, + "step": 3180 + }, + { + "epoch": 5.942164179104478, + "grad_norm": 0.6320916414260864, + "learning_rate": 4.431751258455029e-05, + "loss": 0.7753, + "num_input_tokens_seen": 909288, + "step": 3185 + }, + { + "epoch": 5.951492537313433, + "grad_norm": 0.6697765588760376, + "learning_rate": 4.4291650146326356e-05, + "loss": 0.8303, + "num_input_tokens_seen": 910696, + "step": 3190 + }, + { + "epoch": 5.960820895522388, + "grad_norm": 0.3975605070590973, + "learning_rate": 4.4265736571196645e-05, + "loss": 0.9645, + "num_input_tokens_seen": 912040, + "step": 3195 + }, + { + "epoch": 5.970149253731344, + "grad_norm": 0.4682815670967102, + "learning_rate": 4.423977192785099e-05, + "loss": 0.7756, + "num_input_tokens_seen": 913416, + "step": 3200 + }, + { + "epoch": 5.979477611940299, + "grad_norm": 0.5929524302482605, + "learning_rate": 4.421375628511456e-05, + "loss": 0.7728, + "num_input_tokens_seen": 914824, + "step": 3205 + }, + { + "epoch": 5.9888059701492535, + "grad_norm": 0.37023910880088806, + "learning_rate": 4.4187689711947754e-05, + "loss": 0.7995, + "num_input_tokens_seen": 916264, + "step": 3210 + }, + { + "epoch": 5.9981343283582085, + "grad_norm": 0.5531956553459167, + "learning_rate": 4.416157227744594e-05, + "loss": 0.8085, + "num_input_tokens_seen": 917576, + "step": 3215 + }, + { + "epoch": 6.0, + "eval_loss": 0.7845582962036133, + "eval_runtime": 2.4531, + "eval_samples_per_second": 97.019, + "eval_steps_per_second": 24.459, + "num_input_tokens_seen": 917624, + "step": 3216 + }, + { + "epoch": 6.007462686567164, + "grad_norm": 0.5546311736106873, + "learning_rate": 4.4135404050839326e-05, + "loss": 0.4836, + "num_input_tokens_seen": 919096, + "step": 3220 + }, + { + "epoch": 6.016791044776119, + "grad_norm": 0.591301679611206, + "learning_rate": 4.4109185101492735e-05, + "loss": 0.8881, + "num_input_tokens_seen": 920504, + "step": 3225 + }, + { + "epoch": 6.026119402985074, + "grad_norm": 0.5589866638183594, + "learning_rate": 4.408291549890546e-05, + "loss": 0.8992, + "num_input_tokens_seen": 921816, + "step": 3230 + }, + { + "epoch": 6.03544776119403, + "grad_norm": 0.3907211124897003, + "learning_rate": 4.4056595312711066e-05, + "loss": 0.7159, + "num_input_tokens_seen": 923224, + "step": 3235 + }, + { + "epoch": 6.044776119402985, + "grad_norm": 0.36168015003204346, + "learning_rate": 4.403022461267718e-05, + "loss": 0.7478, + "num_input_tokens_seen": 924760, + "step": 3240 + }, + { + "epoch": 6.05410447761194, + "grad_norm": 0.502166211605072, + "learning_rate": 4.400380346870534e-05, + "loss": 0.8639, + "num_input_tokens_seen": 926072, + "step": 3245 + }, + { + "epoch": 6.063432835820896, + "grad_norm": 0.6342161297798157, + "learning_rate": 4.3977331950830805e-05, + "loss": 0.7547, + "num_input_tokens_seen": 927416, + "step": 3250 + }, + { + "epoch": 6.072761194029851, + "grad_norm": 0.5936429500579834, + "learning_rate": 4.395081012922235e-05, + "loss": 0.99, + "num_input_tokens_seen": 928792, + "step": 3255 + }, + { + "epoch": 6.082089552238806, + "grad_norm": 0.40593409538269043, + "learning_rate": 4.392423807418209e-05, + "loss": 0.7088, + "num_input_tokens_seen": 930200, + "step": 3260 + }, + { + "epoch": 6.0914179104477615, + "grad_norm": 0.4519513249397278, + "learning_rate": 4.38976158561453e-05, + "loss": 0.7968, + "num_input_tokens_seen": 931704, + "step": 3265 + }, + { + "epoch": 6.100746268656716, + "grad_norm": 0.2962663173675537, + "learning_rate": 4.3870943545680225e-05, + "loss": 0.6891, + "num_input_tokens_seen": 932920, + "step": 3270 + }, + { + "epoch": 6.110074626865671, + "grad_norm": 0.40204671025276184, + "learning_rate": 4.384422121348789e-05, + "loss": 0.5735, + "num_input_tokens_seen": 934552, + "step": 3275 + }, + { + "epoch": 6.119402985074627, + "grad_norm": 0.38663190603256226, + "learning_rate": 4.381744893040192e-05, + "loss": 0.7048, + "num_input_tokens_seen": 935960, + "step": 3280 + }, + { + "epoch": 6.128731343283582, + "grad_norm": 0.2637920081615448, + "learning_rate": 4.379062676738832e-05, + "loss": 0.6591, + "num_input_tokens_seen": 937528, + "step": 3285 + }, + { + "epoch": 6.138059701492537, + "grad_norm": 0.6439189910888672, + "learning_rate": 4.3763754795545355e-05, + "loss": 0.6028, + "num_input_tokens_seen": 938904, + "step": 3290 + }, + { + "epoch": 6.147388059701493, + "grad_norm": 0.40123599767684937, + "learning_rate": 4.373683308610328e-05, + "loss": 0.8444, + "num_input_tokens_seen": 940344, + "step": 3295 + }, + { + "epoch": 6.156716417910448, + "grad_norm": 0.5368743538856506, + "learning_rate": 4.370986171042422e-05, + "loss": 1.0517, + "num_input_tokens_seen": 941624, + "step": 3300 + }, + { + "epoch": 6.166044776119403, + "grad_norm": 0.2689199149608612, + "learning_rate": 4.368284074000193e-05, + "loss": 0.9484, + "num_input_tokens_seen": 943032, + "step": 3305 + }, + { + "epoch": 6.175373134328359, + "grad_norm": 0.40329205989837646, + "learning_rate": 4.3655770246461645e-05, + "loss": 0.8142, + "num_input_tokens_seen": 944408, + "step": 3310 + }, + { + "epoch": 6.184701492537314, + "grad_norm": 0.29619100689888, + "learning_rate": 4.362865030155986e-05, + "loss": 0.7901, + "num_input_tokens_seen": 946008, + "step": 3315 + }, + { + "epoch": 6.1940298507462686, + "grad_norm": 0.8423734903335571, + "learning_rate": 4.360148097718416e-05, + "loss": 0.8099, + "num_input_tokens_seen": 947224, + "step": 3320 + }, + { + "epoch": 6.2033582089552235, + "grad_norm": 0.384807825088501, + "learning_rate": 4.3574262345353015e-05, + "loss": 0.9094, + "num_input_tokens_seen": 948632, + "step": 3325 + }, + { + "epoch": 6.212686567164179, + "grad_norm": 0.4576065242290497, + "learning_rate": 4.354699447821558e-05, + "loss": 0.897, + "num_input_tokens_seen": 949976, + "step": 3330 + }, + { + "epoch": 6.222014925373134, + "grad_norm": 0.6495310664176941, + "learning_rate": 4.3519677448051576e-05, + "loss": 0.7452, + "num_input_tokens_seen": 951416, + "step": 3335 + }, + { + "epoch": 6.231343283582089, + "grad_norm": 0.5229567289352417, + "learning_rate": 4.3492311327270987e-05, + "loss": 0.9521, + "num_input_tokens_seen": 952888, + "step": 3340 + }, + { + "epoch": 6.240671641791045, + "grad_norm": 0.27482396364212036, + "learning_rate": 4.346489618841393e-05, + "loss": 0.7789, + "num_input_tokens_seen": 954552, + "step": 3345 + }, + { + "epoch": 6.25, + "grad_norm": 0.4559909999370575, + "learning_rate": 4.3437432104150466e-05, + "loss": 0.7075, + "num_input_tokens_seen": 956024, + "step": 3350 + }, + { + "epoch": 6.259328358208955, + "grad_norm": 0.3086194396018982, + "learning_rate": 4.34099191472804e-05, + "loss": 0.4917, + "num_input_tokens_seen": 957528, + "step": 3355 + }, + { + "epoch": 6.268656716417911, + "grad_norm": 0.6573795080184937, + "learning_rate": 4.338235739073309e-05, + "loss": 1.0902, + "num_input_tokens_seen": 958872, + "step": 3360 + }, + { + "epoch": 6.277985074626866, + "grad_norm": 0.5657520890235901, + "learning_rate": 4.335474690756722e-05, + "loss": 1.2713, + "num_input_tokens_seen": 960152, + "step": 3365 + }, + { + "epoch": 6.287313432835821, + "grad_norm": 0.7382471561431885, + "learning_rate": 4.3327087770970674e-05, + "loss": 1.1459, + "num_input_tokens_seen": 961592, + "step": 3370 + }, + { + "epoch": 6.2966417910447765, + "grad_norm": 0.6482427716255188, + "learning_rate": 4.329938005426027e-05, + "loss": 1.0662, + "num_input_tokens_seen": 962968, + "step": 3375 + }, + { + "epoch": 6.3059701492537314, + "grad_norm": 0.6091682314872742, + "learning_rate": 4.3271623830881625e-05, + "loss": 0.4875, + "num_input_tokens_seen": 964376, + "step": 3380 + }, + { + "epoch": 6.315298507462686, + "grad_norm": 0.41217947006225586, + "learning_rate": 4.324381917440891e-05, + "loss": 0.8551, + "num_input_tokens_seen": 965880, + "step": 3385 + }, + { + "epoch": 6.324626865671641, + "grad_norm": 0.6099388003349304, + "learning_rate": 4.321596615854469e-05, + "loss": 1.0389, + "num_input_tokens_seen": 967128, + "step": 3390 + }, + { + "epoch": 6.333955223880597, + "grad_norm": 0.4437558650970459, + "learning_rate": 4.318806485711972e-05, + "loss": 0.862, + "num_input_tokens_seen": 968600, + "step": 3395 + }, + { + "epoch": 6.343283582089552, + "grad_norm": 0.19739139080047607, + "learning_rate": 4.316011534409275e-05, + "loss": 0.7048, + "num_input_tokens_seen": 970296, + "step": 3400 + }, + { + "epoch": 6.352611940298507, + "grad_norm": 0.39132192730903625, + "learning_rate": 4.313211769355031e-05, + "loss": 1.1157, + "num_input_tokens_seen": 971576, + "step": 3405 + }, + { + "epoch": 6.361940298507463, + "grad_norm": 0.5656144022941589, + "learning_rate": 4.310407197970655e-05, + "loss": 0.8207, + "num_input_tokens_seen": 972952, + "step": 3410 + }, + { + "epoch": 6.371268656716418, + "grad_norm": 0.45362043380737305, + "learning_rate": 4.3075978276903e-05, + "loss": 0.6681, + "num_input_tokens_seen": 974296, + "step": 3415 + }, + { + "epoch": 6.380597014925373, + "grad_norm": 0.3806924521923065, + "learning_rate": 4.304783665960842e-05, + "loss": 0.8446, + "num_input_tokens_seen": 975480, + "step": 3420 + }, + { + "epoch": 6.389925373134329, + "grad_norm": 0.5423265695571899, + "learning_rate": 4.3019647202418566e-05, + "loss": 0.7487, + "num_input_tokens_seen": 976888, + "step": 3425 + }, + { + "epoch": 6.399253731343284, + "grad_norm": 0.3596850335597992, + "learning_rate": 4.2991409980055996e-05, + "loss": 0.778, + "num_input_tokens_seen": 978328, + "step": 3430 + }, + { + "epoch": 6.4085820895522385, + "grad_norm": 0.451376348733902, + "learning_rate": 4.2963125067369894e-05, + "loss": 0.7288, + "num_input_tokens_seen": 979672, + "step": 3435 + }, + { + "epoch": 6.417910447761194, + "grad_norm": 0.5142702460289001, + "learning_rate": 4.293479253933584e-05, + "loss": 0.7397, + "num_input_tokens_seen": 981112, + "step": 3440 + }, + { + "epoch": 6.427238805970149, + "grad_norm": 0.7385170459747314, + "learning_rate": 4.2906412471055675e-05, + "loss": 0.7874, + "num_input_tokens_seen": 982552, + "step": 3445 + }, + { + "epoch": 6.436567164179104, + "grad_norm": 0.2863917052745819, + "learning_rate": 4.287798493775719e-05, + "loss": 0.6362, + "num_input_tokens_seen": 983800, + "step": 3450 + }, + { + "epoch": 6.44589552238806, + "grad_norm": 0.6086423993110657, + "learning_rate": 4.2849510014794045e-05, + "loss": 0.8136, + "num_input_tokens_seen": 985240, + "step": 3455 + }, + { + "epoch": 6.455223880597015, + "grad_norm": 0.7482091784477234, + "learning_rate": 4.2820987777645506e-05, + "loss": 0.8652, + "num_input_tokens_seen": 986616, + "step": 3460 + }, + { + "epoch": 6.46455223880597, + "grad_norm": 0.6426039934158325, + "learning_rate": 4.2792418301916224e-05, + "loss": 0.7879, + "num_input_tokens_seen": 987864, + "step": 3465 + }, + { + "epoch": 6.473880597014926, + "grad_norm": 0.7978320717811584, + "learning_rate": 4.2763801663336114e-05, + "loss": 0.7328, + "num_input_tokens_seen": 989272, + "step": 3470 + }, + { + "epoch": 6.483208955223881, + "grad_norm": 0.5092916488647461, + "learning_rate": 4.273513793776006e-05, + "loss": 0.7991, + "num_input_tokens_seen": 990840, + "step": 3475 + }, + { + "epoch": 6.492537313432836, + "grad_norm": 0.3728201687335968, + "learning_rate": 4.2706427201167806e-05, + "loss": 0.5951, + "num_input_tokens_seen": 992216, + "step": 3480 + }, + { + "epoch": 6.5018656716417915, + "grad_norm": 0.4233628511428833, + "learning_rate": 4.267766952966369e-05, + "loss": 0.6184, + "num_input_tokens_seen": 993592, + "step": 3485 + }, + { + "epoch": 6.5111940298507465, + "grad_norm": 0.6990503072738647, + "learning_rate": 4.264886499947645e-05, + "loss": 0.903, + "num_input_tokens_seen": 995064, + "step": 3490 + }, + { + "epoch": 6.520522388059701, + "grad_norm": 0.4619406759738922, + "learning_rate": 4.262001368695904e-05, + "loss": 0.7989, + "num_input_tokens_seen": 996376, + "step": 3495 + }, + { + "epoch": 6.529850746268656, + "grad_norm": 0.519572913646698, + "learning_rate": 4.2591115668588436e-05, + "loss": 0.7607, + "num_input_tokens_seen": 997752, + "step": 3500 + }, + { + "epoch": 6.539179104477612, + "grad_norm": 0.7582955360412598, + "learning_rate": 4.25621710209654e-05, + "loss": 0.7218, + "num_input_tokens_seen": 998936, + "step": 3505 + }, + { + "epoch": 6.548507462686567, + "grad_norm": 0.360200971364975, + "learning_rate": 4.25331798208143e-05, + "loss": 0.6448, + "num_input_tokens_seen": 1000376, + "step": 3510 + }, + { + "epoch": 6.557835820895522, + "grad_norm": 0.6518470048904419, + "learning_rate": 4.2504142144982916e-05, + "loss": 0.5104, + "num_input_tokens_seen": 1002136, + "step": 3515 + }, + { + "epoch": 6.567164179104478, + "grad_norm": 0.3732702434062958, + "learning_rate": 4.2475058070442195e-05, + "loss": 0.6652, + "num_input_tokens_seen": 1003640, + "step": 3520 + }, + { + "epoch": 6.576492537313433, + "grad_norm": 0.8214306235313416, + "learning_rate": 4.2445927674286114e-05, + "loss": 0.9237, + "num_input_tokens_seen": 1004920, + "step": 3525 + }, + { + "epoch": 6.585820895522388, + "grad_norm": 0.5833376049995422, + "learning_rate": 4.241675103373139e-05, + "loss": 0.8792, + "num_input_tokens_seen": 1006328, + "step": 3530 + }, + { + "epoch": 6.595149253731344, + "grad_norm": 0.5886469483375549, + "learning_rate": 4.238752822611735e-05, + "loss": 0.7002, + "num_input_tokens_seen": 1007800, + "step": 3535 + }, + { + "epoch": 6.604477611940299, + "grad_norm": 0.5036680102348328, + "learning_rate": 4.2358259328905704e-05, + "loss": 0.8158, + "num_input_tokens_seen": 1009432, + "step": 3540 + }, + { + "epoch": 6.6138059701492535, + "grad_norm": 0.4163619577884674, + "learning_rate": 4.23289444196803e-05, + "loss": 0.5978, + "num_input_tokens_seen": 1010968, + "step": 3545 + }, + { + "epoch": 6.6231343283582085, + "grad_norm": 0.3400265574455261, + "learning_rate": 4.2299583576146984e-05, + "loss": 0.5321, + "num_input_tokens_seen": 1012344, + "step": 3550 + }, + { + "epoch": 6.632462686567164, + "grad_norm": 0.3447887897491455, + "learning_rate": 4.227017687613335e-05, + "loss": 0.6147, + "num_input_tokens_seen": 1013944, + "step": 3555 + }, + { + "epoch": 6.641791044776119, + "grad_norm": 0.5130504369735718, + "learning_rate": 4.224072439758853e-05, + "loss": 1.0356, + "num_input_tokens_seen": 1015480, + "step": 3560 + }, + { + "epoch": 6.651119402985074, + "grad_norm": 0.6612918376922607, + "learning_rate": 4.2211226218583037e-05, + "loss": 0.9579, + "num_input_tokens_seen": 1016888, + "step": 3565 + }, + { + "epoch": 6.66044776119403, + "grad_norm": 0.2880975902080536, + "learning_rate": 4.2181682417308475e-05, + "loss": 0.5091, + "num_input_tokens_seen": 1018552, + "step": 3570 + }, + { + "epoch": 6.669776119402985, + "grad_norm": 0.5477833151817322, + "learning_rate": 4.2152093072077435e-05, + "loss": 0.5853, + "num_input_tokens_seen": 1019960, + "step": 3575 + }, + { + "epoch": 6.67910447761194, + "grad_norm": 0.5366309285163879, + "learning_rate": 4.21224582613232e-05, + "loss": 0.9282, + "num_input_tokens_seen": 1021432, + "step": 3580 + }, + { + "epoch": 6.688432835820896, + "grad_norm": 0.3878781795501709, + "learning_rate": 4.2092778063599555e-05, + "loss": 0.8173, + "num_input_tokens_seen": 1022712, + "step": 3585 + }, + { + "epoch": 6.697761194029851, + "grad_norm": 0.5970755815505981, + "learning_rate": 4.206305255758063e-05, + "loss": 0.8484, + "num_input_tokens_seen": 1024184, + "step": 3590 + }, + { + "epoch": 6.707089552238806, + "grad_norm": 0.5297569632530212, + "learning_rate": 4.203328182206064e-05, + "loss": 0.753, + "num_input_tokens_seen": 1025880, + "step": 3595 + }, + { + "epoch": 6.7164179104477615, + "grad_norm": 0.5931053161621094, + "learning_rate": 4.200346593595368e-05, + "loss": 0.8469, + "num_input_tokens_seen": 1027512, + "step": 3600 + }, + { + "epoch": 6.725746268656716, + "grad_norm": 0.30700311064720154, + "learning_rate": 4.197360497829355e-05, + "loss": 0.5578, + "num_input_tokens_seen": 1029048, + "step": 3605 + }, + { + "epoch": 6.735074626865671, + "grad_norm": 0.5141916871070862, + "learning_rate": 4.19436990282335e-05, + "loss": 0.7547, + "num_input_tokens_seen": 1030520, + "step": 3610 + }, + { + "epoch": 6.744402985074627, + "grad_norm": 0.3151283860206604, + "learning_rate": 4.191374816504605e-05, + "loss": 0.6036, + "num_input_tokens_seen": 1032024, + "step": 3615 + }, + { + "epoch": 6.753731343283582, + "grad_norm": 0.422306627035141, + "learning_rate": 4.1883752468122776e-05, + "loss": 0.7448, + "num_input_tokens_seen": 1033304, + "step": 3620 + }, + { + "epoch": 6.763059701492537, + "grad_norm": 0.523743212223053, + "learning_rate": 4.18537120169741e-05, + "loss": 0.8554, + "num_input_tokens_seen": 1034744, + "step": 3625 + }, + { + "epoch": 6.772388059701493, + "grad_norm": 0.4447706639766693, + "learning_rate": 4.1823626891229055e-05, + "loss": 0.7277, + "num_input_tokens_seen": 1036216, + "step": 3630 + }, + { + "epoch": 6.781716417910448, + "grad_norm": 0.6233542561531067, + "learning_rate": 4.179349717063512e-05, + "loss": 0.8387, + "num_input_tokens_seen": 1037496, + "step": 3635 + }, + { + "epoch": 6.791044776119403, + "grad_norm": 0.3577609658241272, + "learning_rate": 4.1763322935057974e-05, + "loss": 0.9415, + "num_input_tokens_seen": 1038808, + "step": 3640 + }, + { + "epoch": 6.800373134328359, + "grad_norm": 0.4807007908821106, + "learning_rate": 4.173310426448128e-05, + "loss": 0.6804, + "num_input_tokens_seen": 1040248, + "step": 3645 + }, + { + "epoch": 6.809701492537314, + "grad_norm": 0.4388648271560669, + "learning_rate": 4.1702841239006496e-05, + "loss": 0.942, + "num_input_tokens_seen": 1041592, + "step": 3650 + }, + { + "epoch": 6.8190298507462686, + "grad_norm": 0.4591957628726959, + "learning_rate": 4.167253393885266e-05, + "loss": 0.8529, + "num_input_tokens_seen": 1042936, + "step": 3655 + }, + { + "epoch": 6.8283582089552235, + "grad_norm": 0.5531473159790039, + "learning_rate": 4.164218244435617e-05, + "loss": 0.9311, + "num_input_tokens_seen": 1044248, + "step": 3660 + }, + { + "epoch": 6.837686567164179, + "grad_norm": 0.5402145385742188, + "learning_rate": 4.161178683597054e-05, + "loss": 0.614, + "num_input_tokens_seen": 1045720, + "step": 3665 + }, + { + "epoch": 6.847014925373134, + "grad_norm": 0.3551274836063385, + "learning_rate": 4.158134719426627e-05, + "loss": 0.6729, + "num_input_tokens_seen": 1047192, + "step": 3670 + }, + { + "epoch": 6.856343283582089, + "grad_norm": 0.5176477432250977, + "learning_rate": 4.155086359993054e-05, + "loss": 1.1027, + "num_input_tokens_seen": 1048440, + "step": 3675 + }, + { + "epoch": 6.865671641791045, + "grad_norm": 0.48509055376052856, + "learning_rate": 4.152033613376704e-05, + "loss": 0.9788, + "num_input_tokens_seen": 1049848, + "step": 3680 + }, + { + "epoch": 6.875, + "grad_norm": 0.3549090027809143, + "learning_rate": 4.1489764876695775e-05, + "loss": 0.6638, + "num_input_tokens_seen": 1051384, + "step": 3685 + }, + { + "epoch": 6.884328358208955, + "grad_norm": 0.5797954797744751, + "learning_rate": 4.145914990975281e-05, + "loss": 0.8111, + "num_input_tokens_seen": 1052696, + "step": 3690 + }, + { + "epoch": 6.893656716417911, + "grad_norm": 0.37401270866394043, + "learning_rate": 4.1428491314090076e-05, + "loss": 0.6795, + "num_input_tokens_seen": 1054040, + "step": 3695 + }, + { + "epoch": 6.902985074626866, + "grad_norm": 0.5340243577957153, + "learning_rate": 4.1397789170975154e-05, + "loss": 0.7328, + "num_input_tokens_seen": 1055640, + "step": 3700 + }, + { + "epoch": 6.912313432835821, + "grad_norm": 0.41514602303504944, + "learning_rate": 4.1367043561791055e-05, + "loss": 0.7805, + "num_input_tokens_seen": 1057144, + "step": 3705 + }, + { + "epoch": 6.9216417910447765, + "grad_norm": 0.42452138662338257, + "learning_rate": 4.1336254568036e-05, + "loss": 0.5363, + "num_input_tokens_seen": 1058680, + "step": 3710 + }, + { + "epoch": 6.9309701492537314, + "grad_norm": 0.6318342089653015, + "learning_rate": 4.130542227132323e-05, + "loss": 0.6371, + "num_input_tokens_seen": 1060120, + "step": 3715 + }, + { + "epoch": 6.940298507462686, + "grad_norm": 0.4124554693698883, + "learning_rate": 4.127454675338076e-05, + "loss": 0.5887, + "num_input_tokens_seen": 1061464, + "step": 3720 + }, + { + "epoch": 6.949626865671641, + "grad_norm": 0.5735503435134888, + "learning_rate": 4.124362809605117e-05, + "loss": 0.876, + "num_input_tokens_seen": 1063128, + "step": 3725 + }, + { + "epoch": 6.958955223880597, + "grad_norm": 0.8561661839485168, + "learning_rate": 4.1212666381291386e-05, + "loss": 0.6594, + "num_input_tokens_seen": 1064440, + "step": 3730 + }, + { + "epoch": 6.968283582089552, + "grad_norm": 0.5079460740089417, + "learning_rate": 4.118166169117248e-05, + "loss": 0.8284, + "num_input_tokens_seen": 1065848, + "step": 3735 + }, + { + "epoch": 6.977611940298507, + "grad_norm": 0.4846682846546173, + "learning_rate": 4.115061410787943e-05, + "loss": 0.6047, + "num_input_tokens_seen": 1067224, + "step": 3740 + }, + { + "epoch": 6.986940298507463, + "grad_norm": 0.4107177257537842, + "learning_rate": 4.1119523713710904e-05, + "loss": 0.8082, + "num_input_tokens_seen": 1068760, + "step": 3745 + }, + { + "epoch": 6.996268656716418, + "grad_norm": 0.4177115857601166, + "learning_rate": 4.1088390591079097e-05, + "loss": 0.7514, + "num_input_tokens_seen": 1070168, + "step": 3750 + }, + { + "epoch": 7.0, + "eval_loss": 0.763595700263977, + "eval_runtime": 2.4643, + "eval_samples_per_second": 96.577, + "eval_steps_per_second": 24.347, + "num_input_tokens_seen": 1070488, + "step": 3752 + }, + { + "epoch": 7.005597014925373, + "grad_norm": 0.37497758865356445, + "learning_rate": 4.1057214822509394e-05, + "loss": 0.9672, + "num_input_tokens_seen": 1071320, + "step": 3755 + }, + { + "epoch": 7.014925373134329, + "grad_norm": 0.5106077790260315, + "learning_rate": 4.1025996490640276e-05, + "loss": 0.9643, + "num_input_tokens_seen": 1072728, + "step": 3760 + }, + { + "epoch": 7.024253731343284, + "grad_norm": 0.6881603598594666, + "learning_rate": 4.099473567822303e-05, + "loss": 0.8184, + "num_input_tokens_seen": 1074040, + "step": 3765 + }, + { + "epoch": 7.0335820895522385, + "grad_norm": 0.5341981053352356, + "learning_rate": 4.096343246812155e-05, + "loss": 0.9159, + "num_input_tokens_seen": 1075576, + "step": 3770 + }, + { + "epoch": 7.042910447761194, + "grad_norm": 0.561252772808075, + "learning_rate": 4.093208694331211e-05, + "loss": 1.1014, + "num_input_tokens_seen": 1076856, + "step": 3775 + }, + { + "epoch": 7.052238805970149, + "grad_norm": 0.39010921120643616, + "learning_rate": 4.090069918688315e-05, + "loss": 0.8042, + "num_input_tokens_seen": 1078200, + "step": 3780 + }, + { + "epoch": 7.061567164179104, + "grad_norm": 0.5346941947937012, + "learning_rate": 4.0869269282035057e-05, + "loss": 0.706, + "num_input_tokens_seen": 1079736, + "step": 3785 + }, + { + "epoch": 7.07089552238806, + "grad_norm": 0.45321395993232727, + "learning_rate": 4.083779731207994e-05, + "loss": 0.7544, + "num_input_tokens_seen": 1081048, + "step": 3790 + }, + { + "epoch": 7.080223880597015, + "grad_norm": 0.2481449842453003, + "learning_rate": 4.080628336044142e-05, + "loss": 0.5595, + "num_input_tokens_seen": 1082680, + "step": 3795 + }, + { + "epoch": 7.08955223880597, + "grad_norm": 0.513296902179718, + "learning_rate": 4.077472751065439e-05, + "loss": 0.8085, + "num_input_tokens_seen": 1084088, + "step": 3800 + }, + { + "epoch": 7.098880597014926, + "grad_norm": 0.45849934220314026, + "learning_rate": 4.074312984636479e-05, + "loss": 0.6317, + "num_input_tokens_seen": 1085592, + "step": 3805 + }, + { + "epoch": 7.108208955223881, + "grad_norm": 0.5395163893699646, + "learning_rate": 4.0711490451329435e-05, + "loss": 0.8174, + "num_input_tokens_seen": 1086904, + "step": 3810 + }, + { + "epoch": 7.117537313432836, + "grad_norm": 0.3488948345184326, + "learning_rate": 4.0679809409415734e-05, + "loss": 0.9106, + "num_input_tokens_seen": 1088344, + "step": 3815 + }, + { + "epoch": 7.126865671641791, + "grad_norm": 0.4984322786331177, + "learning_rate": 4.064808680460148e-05, + "loss": 0.5453, + "num_input_tokens_seen": 1089912, + "step": 3820 + }, + { + "epoch": 7.1361940298507465, + "grad_norm": 0.2954554259777069, + "learning_rate": 4.0616322720974664e-05, + "loss": 0.7632, + "num_input_tokens_seen": 1091736, + "step": 3825 + }, + { + "epoch": 7.145522388059701, + "grad_norm": 0.551169216632843, + "learning_rate": 4.05845172427332e-05, + "loss": 0.8062, + "num_input_tokens_seen": 1093176, + "step": 3830 + }, + { + "epoch": 7.154850746268656, + "grad_norm": 0.39911654591560364, + "learning_rate": 4.055267045418476e-05, + "loss": 0.8236, + "num_input_tokens_seen": 1094712, + "step": 3835 + }, + { + "epoch": 7.164179104477612, + "grad_norm": 0.34627410769462585, + "learning_rate": 4.052078243974648e-05, + "loss": 0.8143, + "num_input_tokens_seen": 1096312, + "step": 3840 + }, + { + "epoch": 7.173507462686567, + "grad_norm": 0.37829849123954773, + "learning_rate": 4.0488853283944806e-05, + "loss": 0.7511, + "num_input_tokens_seen": 1097592, + "step": 3845 + }, + { + "epoch": 7.182835820895522, + "grad_norm": 0.47236159443855286, + "learning_rate": 4.045688307141523e-05, + "loss": 0.722, + "num_input_tokens_seen": 1099000, + "step": 3850 + }, + { + "epoch": 7.192164179104478, + "grad_norm": 0.4111486077308655, + "learning_rate": 4.042487188690205e-05, + "loss": 0.7523, + "num_input_tokens_seen": 1100312, + "step": 3855 + }, + { + "epoch": 7.201492537313433, + "grad_norm": 0.403230756521225, + "learning_rate": 4.039281981525821e-05, + "loss": 0.7752, + "num_input_tokens_seen": 1101816, + "step": 3860 + }, + { + "epoch": 7.210820895522388, + "grad_norm": 0.33522921800613403, + "learning_rate": 4.036072694144501e-05, + "loss": 0.697, + "num_input_tokens_seen": 1103224, + "step": 3865 + }, + { + "epoch": 7.220149253731344, + "grad_norm": 0.38257619738578796, + "learning_rate": 4.032859335053189e-05, + "loss": 0.5024, + "num_input_tokens_seen": 1104664, + "step": 3870 + }, + { + "epoch": 7.229477611940299, + "grad_norm": 0.6517744064331055, + "learning_rate": 4.029641912769628e-05, + "loss": 0.7837, + "num_input_tokens_seen": 1106104, + "step": 3875 + }, + { + "epoch": 7.2388059701492535, + "grad_norm": 0.8163742423057556, + "learning_rate": 4.026420435822325e-05, + "loss": 0.7571, + "num_input_tokens_seen": 1107416, + "step": 3880 + }, + { + "epoch": 7.248134328358209, + "grad_norm": 0.4031738340854645, + "learning_rate": 4.0231949127505365e-05, + "loss": 0.9147, + "num_input_tokens_seen": 1108856, + "step": 3885 + }, + { + "epoch": 7.257462686567164, + "grad_norm": 0.4834558963775635, + "learning_rate": 4.0199653521042464e-05, + "loss": 0.6604, + "num_input_tokens_seen": 1110392, + "step": 3890 + }, + { + "epoch": 7.266791044776119, + "grad_norm": 0.566651463508606, + "learning_rate": 4.0167317624441393e-05, + "loss": 0.6995, + "num_input_tokens_seen": 1111704, + "step": 3895 + }, + { + "epoch": 7.276119402985074, + "grad_norm": 0.6094080209732056, + "learning_rate": 4.013494152341579e-05, + "loss": 0.7178, + "num_input_tokens_seen": 1113144, + "step": 3900 + }, + { + "epoch": 7.28544776119403, + "grad_norm": 0.5256794095039368, + "learning_rate": 4.010252530378589e-05, + "loss": 0.8498, + "num_input_tokens_seen": 1114488, + "step": 3905 + }, + { + "epoch": 7.294776119402985, + "grad_norm": 0.4974917471408844, + "learning_rate": 4.007006905147824e-05, + "loss": 0.9775, + "num_input_tokens_seen": 1115928, + "step": 3910 + }, + { + "epoch": 7.30410447761194, + "grad_norm": 0.34660622477531433, + "learning_rate": 4.003757285252554e-05, + "loss": 0.7786, + "num_input_tokens_seen": 1117336, + "step": 3915 + }, + { + "epoch": 7.313432835820896, + "grad_norm": 0.5105462074279785, + "learning_rate": 4.0005036793066316e-05, + "loss": 0.5355, + "num_input_tokens_seen": 1118872, + "step": 3920 + }, + { + "epoch": 7.322761194029851, + "grad_norm": 0.3051951229572296, + "learning_rate": 3.997246095934483e-05, + "loss": 0.6027, + "num_input_tokens_seen": 1120568, + "step": 3925 + }, + { + "epoch": 7.332089552238806, + "grad_norm": 0.563103973865509, + "learning_rate": 3.993984543771071e-05, + "loss": 0.8105, + "num_input_tokens_seen": 1122040, + "step": 3930 + }, + { + "epoch": 7.3414179104477615, + "grad_norm": 0.45495352149009705, + "learning_rate": 3.990719031461884e-05, + "loss": 0.6735, + "num_input_tokens_seen": 1123608, + "step": 3935 + }, + { + "epoch": 7.350746268656716, + "grad_norm": 0.48016631603240967, + "learning_rate": 3.987449567662902e-05, + "loss": 0.9157, + "num_input_tokens_seen": 1124984, + "step": 3940 + }, + { + "epoch": 7.360074626865671, + "grad_norm": 0.5409470200538635, + "learning_rate": 3.9841761610405845e-05, + "loss": 0.8911, + "num_input_tokens_seen": 1126328, + "step": 3945 + }, + { + "epoch": 7.369402985074627, + "grad_norm": 0.5551537275314331, + "learning_rate": 3.980898820271839e-05, + "loss": 0.6827, + "num_input_tokens_seen": 1127704, + "step": 3950 + }, + { + "epoch": 7.378731343283582, + "grad_norm": 0.24416472017765045, + "learning_rate": 3.977617554044004e-05, + "loss": 0.7801, + "num_input_tokens_seen": 1129368, + "step": 3955 + }, + { + "epoch": 7.388059701492537, + "grad_norm": 0.7265456318855286, + "learning_rate": 3.9743323710548196e-05, + "loss": 0.7552, + "num_input_tokens_seen": 1130808, + "step": 3960 + }, + { + "epoch": 7.397388059701493, + "grad_norm": 0.6038731932640076, + "learning_rate": 3.971043280012413e-05, + "loss": 0.6165, + "num_input_tokens_seen": 1132120, + "step": 3965 + }, + { + "epoch": 7.406716417910448, + "grad_norm": 0.3263598084449768, + "learning_rate": 3.967750289635266e-05, + "loss": 0.6201, + "num_input_tokens_seen": 1133592, + "step": 3970 + }, + { + "epoch": 7.416044776119403, + "grad_norm": 0.5111116170883179, + "learning_rate": 3.9644534086521986e-05, + "loss": 0.8541, + "num_input_tokens_seen": 1134968, + "step": 3975 + }, + { + "epoch": 7.425373134328359, + "grad_norm": 0.5162538290023804, + "learning_rate": 3.961152645802343e-05, + "loss": 0.7761, + "num_input_tokens_seen": 1136440, + "step": 3980 + }, + { + "epoch": 7.434701492537314, + "grad_norm": 0.321845680475235, + "learning_rate": 3.9578480098351244e-05, + "loss": 1.1045, + "num_input_tokens_seen": 1137912, + "step": 3985 + }, + { + "epoch": 7.4440298507462686, + "grad_norm": 0.36903682351112366, + "learning_rate": 3.95453950951023e-05, + "loss": 0.9371, + "num_input_tokens_seen": 1139256, + "step": 3990 + }, + { + "epoch": 7.4533582089552235, + "grad_norm": 0.41654449701309204, + "learning_rate": 3.951227153597592e-05, + "loss": 0.7886, + "num_input_tokens_seen": 1140792, + "step": 3995 + }, + { + "epoch": 7.462686567164179, + "grad_norm": 0.4825894236564636, + "learning_rate": 3.947910950877364e-05, + "loss": 0.7981, + "num_input_tokens_seen": 1142200, + "step": 4000 + }, + { + "epoch": 7.472014925373134, + "grad_norm": 0.7218319177627563, + "learning_rate": 3.944590910139895e-05, + "loss": 0.8598, + "num_input_tokens_seen": 1143416, + "step": 4005 + }, + { + "epoch": 7.481343283582089, + "grad_norm": 0.6219455599784851, + "learning_rate": 3.941267040185707e-05, + "loss": 0.9454, + "num_input_tokens_seen": 1144792, + "step": 4010 + }, + { + "epoch": 7.490671641791045, + "grad_norm": 0.4662356972694397, + "learning_rate": 3.937939349825475e-05, + "loss": 0.7495, + "num_input_tokens_seen": 1146232, + "step": 4015 + }, + { + "epoch": 7.5, + "grad_norm": 0.39308786392211914, + "learning_rate": 3.934607847879999e-05, + "loss": 0.7953, + "num_input_tokens_seen": 1147512, + "step": 4020 + }, + { + "epoch": 7.509328358208955, + "grad_norm": 0.37475642561912537, + "learning_rate": 3.93127254318018e-05, + "loss": 0.6465, + "num_input_tokens_seen": 1148952, + "step": 4025 + }, + { + "epoch": 7.518656716417911, + "grad_norm": 0.579609215259552, + "learning_rate": 3.927933444567004e-05, + "loss": 0.8693, + "num_input_tokens_seen": 1150424, + "step": 4030 + }, + { + "epoch": 7.527985074626866, + "grad_norm": 0.5593960881233215, + "learning_rate": 3.924590560891511e-05, + "loss": 0.6432, + "num_input_tokens_seen": 1151992, + "step": 4035 + }, + { + "epoch": 7.537313432835821, + "grad_norm": 0.36352014541625977, + "learning_rate": 3.921243901014773e-05, + "loss": 0.7322, + "num_input_tokens_seen": 1153432, + "step": 4040 + }, + { + "epoch": 7.5466417910447765, + "grad_norm": 0.4350105822086334, + "learning_rate": 3.9178934738078745e-05, + "loss": 0.7535, + "num_input_tokens_seen": 1155000, + "step": 4045 + }, + { + "epoch": 7.5559701492537314, + "grad_norm": 0.6115725636482239, + "learning_rate": 3.914539288151884e-05, + "loss": 0.7847, + "num_input_tokens_seen": 1156376, + "step": 4050 + }, + { + "epoch": 7.565298507462686, + "grad_norm": 0.42759057879447937, + "learning_rate": 3.9111813529378336e-05, + "loss": 0.6157, + "num_input_tokens_seen": 1157752, + "step": 4055 + }, + { + "epoch": 7.574626865671641, + "grad_norm": 0.6640132069587708, + "learning_rate": 3.907819677066694e-05, + "loss": 0.6449, + "num_input_tokens_seen": 1159128, + "step": 4060 + }, + { + "epoch": 7.583955223880597, + "grad_norm": 0.4634123146533966, + "learning_rate": 3.904454269449351e-05, + "loss": 0.8198, + "num_input_tokens_seen": 1160440, + "step": 4065 + }, + { + "epoch": 7.593283582089552, + "grad_norm": 0.30753156542778015, + "learning_rate": 3.901085139006585e-05, + "loss": 0.6924, + "num_input_tokens_seen": 1161912, + "step": 4070 + }, + { + "epoch": 7.602611940298507, + "grad_norm": 0.4505603611469269, + "learning_rate": 3.8977122946690395e-05, + "loss": 0.7345, + "num_input_tokens_seen": 1163288, + "step": 4075 + }, + { + "epoch": 7.611940298507463, + "grad_norm": 0.4296673536300659, + "learning_rate": 3.894335745377208e-05, + "loss": 0.8663, + "num_input_tokens_seen": 1164600, + "step": 4080 + }, + { + "epoch": 7.621268656716418, + "grad_norm": 0.6355675458908081, + "learning_rate": 3.8909555000814e-05, + "loss": 0.7642, + "num_input_tokens_seen": 1165784, + "step": 4085 + }, + { + "epoch": 7.630597014925373, + "grad_norm": 0.5366876125335693, + "learning_rate": 3.8875715677417255e-05, + "loss": 0.875, + "num_input_tokens_seen": 1167160, + "step": 4090 + }, + { + "epoch": 7.639925373134329, + "grad_norm": 0.4233913719654083, + "learning_rate": 3.884183957328067e-05, + "loss": 0.6927, + "num_input_tokens_seen": 1168440, + "step": 4095 + }, + { + "epoch": 7.649253731343284, + "grad_norm": 0.3721265494823456, + "learning_rate": 3.880792677820056e-05, + "loss": 0.8127, + "num_input_tokens_seen": 1169816, + "step": 4100 + }, + { + "epoch": 7.6585820895522385, + "grad_norm": 0.3702072501182556, + "learning_rate": 3.87739773820705e-05, + "loss": 0.5751, + "num_input_tokens_seen": 1171320, + "step": 4105 + }, + { + "epoch": 7.667910447761194, + "grad_norm": 0.3970547914505005, + "learning_rate": 3.873999147488108e-05, + "loss": 0.7372, + "num_input_tokens_seen": 1172632, + "step": 4110 + }, + { + "epoch": 7.677238805970149, + "grad_norm": 0.4552821218967438, + "learning_rate": 3.8705969146719686e-05, + "loss": 0.6015, + "num_input_tokens_seen": 1174136, + "step": 4115 + }, + { + "epoch": 7.686567164179104, + "grad_norm": 0.2848256826400757, + "learning_rate": 3.8671910487770223e-05, + "loss": 0.5724, + "num_input_tokens_seen": 1175608, + "step": 4120 + }, + { + "epoch": 7.69589552238806, + "grad_norm": 0.6300233006477356, + "learning_rate": 3.863781558831292e-05, + "loss": 0.727, + "num_input_tokens_seen": 1176920, + "step": 4125 + }, + { + "epoch": 7.705223880597015, + "grad_norm": 0.4921245872974396, + "learning_rate": 3.8603684538724055e-05, + "loss": 0.8085, + "num_input_tokens_seen": 1178200, + "step": 4130 + }, + { + "epoch": 7.71455223880597, + "grad_norm": 0.531871497631073, + "learning_rate": 3.856951742947574e-05, + "loss": 0.8631, + "num_input_tokens_seen": 1179576, + "step": 4135 + }, + { + "epoch": 7.723880597014926, + "grad_norm": 0.3901126980781555, + "learning_rate": 3.8535314351135674e-05, + "loss": 0.7274, + "num_input_tokens_seen": 1180888, + "step": 4140 + }, + { + "epoch": 7.733208955223881, + "grad_norm": 0.6321995258331299, + "learning_rate": 3.850107539436689e-05, + "loss": 0.9778, + "num_input_tokens_seen": 1182456, + "step": 4145 + }, + { + "epoch": 7.742537313432836, + "grad_norm": 0.4975540041923523, + "learning_rate": 3.8466800649927536e-05, + "loss": 0.6739, + "num_input_tokens_seen": 1183928, + "step": 4150 + }, + { + "epoch": 7.7518656716417915, + "grad_norm": 0.5928874015808105, + "learning_rate": 3.84324902086706e-05, + "loss": 0.514, + "num_input_tokens_seen": 1185336, + "step": 4155 + }, + { + "epoch": 7.7611940298507465, + "grad_norm": 0.4211712181568146, + "learning_rate": 3.839814416154371e-05, + "loss": 0.7331, + "num_input_tokens_seen": 1186648, + "step": 4160 + }, + { + "epoch": 7.770522388059701, + "grad_norm": 0.462962806224823, + "learning_rate": 3.83637625995889e-05, + "loss": 0.6055, + "num_input_tokens_seen": 1188056, + "step": 4165 + }, + { + "epoch": 7.779850746268656, + "grad_norm": 0.6073195338249207, + "learning_rate": 3.832934561394229e-05, + "loss": 0.7481, + "num_input_tokens_seen": 1189304, + "step": 4170 + }, + { + "epoch": 7.789179104477612, + "grad_norm": 0.45824307203292847, + "learning_rate": 3.829489329583394e-05, + "loss": 0.8418, + "num_input_tokens_seen": 1191000, + "step": 4175 + }, + { + "epoch": 7.798507462686567, + "grad_norm": 0.5778367519378662, + "learning_rate": 3.8260405736587546e-05, + "loss": 0.6505, + "num_input_tokens_seen": 1192536, + "step": 4180 + }, + { + "epoch": 7.807835820895522, + "grad_norm": 0.5267208218574524, + "learning_rate": 3.822588302762024e-05, + "loss": 0.5529, + "num_input_tokens_seen": 1194104, + "step": 4185 + }, + { + "epoch": 7.817164179104478, + "grad_norm": 0.43253588676452637, + "learning_rate": 3.81913252604423e-05, + "loss": 0.8361, + "num_input_tokens_seen": 1195416, + "step": 4190 + }, + { + "epoch": 7.826492537313433, + "grad_norm": 0.41406071186065674, + "learning_rate": 3.815673252665696e-05, + "loss": 1.1035, + "num_input_tokens_seen": 1196760, + "step": 4195 + }, + { + "epoch": 7.835820895522388, + "grad_norm": 0.5005417466163635, + "learning_rate": 3.812210491796011e-05, + "loss": 0.6823, + "num_input_tokens_seen": 1198200, + "step": 4200 + }, + { + "epoch": 7.845149253731344, + "grad_norm": 0.35260072350502014, + "learning_rate": 3.808744252614012e-05, + "loss": 0.6747, + "num_input_tokens_seen": 1199832, + "step": 4205 + }, + { + "epoch": 7.854477611940299, + "grad_norm": 0.43316134810447693, + "learning_rate": 3.805274544307752e-05, + "loss": 0.7571, + "num_input_tokens_seen": 1201432, + "step": 4210 + }, + { + "epoch": 7.8638059701492535, + "grad_norm": 0.4938918650150299, + "learning_rate": 3.8018013760744844e-05, + "loss": 1.0839, + "num_input_tokens_seen": 1202712, + "step": 4215 + }, + { + "epoch": 7.8731343283582085, + "grad_norm": 0.37277331948280334, + "learning_rate": 3.798324757120629e-05, + "loss": 0.6848, + "num_input_tokens_seen": 1204312, + "step": 4220 + }, + { + "epoch": 7.882462686567164, + "grad_norm": 0.7040799260139465, + "learning_rate": 3.794844696661757e-05, + "loss": 1.2672, + "num_input_tokens_seen": 1205656, + "step": 4225 + }, + { + "epoch": 7.891791044776119, + "grad_norm": 0.42961016297340393, + "learning_rate": 3.7913612039225596e-05, + "loss": 0.6969, + "num_input_tokens_seen": 1207064, + "step": 4230 + }, + { + "epoch": 7.901119402985074, + "grad_norm": 0.4573317766189575, + "learning_rate": 3.787874288136824e-05, + "loss": 0.6731, + "num_input_tokens_seen": 1208632, + "step": 4235 + }, + { + "epoch": 7.91044776119403, + "grad_norm": 0.5546610355377197, + "learning_rate": 3.7843839585474186e-05, + "loss": 0.7797, + "num_input_tokens_seen": 1209976, + "step": 4240 + }, + { + "epoch": 7.919776119402985, + "grad_norm": 0.3828423023223877, + "learning_rate": 3.78089022440625e-05, + "loss": 0.5511, + "num_input_tokens_seen": 1211384, + "step": 4245 + }, + { + "epoch": 7.92910447761194, + "grad_norm": 0.4689513146877289, + "learning_rate": 3.777393094974259e-05, + "loss": 0.7444, + "num_input_tokens_seen": 1212760, + "step": 4250 + }, + { + "epoch": 7.938432835820896, + "grad_norm": 0.3715716302394867, + "learning_rate": 3.773892579521381e-05, + "loss": 0.5952, + "num_input_tokens_seen": 1214456, + "step": 4255 + }, + { + "epoch": 7.947761194029851, + "grad_norm": 0.4805932939052582, + "learning_rate": 3.7703886873265285e-05, + "loss": 0.9542, + "num_input_tokens_seen": 1215832, + "step": 4260 + }, + { + "epoch": 7.957089552238806, + "grad_norm": 0.47423240542411804, + "learning_rate": 3.766881427677563e-05, + "loss": 0.5303, + "num_input_tokens_seen": 1217240, + "step": 4265 + }, + { + "epoch": 7.9664179104477615, + "grad_norm": 0.6770190596580505, + "learning_rate": 3.7633708098712766e-05, + "loss": 0.7661, + "num_input_tokens_seen": 1218584, + "step": 4270 + }, + { + "epoch": 7.975746268656716, + "grad_norm": 0.36155635118484497, + "learning_rate": 3.7598568432133586e-05, + "loss": 0.7341, + "num_input_tokens_seen": 1219960, + "step": 4275 + }, + { + "epoch": 7.985074626865671, + "grad_norm": 0.5968673825263977, + "learning_rate": 3.7563395370183755e-05, + "loss": 0.8537, + "num_input_tokens_seen": 1221272, + "step": 4280 + }, + { + "epoch": 7.994402985074627, + "grad_norm": 0.3957972228527069, + "learning_rate": 3.75281890060975e-05, + "loss": 0.5654, + "num_input_tokens_seen": 1222808, + "step": 4285 + }, + { + "epoch": 8.0, + "eval_loss": 0.7469407320022583, + "eval_runtime": 2.4567, + "eval_samples_per_second": 96.879, + "eval_steps_per_second": 24.423, + "num_input_tokens_seen": 1223384, + "step": 4288 + }, + { + "epoch": 8.003731343283581, + "grad_norm": 0.44372233748435974, + "learning_rate": 3.749294943319728e-05, + "loss": 0.6589, + "num_input_tokens_seen": 1223992, + "step": 4290 + }, + { + "epoch": 8.013059701492537, + "grad_norm": 0.6587451696395874, + "learning_rate": 3.7457676744893594e-05, + "loss": 1.0123, + "num_input_tokens_seen": 1225464, + "step": 4295 + }, + { + "epoch": 8.022388059701493, + "grad_norm": 0.44818198680877686, + "learning_rate": 3.7422371034684735e-05, + "loss": 0.7368, + "num_input_tokens_seen": 1227032, + "step": 4300 + }, + { + "epoch": 8.031716417910447, + "grad_norm": 0.5407925248146057, + "learning_rate": 3.73870323961565e-05, + "loss": 0.7624, + "num_input_tokens_seen": 1228376, + "step": 4305 + }, + { + "epoch": 8.041044776119403, + "grad_norm": 0.5420264601707458, + "learning_rate": 3.735166092298199e-05, + "loss": 0.7623, + "num_input_tokens_seen": 1229688, + "step": 4310 + }, + { + "epoch": 8.050373134328359, + "grad_norm": 0.4349730908870697, + "learning_rate": 3.731625670892135e-05, + "loss": 0.8406, + "num_input_tokens_seen": 1231128, + "step": 4315 + }, + { + "epoch": 8.059701492537313, + "grad_norm": 0.4247768819332123, + "learning_rate": 3.7280819847821504e-05, + "loss": 0.5238, + "num_input_tokens_seen": 1232632, + "step": 4320 + }, + { + "epoch": 8.069029850746269, + "grad_norm": 0.3454703092575073, + "learning_rate": 3.724535043361589e-05, + "loss": 0.9563, + "num_input_tokens_seen": 1233976, + "step": 4325 + }, + { + "epoch": 8.078358208955224, + "grad_norm": 0.3886520266532898, + "learning_rate": 3.720984856032428e-05, + "loss": 0.6612, + "num_input_tokens_seen": 1235608, + "step": 4330 + }, + { + "epoch": 8.087686567164178, + "grad_norm": 0.6764652132987976, + "learning_rate": 3.717431432205244e-05, + "loss": 0.7259, + "num_input_tokens_seen": 1237016, + "step": 4335 + }, + { + "epoch": 8.097014925373134, + "grad_norm": 0.6109235286712646, + "learning_rate": 3.713874781299196e-05, + "loss": 0.7199, + "num_input_tokens_seen": 1238360, + "step": 4340 + }, + { + "epoch": 8.10634328358209, + "grad_norm": 0.7754396796226501, + "learning_rate": 3.710314912741997e-05, + "loss": 0.5769, + "num_input_tokens_seen": 1239768, + "step": 4345 + }, + { + "epoch": 8.115671641791044, + "grad_norm": 0.3627621531486511, + "learning_rate": 3.7067518359698856e-05, + "loss": 0.4943, + "num_input_tokens_seen": 1241144, + "step": 4350 + }, + { + "epoch": 8.125, + "grad_norm": 0.6411849856376648, + "learning_rate": 3.70318556042761e-05, + "loss": 0.6181, + "num_input_tokens_seen": 1242488, + "step": 4355 + }, + { + "epoch": 8.134328358208956, + "grad_norm": 0.3977873623371124, + "learning_rate": 3.6996160955683924e-05, + "loss": 0.6077, + "num_input_tokens_seen": 1244024, + "step": 4360 + }, + { + "epoch": 8.14365671641791, + "grad_norm": 0.3800828754901886, + "learning_rate": 3.696043450853912e-05, + "loss": 0.9402, + "num_input_tokens_seen": 1245368, + "step": 4365 + }, + { + "epoch": 8.152985074626866, + "grad_norm": 0.6364455819129944, + "learning_rate": 3.692467635754276e-05, + "loss": 0.8738, + "num_input_tokens_seen": 1246712, + "step": 4370 + }, + { + "epoch": 8.162313432835822, + "grad_norm": 0.5204750895500183, + "learning_rate": 3.688888659747995e-05, + "loss": 0.7706, + "num_input_tokens_seen": 1248216, + "step": 4375 + }, + { + "epoch": 8.171641791044776, + "grad_norm": 0.5063745975494385, + "learning_rate": 3.6853065323219606e-05, + "loss": 0.6705, + "num_input_tokens_seen": 1249688, + "step": 4380 + }, + { + "epoch": 8.180970149253731, + "grad_norm": 0.3386964201927185, + "learning_rate": 3.681721262971413e-05, + "loss": 0.7686, + "num_input_tokens_seen": 1251096, + "step": 4385 + }, + { + "epoch": 8.190298507462687, + "grad_norm": 0.45624569058418274, + "learning_rate": 3.678132861199927e-05, + "loss": 0.5598, + "num_input_tokens_seen": 1252664, + "step": 4390 + }, + { + "epoch": 8.199626865671641, + "grad_norm": 0.4004260301589966, + "learning_rate": 3.674541336519376e-05, + "loss": 0.7782, + "num_input_tokens_seen": 1254136, + "step": 4395 + }, + { + "epoch": 8.208955223880597, + "grad_norm": 0.5060660243034363, + "learning_rate": 3.670946698449912e-05, + "loss": 0.6066, + "num_input_tokens_seen": 1255608, + "step": 4400 + }, + { + "epoch": 8.218283582089553, + "grad_norm": 0.4452071487903595, + "learning_rate": 3.667348956519943e-05, + "loss": 0.7137, + "num_input_tokens_seen": 1257080, + "step": 4405 + }, + { + "epoch": 8.227611940298507, + "grad_norm": 0.4885548949241638, + "learning_rate": 3.6637481202661006e-05, + "loss": 0.8008, + "num_input_tokens_seen": 1258520, + "step": 4410 + }, + { + "epoch": 8.236940298507463, + "grad_norm": 0.5083829164505005, + "learning_rate": 3.660144199233221e-05, + "loss": 0.6808, + "num_input_tokens_seen": 1259896, + "step": 4415 + }, + { + "epoch": 8.246268656716419, + "grad_norm": 0.28621456027030945, + "learning_rate": 3.656537202974315e-05, + "loss": 0.8842, + "num_input_tokens_seen": 1261528, + "step": 4420 + }, + { + "epoch": 8.255597014925373, + "grad_norm": 0.6495375633239746, + "learning_rate": 3.652927141050548e-05, + "loss": 0.7732, + "num_input_tokens_seen": 1262808, + "step": 4425 + }, + { + "epoch": 8.264925373134329, + "grad_norm": 0.5401778817176819, + "learning_rate": 3.649314023031208e-05, + "loss": 0.7968, + "num_input_tokens_seen": 1264152, + "step": 4430 + }, + { + "epoch": 8.274253731343283, + "grad_norm": 0.31370964646339417, + "learning_rate": 3.64569785849369e-05, + "loss": 0.7312, + "num_input_tokens_seen": 1265656, + "step": 4435 + }, + { + "epoch": 8.283582089552239, + "grad_norm": 0.5473148226737976, + "learning_rate": 3.642078657023456e-05, + "loss": 0.9218, + "num_input_tokens_seen": 1266968, + "step": 4440 + }, + { + "epoch": 8.292910447761194, + "grad_norm": 0.5221791863441467, + "learning_rate": 3.638456428214024e-05, + "loss": 0.9793, + "num_input_tokens_seen": 1268536, + "step": 4445 + }, + { + "epoch": 8.302238805970148, + "grad_norm": 0.5974968671798706, + "learning_rate": 3.6348311816669366e-05, + "loss": 0.9037, + "num_input_tokens_seen": 1269880, + "step": 4450 + }, + { + "epoch": 8.311567164179104, + "grad_norm": 0.38046640157699585, + "learning_rate": 3.6312029269917335e-05, + "loss": 0.7008, + "num_input_tokens_seen": 1271352, + "step": 4455 + }, + { + "epoch": 8.32089552238806, + "grad_norm": 0.4532108008861542, + "learning_rate": 3.627571673805927e-05, + "loss": 0.8487, + "num_input_tokens_seen": 1272824, + "step": 4460 + }, + { + "epoch": 8.330223880597014, + "grad_norm": 0.5598095655441284, + "learning_rate": 3.623937431734982e-05, + "loss": 0.6261, + "num_input_tokens_seen": 1274232, + "step": 4465 + }, + { + "epoch": 8.33955223880597, + "grad_norm": 0.6013551354408264, + "learning_rate": 3.6203002104122824e-05, + "loss": 0.8018, + "num_input_tokens_seen": 1275608, + "step": 4470 + }, + { + "epoch": 8.348880597014926, + "grad_norm": 0.6728977560997009, + "learning_rate": 3.61666001947911e-05, + "loss": 0.6805, + "num_input_tokens_seen": 1276888, + "step": 4475 + }, + { + "epoch": 8.35820895522388, + "grad_norm": 0.335165411233902, + "learning_rate": 3.61301686858462e-05, + "loss": 0.8302, + "num_input_tokens_seen": 1278424, + "step": 4480 + }, + { + "epoch": 8.367537313432836, + "grad_norm": 0.4760945737361908, + "learning_rate": 3.609370767385811e-05, + "loss": 0.8294, + "num_input_tokens_seen": 1279576, + "step": 4485 + }, + { + "epoch": 8.376865671641792, + "grad_norm": 0.4704505503177643, + "learning_rate": 3.6057217255475034e-05, + "loss": 0.6724, + "num_input_tokens_seen": 1280888, + "step": 4490 + }, + { + "epoch": 8.386194029850746, + "grad_norm": 0.4582040309906006, + "learning_rate": 3.6020697527423134e-05, + "loss": 0.6323, + "num_input_tokens_seen": 1282488, + "step": 4495 + }, + { + "epoch": 8.395522388059701, + "grad_norm": 0.6542097926139832, + "learning_rate": 3.598414858650625e-05, + "loss": 0.8923, + "num_input_tokens_seen": 1283928, + "step": 4500 + }, + { + "epoch": 8.404850746268657, + "grad_norm": 1.0028257369995117, + "learning_rate": 3.594757052960566e-05, + "loss": 0.9741, + "num_input_tokens_seen": 1285240, + "step": 4505 + }, + { + "epoch": 8.414179104477611, + "grad_norm": 0.4959922432899475, + "learning_rate": 3.591096345367982e-05, + "loss": 0.6928, + "num_input_tokens_seen": 1286680, + "step": 4510 + }, + { + "epoch": 8.423507462686567, + "grad_norm": 0.5660980939865112, + "learning_rate": 3.587432745576411e-05, + "loss": 0.7021, + "num_input_tokens_seen": 1288216, + "step": 4515 + }, + { + "epoch": 8.432835820895523, + "grad_norm": 0.46799421310424805, + "learning_rate": 3.583766263297058e-05, + "loss": 0.7391, + "num_input_tokens_seen": 1289720, + "step": 4520 + }, + { + "epoch": 8.442164179104477, + "grad_norm": 0.41589194536209106, + "learning_rate": 3.580096908248768e-05, + "loss": 0.5462, + "num_input_tokens_seen": 1291096, + "step": 4525 + }, + { + "epoch": 8.451492537313433, + "grad_norm": 0.49680763483047485, + "learning_rate": 3.576424690158e-05, + "loss": 0.6605, + "num_input_tokens_seen": 1292408, + "step": 4530 + }, + { + "epoch": 8.460820895522389, + "grad_norm": 0.4566895067691803, + "learning_rate": 3.572749618758804e-05, + "loss": 0.771, + "num_input_tokens_seen": 1293816, + "step": 4535 + }, + { + "epoch": 8.470149253731343, + "grad_norm": 0.5746862888336182, + "learning_rate": 3.5690717037927926e-05, + "loss": 0.9172, + "num_input_tokens_seen": 1295128, + "step": 4540 + }, + { + "epoch": 8.479477611940299, + "grad_norm": 0.6830249428749084, + "learning_rate": 3.565390955009113e-05, + "loss": 0.8356, + "num_input_tokens_seen": 1296408, + "step": 4545 + }, + { + "epoch": 8.488805970149254, + "grad_norm": 0.48283347487449646, + "learning_rate": 3.561707382164432e-05, + "loss": 0.7491, + "num_input_tokens_seen": 1297720, + "step": 4550 + }, + { + "epoch": 8.498134328358208, + "grad_norm": 0.32665809988975525, + "learning_rate": 3.5580209950228936e-05, + "loss": 0.6092, + "num_input_tokens_seen": 1299160, + "step": 4555 + }, + { + "epoch": 8.507462686567164, + "grad_norm": 0.45016342401504517, + "learning_rate": 3.554331803356107e-05, + "loss": 0.5984, + "num_input_tokens_seen": 1300792, + "step": 4560 + }, + { + "epoch": 8.51679104477612, + "grad_norm": 0.5251902937889099, + "learning_rate": 3.550639816943111e-05, + "loss": 0.663, + "num_input_tokens_seen": 1302360, + "step": 4565 + }, + { + "epoch": 8.526119402985074, + "grad_norm": 0.440004825592041, + "learning_rate": 3.546945045570358e-05, + "loss": 0.7435, + "num_input_tokens_seen": 1303800, + "step": 4570 + }, + { + "epoch": 8.53544776119403, + "grad_norm": 0.3921791911125183, + "learning_rate": 3.543247499031679e-05, + "loss": 0.6935, + "num_input_tokens_seen": 1305176, + "step": 4575 + }, + { + "epoch": 8.544776119402986, + "grad_norm": 0.48383262753486633, + "learning_rate": 3.5395471871282604e-05, + "loss": 0.7054, + "num_input_tokens_seen": 1306648, + "step": 4580 + }, + { + "epoch": 8.55410447761194, + "grad_norm": 0.3118312358856201, + "learning_rate": 3.535844119668622e-05, + "loss": 0.7371, + "num_input_tokens_seen": 1308024, + "step": 4585 + }, + { + "epoch": 8.563432835820896, + "grad_norm": 0.31263700127601624, + "learning_rate": 3.532138306468586e-05, + "loss": 0.6129, + "num_input_tokens_seen": 1309624, + "step": 4590 + }, + { + "epoch": 8.572761194029852, + "grad_norm": 0.4481871724128723, + "learning_rate": 3.528429757351253e-05, + "loss": 0.6076, + "num_input_tokens_seen": 1310968, + "step": 4595 + }, + { + "epoch": 8.582089552238806, + "grad_norm": 0.6296461820602417, + "learning_rate": 3.524718482146975e-05, + "loss": 0.746, + "num_input_tokens_seen": 1312504, + "step": 4600 + }, + { + "epoch": 8.591417910447761, + "grad_norm": 0.31985539197921753, + "learning_rate": 3.521004490693331e-05, + "loss": 0.6305, + "num_input_tokens_seen": 1313720, + "step": 4605 + }, + { + "epoch": 8.600746268656717, + "grad_norm": 0.5834276676177979, + "learning_rate": 3.5172877928351024e-05, + "loss": 0.7767, + "num_input_tokens_seen": 1315032, + "step": 4610 + }, + { + "epoch": 8.610074626865671, + "grad_norm": 0.388621985912323, + "learning_rate": 3.513568398424239e-05, + "loss": 0.8411, + "num_input_tokens_seen": 1316632, + "step": 4615 + }, + { + "epoch": 8.619402985074627, + "grad_norm": 0.36017507314682007, + "learning_rate": 3.509846317319841e-05, + "loss": 0.8054, + "num_input_tokens_seen": 1317944, + "step": 4620 + }, + { + "epoch": 8.628731343283581, + "grad_norm": 0.5079282522201538, + "learning_rate": 3.5061215593881345e-05, + "loss": 0.6924, + "num_input_tokens_seen": 1319352, + "step": 4625 + }, + { + "epoch": 8.638059701492537, + "grad_norm": 0.37411239743232727, + "learning_rate": 3.502394134502435e-05, + "loss": 0.6771, + "num_input_tokens_seen": 1320824, + "step": 4630 + }, + { + "epoch": 8.647388059701493, + "grad_norm": 0.38043907284736633, + "learning_rate": 3.4986640525431286e-05, + "loss": 0.7474, + "num_input_tokens_seen": 1322328, + "step": 4635 + }, + { + "epoch": 8.656716417910447, + "grad_norm": 0.584481418132782, + "learning_rate": 3.494931323397649e-05, + "loss": 0.7348, + "num_input_tokens_seen": 1323768, + "step": 4640 + }, + { + "epoch": 8.666044776119403, + "grad_norm": 0.5442440509796143, + "learning_rate": 3.491195956960441e-05, + "loss": 0.7307, + "num_input_tokens_seen": 1325176, + "step": 4645 + }, + { + "epoch": 8.675373134328359, + "grad_norm": 0.6432859301567078, + "learning_rate": 3.4874579631329443e-05, + "loss": 0.869, + "num_input_tokens_seen": 1326744, + "step": 4650 + }, + { + "epoch": 8.684701492537313, + "grad_norm": 0.5043646693229675, + "learning_rate": 3.483717351823561e-05, + "loss": 0.6278, + "num_input_tokens_seen": 1328120, + "step": 4655 + }, + { + "epoch": 8.694029850746269, + "grad_norm": 0.4595416188240051, + "learning_rate": 3.479974132947632e-05, + "loss": 0.7405, + "num_input_tokens_seen": 1329624, + "step": 4660 + }, + { + "epoch": 8.703358208955224, + "grad_norm": 0.4896150827407837, + "learning_rate": 3.47622831642741e-05, + "loss": 0.708, + "num_input_tokens_seen": 1331096, + "step": 4665 + }, + { + "epoch": 8.712686567164178, + "grad_norm": 0.5238153338432312, + "learning_rate": 3.472479912192034e-05, + "loss": 0.8574, + "num_input_tokens_seen": 1332280, + "step": 4670 + }, + { + "epoch": 8.722014925373134, + "grad_norm": 0.6597668528556824, + "learning_rate": 3.468728930177501e-05, + "loss": 0.7845, + "num_input_tokens_seen": 1334072, + "step": 4675 + }, + { + "epoch": 8.73134328358209, + "grad_norm": 0.43389055132865906, + "learning_rate": 3.464975380326643e-05, + "loss": 0.6211, + "num_input_tokens_seen": 1335352, + "step": 4680 + }, + { + "epoch": 8.740671641791044, + "grad_norm": 0.5396770238876343, + "learning_rate": 3.461219272589097e-05, + "loss": 0.8204, + "num_input_tokens_seen": 1336760, + "step": 4685 + }, + { + "epoch": 8.75, + "grad_norm": 0.4146003723144531, + "learning_rate": 3.45746061692128e-05, + "loss": 0.6097, + "num_input_tokens_seen": 1338328, + "step": 4690 + }, + { + "epoch": 8.759328358208956, + "grad_norm": 0.6485005021095276, + "learning_rate": 3.4536994232863637e-05, + "loss": 0.5539, + "num_input_tokens_seen": 1339800, + "step": 4695 + }, + { + "epoch": 8.76865671641791, + "grad_norm": 0.27225005626678467, + "learning_rate": 3.4499357016542485e-05, + "loss": 0.6296, + "num_input_tokens_seen": 1341304, + "step": 4700 + }, + { + "epoch": 8.777985074626866, + "grad_norm": 0.42192327976226807, + "learning_rate": 3.446169462001534e-05, + "loss": 0.7004, + "num_input_tokens_seen": 1342712, + "step": 4705 + }, + { + "epoch": 8.787313432835822, + "grad_norm": 0.39788004755973816, + "learning_rate": 3.4424007143114944e-05, + "loss": 0.7386, + "num_input_tokens_seen": 1344216, + "step": 4710 + }, + { + "epoch": 8.796641791044776, + "grad_norm": 0.6502968668937683, + "learning_rate": 3.438629468574052e-05, + "loss": 0.8224, + "num_input_tokens_seen": 1345624, + "step": 4715 + }, + { + "epoch": 8.805970149253731, + "grad_norm": 0.4159415364265442, + "learning_rate": 3.4348557347857526e-05, + "loss": 1.0789, + "num_input_tokens_seen": 1347032, + "step": 4720 + }, + { + "epoch": 8.815298507462687, + "grad_norm": 0.5499621629714966, + "learning_rate": 3.431079522949734e-05, + "loss": 0.9519, + "num_input_tokens_seen": 1348504, + "step": 4725 + }, + { + "epoch": 8.824626865671641, + "grad_norm": 0.4382382929325104, + "learning_rate": 3.427300843075706e-05, + "loss": 0.7057, + "num_input_tokens_seen": 1349944, + "step": 4730 + }, + { + "epoch": 8.833955223880597, + "grad_norm": 0.45231369137763977, + "learning_rate": 3.423519705179918e-05, + "loss": 0.8534, + "num_input_tokens_seen": 1351192, + "step": 4735 + }, + { + "epoch": 8.843283582089553, + "grad_norm": 0.5075811147689819, + "learning_rate": 3.419736119285136e-05, + "loss": 0.5939, + "num_input_tokens_seen": 1352504, + "step": 4740 + }, + { + "epoch": 8.852611940298507, + "grad_norm": 0.4507134258747101, + "learning_rate": 3.415950095420616e-05, + "loss": 0.7018, + "num_input_tokens_seen": 1353912, + "step": 4745 + }, + { + "epoch": 8.861940298507463, + "grad_norm": 0.4786553680896759, + "learning_rate": 3.4121616436220734e-05, + "loss": 0.5715, + "num_input_tokens_seen": 1355416, + "step": 4750 + }, + { + "epoch": 8.871268656716419, + "grad_norm": 0.6489531397819519, + "learning_rate": 3.4083707739316614e-05, + "loss": 0.8025, + "num_input_tokens_seen": 1356984, + "step": 4755 + }, + { + "epoch": 8.880597014925373, + "grad_norm": 0.5135108232498169, + "learning_rate": 3.404577496397944e-05, + "loss": 0.8771, + "num_input_tokens_seen": 1358296, + "step": 4760 + }, + { + "epoch": 8.889925373134329, + "grad_norm": 0.2643517851829529, + "learning_rate": 3.400781821075865e-05, + "loss": 0.6224, + "num_input_tokens_seen": 1359864, + "step": 4765 + }, + { + "epoch": 8.899253731343283, + "grad_norm": 0.5333414673805237, + "learning_rate": 3.396983758026724e-05, + "loss": 0.4954, + "num_input_tokens_seen": 1361272, + "step": 4770 + }, + { + "epoch": 8.908582089552239, + "grad_norm": 0.7326984405517578, + "learning_rate": 3.393183317318151e-05, + "loss": 0.6506, + "num_input_tokens_seen": 1362904, + "step": 4775 + }, + { + "epoch": 8.917910447761194, + "grad_norm": 0.6346047520637512, + "learning_rate": 3.389380509024081e-05, + "loss": 0.7818, + "num_input_tokens_seen": 1364312, + "step": 4780 + }, + { + "epoch": 8.927238805970148, + "grad_norm": 0.49178656935691833, + "learning_rate": 3.385575343224718e-05, + "loss": 0.7468, + "num_input_tokens_seen": 1365816, + "step": 4785 + }, + { + "epoch": 8.936567164179104, + "grad_norm": 0.5629570484161377, + "learning_rate": 3.381767830006522e-05, + "loss": 0.8734, + "num_input_tokens_seen": 1367224, + "step": 4790 + }, + { + "epoch": 8.94589552238806, + "grad_norm": 0.4279366433620453, + "learning_rate": 3.377957979462172e-05, + "loss": 0.7706, + "num_input_tokens_seen": 1368472, + "step": 4795 + }, + { + "epoch": 8.955223880597014, + "grad_norm": 0.5362611413002014, + "learning_rate": 3.3741458016905436e-05, + "loss": 0.6972, + "num_input_tokens_seen": 1369784, + "step": 4800 + }, + { + "epoch": 8.96455223880597, + "grad_norm": 0.4480445384979248, + "learning_rate": 3.37033130679668e-05, + "loss": 0.6869, + "num_input_tokens_seen": 1371032, + "step": 4805 + }, + { + "epoch": 8.973880597014926, + "grad_norm": 0.37841928005218506, + "learning_rate": 3.366514504891769e-05, + "loss": 0.5896, + "num_input_tokens_seen": 1372440, + "step": 4810 + }, + { + "epoch": 8.98320895522388, + "grad_norm": 0.5413593649864197, + "learning_rate": 3.36269540609311e-05, + "loss": 0.6737, + "num_input_tokens_seen": 1373784, + "step": 4815 + }, + { + "epoch": 8.992537313432836, + "grad_norm": 0.35998380184173584, + "learning_rate": 3.358874020524094e-05, + "loss": 0.6064, + "num_input_tokens_seen": 1375352, + "step": 4820 + }, + { + "epoch": 9.0, + "eval_loss": 0.7342267632484436, + "eval_runtime": 2.4619, + "eval_samples_per_second": 96.672, + "eval_steps_per_second": 24.371, + "num_input_tokens_seen": 1376240, + "step": 4824 + }, + { + "epoch": 9.001865671641792, + "grad_norm": 0.5112494826316833, + "learning_rate": 3.355050358314172e-05, + "loss": 0.5968, + "num_input_tokens_seen": 1376560, + "step": 4825 + }, + { + "epoch": 9.011194029850746, + "grad_norm": 0.3878052532672882, + "learning_rate": 3.3512244295988306e-05, + "loss": 0.7706, + "num_input_tokens_seen": 1378160, + "step": 4830 + }, + { + "epoch": 9.020522388059701, + "grad_norm": 0.4483376145362854, + "learning_rate": 3.3473962445195646e-05, + "loss": 0.9003, + "num_input_tokens_seen": 1379568, + "step": 4835 + }, + { + "epoch": 9.029850746268657, + "grad_norm": 0.4799887239933014, + "learning_rate": 3.343565813223847e-05, + "loss": 0.6477, + "num_input_tokens_seen": 1381168, + "step": 4840 + }, + { + "epoch": 9.039179104477611, + "grad_norm": 0.5365697741508484, + "learning_rate": 3.339733145865109e-05, + "loss": 0.8033, + "num_input_tokens_seen": 1382736, + "step": 4845 + }, + { + "epoch": 9.048507462686567, + "grad_norm": 0.5388579368591309, + "learning_rate": 3.3358982526027065e-05, + "loss": 0.6234, + "num_input_tokens_seen": 1384112, + "step": 4850 + }, + { + "epoch": 9.057835820895523, + "grad_norm": 0.4152458608150482, + "learning_rate": 3.3320611436018955e-05, + "loss": 0.7209, + "num_input_tokens_seen": 1385520, + "step": 4855 + }, + { + "epoch": 9.067164179104477, + "grad_norm": 0.5960867404937744, + "learning_rate": 3.328221829033807e-05, + "loss": 0.8005, + "num_input_tokens_seen": 1387120, + "step": 4860 + }, + { + "epoch": 9.076492537313433, + "grad_norm": 0.3867981433868408, + "learning_rate": 3.324380319075416e-05, + "loss": 0.8128, + "num_input_tokens_seen": 1388688, + "step": 4865 + }, + { + "epoch": 9.085820895522389, + "grad_norm": 0.44487157464027405, + "learning_rate": 3.32053662390952e-05, + "loss": 0.5836, + "num_input_tokens_seen": 1390128, + "step": 4870 + }, + { + "epoch": 9.095149253731343, + "grad_norm": 0.4415973424911499, + "learning_rate": 3.316690753724706e-05, + "loss": 0.8986, + "num_input_tokens_seen": 1391632, + "step": 4875 + }, + { + "epoch": 9.104477611940299, + "grad_norm": 0.5161446332931519, + "learning_rate": 3.312842718715328e-05, + "loss": 0.6248, + "num_input_tokens_seen": 1393072, + "step": 4880 + }, + { + "epoch": 9.113805970149254, + "grad_norm": 0.4153090715408325, + "learning_rate": 3.308992529081477e-05, + "loss": 0.7734, + "num_input_tokens_seen": 1394640, + "step": 4885 + }, + { + "epoch": 9.123134328358208, + "grad_norm": 0.4326528310775757, + "learning_rate": 3.3051401950289566e-05, + "loss": 0.6166, + "num_input_tokens_seen": 1396080, + "step": 4890 + }, + { + "epoch": 9.132462686567164, + "grad_norm": 0.3988601565361023, + "learning_rate": 3.301285726769255e-05, + "loss": 0.7203, + "num_input_tokens_seen": 1397584, + "step": 4895 + }, + { + "epoch": 9.14179104477612, + "grad_norm": 0.4928361475467682, + "learning_rate": 3.297429134519516e-05, + "loss": 0.9315, + "num_input_tokens_seen": 1398832, + "step": 4900 + }, + { + "epoch": 9.151119402985074, + "grad_norm": 0.3129136562347412, + "learning_rate": 3.293570428502515e-05, + "loss": 0.7694, + "num_input_tokens_seen": 1400272, + "step": 4905 + }, + { + "epoch": 9.16044776119403, + "grad_norm": 0.644656777381897, + "learning_rate": 3.2897096189466284e-05, + "loss": 0.7096, + "num_input_tokens_seen": 1401680, + "step": 4910 + }, + { + "epoch": 9.169776119402986, + "grad_norm": 0.5794469714164734, + "learning_rate": 3.2858467160858116e-05, + "loss": 0.9232, + "num_input_tokens_seen": 1402992, + "step": 4915 + }, + { + "epoch": 9.17910447761194, + "grad_norm": 0.6690397262573242, + "learning_rate": 3.281981730159567e-05, + "loss": 0.5672, + "num_input_tokens_seen": 1404528, + "step": 4920 + }, + { + "epoch": 9.188432835820896, + "grad_norm": 0.23921272158622742, + "learning_rate": 3.278114671412917e-05, + "loss": 0.5596, + "num_input_tokens_seen": 1406288, + "step": 4925 + }, + { + "epoch": 9.197761194029852, + "grad_norm": 0.3707886040210724, + "learning_rate": 3.274245550096382e-05, + "loss": 0.6642, + "num_input_tokens_seen": 1407888, + "step": 4930 + }, + { + "epoch": 9.207089552238806, + "grad_norm": 0.26984134316444397, + "learning_rate": 3.2703743764659475e-05, + "loss": 0.4305, + "num_input_tokens_seen": 1409488, + "step": 4935 + }, + { + "epoch": 9.216417910447761, + "grad_norm": 0.5667305588722229, + "learning_rate": 3.266501160783039e-05, + "loss": 0.8029, + "num_input_tokens_seen": 1410864, + "step": 4940 + }, + { + "epoch": 9.225746268656717, + "grad_norm": 0.48994648456573486, + "learning_rate": 3.262625913314496e-05, + "loss": 0.6399, + "num_input_tokens_seen": 1412592, + "step": 4945 + }, + { + "epoch": 9.235074626865671, + "grad_norm": 0.4791450500488281, + "learning_rate": 3.2587486443325424e-05, + "loss": 0.9019, + "num_input_tokens_seen": 1413968, + "step": 4950 + }, + { + "epoch": 9.244402985074627, + "grad_norm": 0.5281780362129211, + "learning_rate": 3.254869364114763e-05, + "loss": 0.7158, + "num_input_tokens_seen": 1415248, + "step": 4955 + }, + { + "epoch": 9.253731343283581, + "grad_norm": 0.45974698662757874, + "learning_rate": 3.25098808294407e-05, + "loss": 0.7734, + "num_input_tokens_seen": 1416752, + "step": 4960 + }, + { + "epoch": 9.263059701492537, + "grad_norm": 0.38772836327552795, + "learning_rate": 3.2471048111086825e-05, + "loss": 0.6131, + "num_input_tokens_seen": 1418128, + "step": 4965 + }, + { + "epoch": 9.272388059701493, + "grad_norm": 0.6069932579994202, + "learning_rate": 3.2432195589020976e-05, + "loss": 0.5509, + "num_input_tokens_seen": 1419440, + "step": 4970 + }, + { + "epoch": 9.281716417910447, + "grad_norm": 0.6035856604576111, + "learning_rate": 3.2393323366230575e-05, + "loss": 1.0042, + "num_input_tokens_seen": 1420880, + "step": 4975 + }, + { + "epoch": 9.291044776119403, + "grad_norm": 0.4974258840084076, + "learning_rate": 3.2354431545755296e-05, + "loss": 0.587, + "num_input_tokens_seen": 1422320, + "step": 4980 + }, + { + "epoch": 9.300373134328359, + "grad_norm": 0.513087272644043, + "learning_rate": 3.231552023068675e-05, + "loss": 0.7599, + "num_input_tokens_seen": 1423632, + "step": 4985 + }, + { + "epoch": 9.309701492537313, + "grad_norm": 0.46063247323036194, + "learning_rate": 3.227658952416822e-05, + "loss": 0.8168, + "num_input_tokens_seen": 1424944, + "step": 4990 + }, + { + "epoch": 9.319029850746269, + "grad_norm": 0.4131579101085663, + "learning_rate": 3.223763952939442e-05, + "loss": 0.6311, + "num_input_tokens_seen": 1426288, + "step": 4995 + }, + { + "epoch": 9.328358208955224, + "grad_norm": 0.3575219511985779, + "learning_rate": 3.219867034961114e-05, + "loss": 0.7723, + "num_input_tokens_seen": 1427568, + "step": 5000 + }, + { + "epoch": 9.337686567164178, + "grad_norm": 0.3366541862487793, + "learning_rate": 3.215968208811508e-05, + "loss": 0.7269, + "num_input_tokens_seen": 1428880, + "step": 5005 + }, + { + "epoch": 9.347014925373134, + "grad_norm": 0.3262519836425781, + "learning_rate": 3.2120674848253475e-05, + "loss": 0.5873, + "num_input_tokens_seen": 1430576, + "step": 5010 + }, + { + "epoch": 9.35634328358209, + "grad_norm": 0.5173710584640503, + "learning_rate": 3.2081648733423893e-05, + "loss": 0.7544, + "num_input_tokens_seen": 1432016, + "step": 5015 + }, + { + "epoch": 9.365671641791044, + "grad_norm": 0.48029854893684387, + "learning_rate": 3.204260384707393e-05, + "loss": 0.6969, + "num_input_tokens_seen": 1433520, + "step": 5020 + }, + { + "epoch": 9.375, + "grad_norm": 0.47161218523979187, + "learning_rate": 3.200354029270091e-05, + "loss": 0.6427, + "num_input_tokens_seen": 1434896, + "step": 5025 + }, + { + "epoch": 9.384328358208956, + "grad_norm": 0.3155362904071808, + "learning_rate": 3.196445817385171e-05, + "loss": 0.7638, + "num_input_tokens_seen": 1436208, + "step": 5030 + }, + { + "epoch": 9.39365671641791, + "grad_norm": 0.6714115142822266, + "learning_rate": 3.192535759412233e-05, + "loss": 0.6016, + "num_input_tokens_seen": 1437552, + "step": 5035 + }, + { + "epoch": 9.402985074626866, + "grad_norm": 0.5492150783538818, + "learning_rate": 3.188623865715778e-05, + "loss": 0.9675, + "num_input_tokens_seen": 1438832, + "step": 5040 + }, + { + "epoch": 9.412313432835822, + "grad_norm": 0.5668659806251526, + "learning_rate": 3.184710146665169e-05, + "loss": 0.6318, + "num_input_tokens_seen": 1440240, + "step": 5045 + }, + { + "epoch": 9.421641791044776, + "grad_norm": 0.5631800293922424, + "learning_rate": 3.180794612634608e-05, + "loss": 0.5545, + "num_input_tokens_seen": 1441584, + "step": 5050 + }, + { + "epoch": 9.430970149253731, + "grad_norm": 0.8040881156921387, + "learning_rate": 3.176877274003108e-05, + "loss": 0.6086, + "num_input_tokens_seen": 1443120, + "step": 5055 + }, + { + "epoch": 9.440298507462687, + "grad_norm": 0.44266167283058167, + "learning_rate": 3.172958141154466e-05, + "loss": 0.4567, + "num_input_tokens_seen": 1444784, + "step": 5060 + }, + { + "epoch": 9.449626865671641, + "grad_norm": 0.3545379042625427, + "learning_rate": 3.1690372244772356e-05, + "loss": 0.7207, + "num_input_tokens_seen": 1446352, + "step": 5065 + }, + { + "epoch": 9.458955223880597, + "grad_norm": 0.4228437542915344, + "learning_rate": 3.165114534364698e-05, + "loss": 0.7184, + "num_input_tokens_seen": 1447632, + "step": 5070 + }, + { + "epoch": 9.468283582089553, + "grad_norm": 0.8778691291809082, + "learning_rate": 3.161190081214835e-05, + "loss": 0.8601, + "num_input_tokens_seen": 1449040, + "step": 5075 + }, + { + "epoch": 9.477611940298507, + "grad_norm": 0.49555566906929016, + "learning_rate": 3.157263875430302e-05, + "loss": 1.0362, + "num_input_tokens_seen": 1450256, + "step": 5080 + }, + { + "epoch": 9.486940298507463, + "grad_norm": 0.7382689714431763, + "learning_rate": 3.1533359274184e-05, + "loss": 0.897, + "num_input_tokens_seen": 1451600, + "step": 5085 + }, + { + "epoch": 9.496268656716419, + "grad_norm": 0.31658971309661865, + "learning_rate": 3.14940624759105e-05, + "loss": 0.5734, + "num_input_tokens_seen": 1453200, + "step": 5090 + }, + { + "epoch": 9.505597014925373, + "grad_norm": 0.4774719774723053, + "learning_rate": 3.145474846364761e-05, + "loss": 0.6699, + "num_input_tokens_seen": 1454512, + "step": 5095 + }, + { + "epoch": 9.514925373134329, + "grad_norm": 0.5332751870155334, + "learning_rate": 3.1415417341606054e-05, + "loss": 0.7576, + "num_input_tokens_seen": 1455952, + "step": 5100 + }, + { + "epoch": 9.524253731343283, + "grad_norm": 0.40650662779808044, + "learning_rate": 3.1376069214041913e-05, + "loss": 0.8564, + "num_input_tokens_seen": 1457328, + "step": 5105 + }, + { + "epoch": 9.533582089552239, + "grad_norm": 0.5219431519508362, + "learning_rate": 3.1336704185256363e-05, + "loss": 0.6075, + "num_input_tokens_seen": 1458800, + "step": 5110 + }, + { + "epoch": 9.542910447761194, + "grad_norm": 0.4945698082447052, + "learning_rate": 3.129732235959535e-05, + "loss": 0.8173, + "num_input_tokens_seen": 1460144, + "step": 5115 + }, + { + "epoch": 9.552238805970148, + "grad_norm": 0.3504523038864136, + "learning_rate": 3.1257923841449374e-05, + "loss": 0.5503, + "num_input_tokens_seen": 1461456, + "step": 5120 + }, + { + "epoch": 9.561567164179104, + "grad_norm": 0.3395248353481293, + "learning_rate": 3.121850873525315e-05, + "loss": 0.846, + "num_input_tokens_seen": 1462896, + "step": 5125 + }, + { + "epoch": 9.57089552238806, + "grad_norm": 0.35810965299606323, + "learning_rate": 3.1179077145485395e-05, + "loss": 0.8963, + "num_input_tokens_seen": 1464176, + "step": 5130 + }, + { + "epoch": 9.580223880597014, + "grad_norm": 0.5055875778198242, + "learning_rate": 3.1139629176668496e-05, + "loss": 0.83, + "num_input_tokens_seen": 1465552, + "step": 5135 + }, + { + "epoch": 9.58955223880597, + "grad_norm": 0.5442038774490356, + "learning_rate": 3.1100164933368263e-05, + "loss": 1.0391, + "num_input_tokens_seen": 1466896, + "step": 5140 + }, + { + "epoch": 9.598880597014926, + "grad_norm": 0.5341441035270691, + "learning_rate": 3.106068452019365e-05, + "loss": 0.6522, + "num_input_tokens_seen": 1468432, + "step": 5145 + }, + { + "epoch": 9.60820895522388, + "grad_norm": 0.4389152228832245, + "learning_rate": 3.1021188041796476e-05, + "loss": 0.8573, + "num_input_tokens_seen": 1469712, + "step": 5150 + }, + { + "epoch": 9.617537313432836, + "grad_norm": 0.600601315498352, + "learning_rate": 3.098167560287113e-05, + "loss": 0.5085, + "num_input_tokens_seen": 1471152, + "step": 5155 + }, + { + "epoch": 9.626865671641792, + "grad_norm": 0.5866464972496033, + "learning_rate": 3.094214730815433e-05, + "loss": 0.9514, + "num_input_tokens_seen": 1472624, + "step": 5160 + }, + { + "epoch": 9.636194029850746, + "grad_norm": 0.47676122188568115, + "learning_rate": 3.09026032624248e-05, + "loss": 0.841, + "num_input_tokens_seen": 1474000, + "step": 5165 + }, + { + "epoch": 9.645522388059701, + "grad_norm": 0.7214521169662476, + "learning_rate": 3.086304357050302e-05, + "loss": 0.8298, + "num_input_tokens_seen": 1475376, + "step": 5170 + }, + { + "epoch": 9.654850746268657, + "grad_norm": 0.2885942757129669, + "learning_rate": 3.082346833725095e-05, + "loss": 0.6084, + "num_input_tokens_seen": 1476720, + "step": 5175 + }, + { + "epoch": 9.664179104477611, + "grad_norm": 0.22628052532672882, + "learning_rate": 3.078387766757177e-05, + "loss": 0.5386, + "num_input_tokens_seen": 1478192, + "step": 5180 + }, + { + "epoch": 9.673507462686567, + "grad_norm": 0.41684573888778687, + "learning_rate": 3.0744271666409524e-05, + "loss": 0.5459, + "num_input_tokens_seen": 1479792, + "step": 5185 + }, + { + "epoch": 9.682835820895523, + "grad_norm": 0.3898008465766907, + "learning_rate": 3.0704650438748946e-05, + "loss": 0.7306, + "num_input_tokens_seen": 1481104, + "step": 5190 + }, + { + "epoch": 9.692164179104477, + "grad_norm": 0.5324078798294067, + "learning_rate": 3.066501408961509e-05, + "loss": 0.984, + "num_input_tokens_seen": 1482448, + "step": 5195 + }, + { + "epoch": 9.701492537313433, + "grad_norm": 0.39649948477745056, + "learning_rate": 3.062536272407313e-05, + "loss": 0.5648, + "num_input_tokens_seen": 1484048, + "step": 5200 + }, + { + "epoch": 9.710820895522389, + "grad_norm": 0.4984428286552429, + "learning_rate": 3.0585696447228006e-05, + "loss": 0.7738, + "num_input_tokens_seen": 1485424, + "step": 5205 + }, + { + "epoch": 9.720149253731343, + "grad_norm": 0.4170304238796234, + "learning_rate": 3.054601536422423e-05, + "loss": 0.6427, + "num_input_tokens_seen": 1486800, + "step": 5210 + }, + { + "epoch": 9.729477611940299, + "grad_norm": 0.4860069453716278, + "learning_rate": 3.0506319580245536e-05, + "loss": 0.6512, + "num_input_tokens_seen": 1488240, + "step": 5215 + }, + { + "epoch": 9.738805970149254, + "grad_norm": 0.5349262952804565, + "learning_rate": 3.0466609200514605e-05, + "loss": 0.7038, + "num_input_tokens_seen": 1489840, + "step": 5220 + }, + { + "epoch": 9.748134328358208, + "grad_norm": 0.6571551561355591, + "learning_rate": 3.0426884330292842e-05, + "loss": 0.7388, + "num_input_tokens_seen": 1491312, + "step": 5225 + }, + { + "epoch": 9.757462686567164, + "grad_norm": 0.3546343743801117, + "learning_rate": 3.0387145074880053e-05, + "loss": 0.6273, + "num_input_tokens_seen": 1492656, + "step": 5230 + }, + { + "epoch": 9.76679104477612, + "grad_norm": 0.4753497540950775, + "learning_rate": 3.0347391539614156e-05, + "loss": 0.8262, + "num_input_tokens_seen": 1494096, + "step": 5235 + }, + { + "epoch": 9.776119402985074, + "grad_norm": 0.4580038785934448, + "learning_rate": 3.0307623829870952e-05, + "loss": 0.755, + "num_input_tokens_seen": 1495536, + "step": 5240 + }, + { + "epoch": 9.78544776119403, + "grad_norm": 0.42720481753349304, + "learning_rate": 3.0267842051063795e-05, + "loss": 0.7622, + "num_input_tokens_seen": 1497040, + "step": 5245 + }, + { + "epoch": 9.794776119402986, + "grad_norm": 0.3239773213863373, + "learning_rate": 3.022804630864333e-05, + "loss": 0.8564, + "num_input_tokens_seen": 1498576, + "step": 5250 + }, + { + "epoch": 9.80410447761194, + "grad_norm": 0.41649025678634644, + "learning_rate": 3.018823670809724e-05, + "loss": 0.5643, + "num_input_tokens_seen": 1500048, + "step": 5255 + }, + { + "epoch": 9.813432835820896, + "grad_norm": 0.5448052287101746, + "learning_rate": 3.0148413354949902e-05, + "loss": 0.7313, + "num_input_tokens_seen": 1501392, + "step": 5260 + }, + { + "epoch": 9.822761194029852, + "grad_norm": 0.580221951007843, + "learning_rate": 3.0108576354762175e-05, + "loss": 0.768, + "num_input_tokens_seen": 1502832, + "step": 5265 + }, + { + "epoch": 9.832089552238806, + "grad_norm": 0.4044698178768158, + "learning_rate": 3.0068725813131098e-05, + "loss": 0.7289, + "num_input_tokens_seen": 1504176, + "step": 5270 + }, + { + "epoch": 9.841417910447761, + "grad_norm": 0.5041797161102295, + "learning_rate": 3.0028861835689587e-05, + "loss": 0.7276, + "num_input_tokens_seen": 1505584, + "step": 5275 + }, + { + "epoch": 9.850746268656717, + "grad_norm": 0.4914662837982178, + "learning_rate": 2.9988984528106177e-05, + "loss": 0.7049, + "num_input_tokens_seen": 1506992, + "step": 5280 + }, + { + "epoch": 9.860074626865671, + "grad_norm": 0.3528186082839966, + "learning_rate": 2.9949093996084747e-05, + "loss": 0.7377, + "num_input_tokens_seen": 1508464, + "step": 5285 + }, + { + "epoch": 9.869402985074627, + "grad_norm": 0.3934158682823181, + "learning_rate": 2.9909190345364217e-05, + "loss": 0.737, + "num_input_tokens_seen": 1509840, + "step": 5290 + }, + { + "epoch": 9.878731343283581, + "grad_norm": 0.6380987167358398, + "learning_rate": 2.986927368171829e-05, + "loss": 1.0334, + "num_input_tokens_seen": 1511152, + "step": 5295 + }, + { + "epoch": 9.888059701492537, + "grad_norm": 0.3977501392364502, + "learning_rate": 2.982934411095516e-05, + "loss": 0.7865, + "num_input_tokens_seen": 1512560, + "step": 5300 + }, + { + "epoch": 9.897388059701493, + "grad_norm": 0.5469490885734558, + "learning_rate": 2.9789401738917244e-05, + "loss": 0.6823, + "num_input_tokens_seen": 1513968, + "step": 5305 + }, + { + "epoch": 9.906716417910447, + "grad_norm": 0.45253226161003113, + "learning_rate": 2.9749446671480862e-05, + "loss": 0.7541, + "num_input_tokens_seen": 1515376, + "step": 5310 + }, + { + "epoch": 9.916044776119403, + "grad_norm": 0.49398574233055115, + "learning_rate": 2.9709479014556024e-05, + "loss": 0.5598, + "num_input_tokens_seen": 1516720, + "step": 5315 + }, + { + "epoch": 9.925373134328359, + "grad_norm": 0.368939608335495, + "learning_rate": 2.966949887408608e-05, + "loss": 0.701, + "num_input_tokens_seen": 1518224, + "step": 5320 + }, + { + "epoch": 9.934701492537313, + "grad_norm": 0.9970903992652893, + "learning_rate": 2.96295063560475e-05, + "loss": 0.7701, + "num_input_tokens_seen": 1519440, + "step": 5325 + }, + { + "epoch": 9.944029850746269, + "grad_norm": 0.4684460461139679, + "learning_rate": 2.9589501566449534e-05, + "loss": 0.7168, + "num_input_tokens_seen": 1520944, + "step": 5330 + }, + { + "epoch": 9.953358208955224, + "grad_norm": 0.32310616970062256, + "learning_rate": 2.9549484611333983e-05, + "loss": 0.6905, + "num_input_tokens_seen": 1522736, + "step": 5335 + }, + { + "epoch": 9.962686567164178, + "grad_norm": 0.5298346877098083, + "learning_rate": 2.950945559677488e-05, + "loss": 0.764, + "num_input_tokens_seen": 1524336, + "step": 5340 + }, + { + "epoch": 9.972014925373134, + "grad_norm": 0.5397038459777832, + "learning_rate": 2.946941462887824e-05, + "loss": 0.5967, + "num_input_tokens_seen": 1525808, + "step": 5345 + }, + { + "epoch": 9.98134328358209, + "grad_norm": 0.5348801016807556, + "learning_rate": 2.942936181378174e-05, + "loss": 0.7652, + "num_input_tokens_seen": 1527152, + "step": 5350 + }, + { + "epoch": 9.990671641791044, + "grad_norm": 0.5427783727645874, + "learning_rate": 2.9389297257654482e-05, + "loss": 0.9463, + "num_input_tokens_seen": 1528400, + "step": 5355 + }, + { + "epoch": 10.0, + "grad_norm": 0.2525341212749481, + "learning_rate": 2.9349221066696693e-05, + "loss": 0.4365, + "num_input_tokens_seen": 1529640, + "step": 5360 + }, + { + "epoch": 10.0, + "eval_loss": 0.7237700819969177, + "eval_runtime": 2.4637, + "eval_samples_per_second": 96.603, + "eval_steps_per_second": 24.354, + "num_input_tokens_seen": 1529640, + "step": 5360 + }, + { + "epoch": 10.009328358208956, + "grad_norm": 0.363652765750885, + "learning_rate": 2.9309133347139412e-05, + "loss": 0.8791, + "num_input_tokens_seen": 1530888, + "step": 5365 + }, + { + "epoch": 10.01865671641791, + "grad_norm": 0.6078024506568909, + "learning_rate": 2.9269034205244272e-05, + "loss": 0.56, + "num_input_tokens_seen": 1532328, + "step": 5370 + }, + { + "epoch": 10.027985074626866, + "grad_norm": 0.6894094347953796, + "learning_rate": 2.922892374730316e-05, + "loss": 0.8305, + "num_input_tokens_seen": 1533736, + "step": 5375 + }, + { + "epoch": 10.037313432835822, + "grad_norm": 0.5886725783348083, + "learning_rate": 2.9188802079637966e-05, + "loss": 0.894, + "num_input_tokens_seen": 1535176, + "step": 5380 + }, + { + "epoch": 10.046641791044776, + "grad_norm": 0.5374459028244019, + "learning_rate": 2.9148669308600296e-05, + "loss": 1.0099, + "num_input_tokens_seen": 1536520, + "step": 5385 + }, + { + "epoch": 10.055970149253731, + "grad_norm": 0.3218475580215454, + "learning_rate": 2.910852554057118e-05, + "loss": 0.4397, + "num_input_tokens_seen": 1538152, + "step": 5390 + }, + { + "epoch": 10.065298507462687, + "grad_norm": 0.7223525047302246, + "learning_rate": 2.9068370881960817e-05, + "loss": 0.7851, + "num_input_tokens_seen": 1539688, + "step": 5395 + }, + { + "epoch": 10.074626865671641, + "grad_norm": 0.5029573440551758, + "learning_rate": 2.902820543920825e-05, + "loss": 0.7446, + "num_input_tokens_seen": 1541160, + "step": 5400 + }, + { + "epoch": 10.083955223880597, + "grad_norm": 0.21922869980335236, + "learning_rate": 2.8988029318781124e-05, + "loss": 0.5891, + "num_input_tokens_seen": 1542664, + "step": 5405 + }, + { + "epoch": 10.093283582089553, + "grad_norm": 0.46875396370887756, + "learning_rate": 2.894784262717538e-05, + "loss": 0.7732, + "num_input_tokens_seen": 1543976, + "step": 5410 + }, + { + "epoch": 10.102611940298507, + "grad_norm": 0.5124669671058655, + "learning_rate": 2.8907645470914978e-05, + "loss": 0.8183, + "num_input_tokens_seen": 1545352, + "step": 5415 + }, + { + "epoch": 10.111940298507463, + "grad_norm": 0.39394158124923706, + "learning_rate": 2.886743795655164e-05, + "loss": 0.4718, + "num_input_tokens_seen": 1546792, + "step": 5420 + }, + { + "epoch": 10.121268656716419, + "grad_norm": 0.3458622395992279, + "learning_rate": 2.8827220190664506e-05, + "loss": 0.6081, + "num_input_tokens_seen": 1548200, + "step": 5425 + }, + { + "epoch": 10.130597014925373, + "grad_norm": 0.45292478799819946, + "learning_rate": 2.8786992279859922e-05, + "loss": 0.7235, + "num_input_tokens_seen": 1549832, + "step": 5430 + }, + { + "epoch": 10.139925373134329, + "grad_norm": 0.5708794593811035, + "learning_rate": 2.8746754330771104e-05, + "loss": 0.8325, + "num_input_tokens_seen": 1551144, + "step": 5435 + }, + { + "epoch": 10.149253731343283, + "grad_norm": 0.32183316349983215, + "learning_rate": 2.8706506450057902e-05, + "loss": 0.7402, + "num_input_tokens_seen": 1552584, + "step": 5440 + }, + { + "epoch": 10.158582089552239, + "grad_norm": 0.4306357800960541, + "learning_rate": 2.8666248744406454e-05, + "loss": 0.5788, + "num_input_tokens_seen": 1554056, + "step": 5445 + }, + { + "epoch": 10.167910447761194, + "grad_norm": 0.4137272238731384, + "learning_rate": 2.862598132052898e-05, + "loss": 0.7662, + "num_input_tokens_seen": 1555560, + "step": 5450 + }, + { + "epoch": 10.177238805970148, + "grad_norm": 0.4524538516998291, + "learning_rate": 2.8585704285163454e-05, + "loss": 0.6451, + "num_input_tokens_seen": 1557160, + "step": 5455 + }, + { + "epoch": 10.186567164179104, + "grad_norm": 0.4747234284877777, + "learning_rate": 2.8545417745073294e-05, + "loss": 0.8216, + "num_input_tokens_seen": 1558600, + "step": 5460 + }, + { + "epoch": 10.19589552238806, + "grad_norm": 0.46810322999954224, + "learning_rate": 2.850512180704715e-05, + "loss": 0.6458, + "num_input_tokens_seen": 1560104, + "step": 5465 + }, + { + "epoch": 10.205223880597014, + "grad_norm": 0.4478859603404999, + "learning_rate": 2.846481657789856e-05, + "loss": 0.5344, + "num_input_tokens_seen": 1561544, + "step": 5470 + }, + { + "epoch": 10.21455223880597, + "grad_norm": 0.5770183801651001, + "learning_rate": 2.8424502164465705e-05, + "loss": 0.8732, + "num_input_tokens_seen": 1562888, + "step": 5475 + }, + { + "epoch": 10.223880597014926, + "grad_norm": 0.5467674136161804, + "learning_rate": 2.838417867361111e-05, + "loss": 0.702, + "num_input_tokens_seen": 1564392, + "step": 5480 + }, + { + "epoch": 10.23320895522388, + "grad_norm": 0.4422265589237213, + "learning_rate": 2.8343846212221354e-05, + "loss": 0.714, + "num_input_tokens_seen": 1565832, + "step": 5485 + }, + { + "epoch": 10.242537313432836, + "grad_norm": 0.46220991015434265, + "learning_rate": 2.8303504887206794e-05, + "loss": 0.6096, + "num_input_tokens_seen": 1567432, + "step": 5490 + }, + { + "epoch": 10.251865671641792, + "grad_norm": 0.452933132648468, + "learning_rate": 2.8263154805501297e-05, + "loss": 0.9737, + "num_input_tokens_seen": 1568744, + "step": 5495 + }, + { + "epoch": 10.261194029850746, + "grad_norm": 0.2999035716056824, + "learning_rate": 2.8222796074061907e-05, + "loss": 0.7094, + "num_input_tokens_seen": 1570216, + "step": 5500 + }, + { + "epoch": 10.270522388059701, + "grad_norm": 0.6962710618972778, + "learning_rate": 2.8182428799868645e-05, + "loss": 0.7779, + "num_input_tokens_seen": 1571592, + "step": 5505 + }, + { + "epoch": 10.279850746268657, + "grad_norm": 0.3356461822986603, + "learning_rate": 2.8142053089924142e-05, + "loss": 0.5947, + "num_input_tokens_seen": 1573064, + "step": 5510 + }, + { + "epoch": 10.289179104477611, + "grad_norm": 0.7310212850570679, + "learning_rate": 2.8101669051253392e-05, + "loss": 0.6792, + "num_input_tokens_seen": 1574408, + "step": 5515 + }, + { + "epoch": 10.298507462686567, + "grad_norm": 0.48236924409866333, + "learning_rate": 2.806127679090349e-05, + "loss": 0.7894, + "num_input_tokens_seen": 1575688, + "step": 5520 + }, + { + "epoch": 10.307835820895523, + "grad_norm": 0.7519316077232361, + "learning_rate": 2.8020876415943287e-05, + "loss": 0.6931, + "num_input_tokens_seen": 1577160, + "step": 5525 + }, + { + "epoch": 10.317164179104477, + "grad_norm": 0.3200331926345825, + "learning_rate": 2.7980468033463175e-05, + "loss": 0.5382, + "num_input_tokens_seen": 1578632, + "step": 5530 + }, + { + "epoch": 10.326492537313433, + "grad_norm": 0.29282453656196594, + "learning_rate": 2.7940051750574763e-05, + "loss": 0.8229, + "num_input_tokens_seen": 1580360, + "step": 5535 + }, + { + "epoch": 10.335820895522389, + "grad_norm": 0.35867422819137573, + "learning_rate": 2.7899627674410587e-05, + "loss": 0.5787, + "num_input_tokens_seen": 1581736, + "step": 5540 + }, + { + "epoch": 10.345149253731343, + "grad_norm": 0.6517281532287598, + "learning_rate": 2.7859195912123874e-05, + "loss": 0.6503, + "num_input_tokens_seen": 1583272, + "step": 5545 + }, + { + "epoch": 10.354477611940299, + "grad_norm": 0.36823412775993347, + "learning_rate": 2.7818756570888194e-05, + "loss": 0.643, + "num_input_tokens_seen": 1584648, + "step": 5550 + }, + { + "epoch": 10.363805970149254, + "grad_norm": 0.3981950283050537, + "learning_rate": 2.7778309757897213e-05, + "loss": 0.5017, + "num_input_tokens_seen": 1586120, + "step": 5555 + }, + { + "epoch": 10.373134328358208, + "grad_norm": 0.4670792818069458, + "learning_rate": 2.77378555803644e-05, + "loss": 0.819, + "num_input_tokens_seen": 1587528, + "step": 5560 + }, + { + "epoch": 10.382462686567164, + "grad_norm": 0.7728878259658813, + "learning_rate": 2.7697394145522775e-05, + "loss": 0.7891, + "num_input_tokens_seen": 1588872, + "step": 5565 + }, + { + "epoch": 10.39179104477612, + "grad_norm": 0.4525209665298462, + "learning_rate": 2.765692556062456e-05, + "loss": 0.7432, + "num_input_tokens_seen": 1590312, + "step": 5570 + }, + { + "epoch": 10.401119402985074, + "grad_norm": 0.7338618040084839, + "learning_rate": 2.7616449932940942e-05, + "loss": 0.5922, + "num_input_tokens_seen": 1591528, + "step": 5575 + }, + { + "epoch": 10.41044776119403, + "grad_norm": 0.46827641129493713, + "learning_rate": 2.7575967369761775e-05, + "loss": 0.617, + "num_input_tokens_seen": 1593096, + "step": 5580 + }, + { + "epoch": 10.419776119402986, + "grad_norm": 0.45147886872291565, + "learning_rate": 2.7535477978395297e-05, + "loss": 1.0728, + "num_input_tokens_seen": 1594280, + "step": 5585 + }, + { + "epoch": 10.42910447761194, + "grad_norm": 0.3405883312225342, + "learning_rate": 2.749498186616785e-05, + "loss": 0.5849, + "num_input_tokens_seen": 1595496, + "step": 5590 + }, + { + "epoch": 10.438432835820896, + "grad_norm": 0.6466345191001892, + "learning_rate": 2.745447914042359e-05, + "loss": 0.6559, + "num_input_tokens_seen": 1596840, + "step": 5595 + }, + { + "epoch": 10.447761194029852, + "grad_norm": 0.3541865944862366, + "learning_rate": 2.7413969908524206e-05, + "loss": 0.6736, + "num_input_tokens_seen": 1598248, + "step": 5600 + }, + { + "epoch": 10.457089552238806, + "grad_norm": 0.40969812870025635, + "learning_rate": 2.7373454277848622e-05, + "loss": 0.9506, + "num_input_tokens_seen": 1599496, + "step": 5605 + }, + { + "epoch": 10.466417910447761, + "grad_norm": 0.5418379902839661, + "learning_rate": 2.733293235579274e-05, + "loss": 0.9394, + "num_input_tokens_seen": 1600968, + "step": 5610 + }, + { + "epoch": 10.475746268656717, + "grad_norm": 0.38557353615760803, + "learning_rate": 2.729240424976911e-05, + "loss": 0.5991, + "num_input_tokens_seen": 1602344, + "step": 5615 + }, + { + "epoch": 10.485074626865671, + "grad_norm": 0.48555323481559753, + "learning_rate": 2.7251870067206715e-05, + "loss": 0.7512, + "num_input_tokens_seen": 1603880, + "step": 5620 + }, + { + "epoch": 10.494402985074627, + "grad_norm": 0.5015121698379517, + "learning_rate": 2.7211329915550615e-05, + "loss": 1.0154, + "num_input_tokens_seen": 1605192, + "step": 5625 + }, + { + "epoch": 10.503731343283581, + "grad_norm": 0.4917770326137543, + "learning_rate": 2.7170783902261692e-05, + "loss": 0.8249, + "num_input_tokens_seen": 1606568, + "step": 5630 + }, + { + "epoch": 10.513059701492537, + "grad_norm": 0.5721375346183777, + "learning_rate": 2.7130232134816397e-05, + "loss": 0.7712, + "num_input_tokens_seen": 1607880, + "step": 5635 + }, + { + "epoch": 10.522388059701493, + "grad_norm": 0.3120163381099701, + "learning_rate": 2.7089674720706387e-05, + "loss": 0.7042, + "num_input_tokens_seen": 1609096, + "step": 5640 + }, + { + "epoch": 10.531716417910447, + "grad_norm": 0.3758350908756256, + "learning_rate": 2.704911176743833e-05, + "loss": 0.6765, + "num_input_tokens_seen": 1610664, + "step": 5645 + }, + { + "epoch": 10.541044776119403, + "grad_norm": 0.3607464134693146, + "learning_rate": 2.7008543382533545e-05, + "loss": 0.7629, + "num_input_tokens_seen": 1612136, + "step": 5650 + }, + { + "epoch": 10.550373134328359, + "grad_norm": 0.492830753326416, + "learning_rate": 2.6967969673527764e-05, + "loss": 0.4677, + "num_input_tokens_seen": 1613512, + "step": 5655 + }, + { + "epoch": 10.559701492537313, + "grad_norm": 0.992146909236908, + "learning_rate": 2.6927390747970843e-05, + "loss": 0.6491, + "num_input_tokens_seen": 1614888, + "step": 5660 + }, + { + "epoch": 10.569029850746269, + "grad_norm": 0.5607097744941711, + "learning_rate": 2.6886806713426434e-05, + "loss": 0.8168, + "num_input_tokens_seen": 1616168, + "step": 5665 + }, + { + "epoch": 10.578358208955224, + "grad_norm": 0.5378861427307129, + "learning_rate": 2.6846217677471765e-05, + "loss": 0.694, + "num_input_tokens_seen": 1617704, + "step": 5670 + }, + { + "epoch": 10.587686567164178, + "grad_norm": 0.4745161831378937, + "learning_rate": 2.6805623747697283e-05, + "loss": 0.8164, + "num_input_tokens_seen": 1619144, + "step": 5675 + }, + { + "epoch": 10.597014925373134, + "grad_norm": 0.520386815071106, + "learning_rate": 2.6765025031706453e-05, + "loss": 0.5692, + "num_input_tokens_seen": 1620616, + "step": 5680 + }, + { + "epoch": 10.60634328358209, + "grad_norm": 0.5510054230690002, + "learning_rate": 2.67244216371154e-05, + "loss": 0.8371, + "num_input_tokens_seen": 1621960, + "step": 5685 + }, + { + "epoch": 10.615671641791044, + "grad_norm": 0.48242998123168945, + "learning_rate": 2.668381367155265e-05, + "loss": 0.8542, + "num_input_tokens_seen": 1623496, + "step": 5690 + }, + { + "epoch": 10.625, + "grad_norm": 0.41213440895080566, + "learning_rate": 2.664320124265885e-05, + "loss": 0.787, + "num_input_tokens_seen": 1624808, + "step": 5695 + }, + { + "epoch": 10.634328358208956, + "grad_norm": 0.47447457909584045, + "learning_rate": 2.660258445808648e-05, + "loss": 0.9606, + "num_input_tokens_seen": 1626088, + "step": 5700 + }, + { + "epoch": 10.64365671641791, + "grad_norm": 0.3711308538913727, + "learning_rate": 2.6561963425499574e-05, + "loss": 0.6773, + "num_input_tokens_seen": 1627400, + "step": 5705 + }, + { + "epoch": 10.652985074626866, + "grad_norm": 0.5159795880317688, + "learning_rate": 2.652133825257339e-05, + "loss": 0.7168, + "num_input_tokens_seen": 1628776, + "step": 5710 + }, + { + "epoch": 10.662313432835822, + "grad_norm": 0.4027276933193207, + "learning_rate": 2.6480709046994218e-05, + "loss": 0.6658, + "num_input_tokens_seen": 1630184, + "step": 5715 + }, + { + "epoch": 10.671641791044776, + "grad_norm": 0.37922146916389465, + "learning_rate": 2.6440075916458982e-05, + "loss": 0.6873, + "num_input_tokens_seen": 1631720, + "step": 5720 + }, + { + "epoch": 10.680970149253731, + "grad_norm": 0.326651006937027, + "learning_rate": 2.6399438968675055e-05, + "loss": 0.4891, + "num_input_tokens_seen": 1633224, + "step": 5725 + }, + { + "epoch": 10.690298507462687, + "grad_norm": 0.30386635661125183, + "learning_rate": 2.635879831135989e-05, + "loss": 0.5003, + "num_input_tokens_seen": 1634792, + "step": 5730 + }, + { + "epoch": 10.699626865671641, + "grad_norm": 0.6038807034492493, + "learning_rate": 2.6318154052240807e-05, + "loss": 0.8002, + "num_input_tokens_seen": 1636104, + "step": 5735 + }, + { + "epoch": 10.708955223880597, + "grad_norm": 0.4443417191505432, + "learning_rate": 2.6277506299054645e-05, + "loss": 0.5851, + "num_input_tokens_seen": 1637672, + "step": 5740 + }, + { + "epoch": 10.718283582089553, + "grad_norm": 0.44301965832710266, + "learning_rate": 2.6236855159547525e-05, + "loss": 0.846, + "num_input_tokens_seen": 1639016, + "step": 5745 + }, + { + "epoch": 10.727611940298507, + "grad_norm": 0.3685060441493988, + "learning_rate": 2.6196200741474534e-05, + "loss": 0.7317, + "num_input_tokens_seen": 1640584, + "step": 5750 + }, + { + "epoch": 10.736940298507463, + "grad_norm": 0.5438111424446106, + "learning_rate": 2.6155543152599455e-05, + "loss": 0.8338, + "num_input_tokens_seen": 1642120, + "step": 5755 + }, + { + "epoch": 10.746268656716419, + "grad_norm": 0.5570796728134155, + "learning_rate": 2.611488250069447e-05, + "loss": 0.7041, + "num_input_tokens_seen": 1643656, + "step": 5760 + }, + { + "epoch": 10.755597014925373, + "grad_norm": 0.30568844079971313, + "learning_rate": 2.6074218893539885e-05, + "loss": 0.5825, + "num_input_tokens_seen": 1644904, + "step": 5765 + }, + { + "epoch": 10.764925373134329, + "grad_norm": 0.4486890435218811, + "learning_rate": 2.6033552438923837e-05, + "loss": 0.6124, + "num_input_tokens_seen": 1646280, + "step": 5770 + }, + { + "epoch": 10.774253731343283, + "grad_norm": 0.4467718005180359, + "learning_rate": 2.5992883244642014e-05, + "loss": 0.6119, + "num_input_tokens_seen": 1647816, + "step": 5775 + }, + { + "epoch": 10.783582089552239, + "grad_norm": 0.4625658094882965, + "learning_rate": 2.5952211418497358e-05, + "loss": 0.496, + "num_input_tokens_seen": 1649288, + "step": 5780 + }, + { + "epoch": 10.792910447761194, + "grad_norm": 0.3831215798854828, + "learning_rate": 2.5911537068299802e-05, + "loss": 0.6014, + "num_input_tokens_seen": 1650760, + "step": 5785 + }, + { + "epoch": 10.802238805970148, + "grad_norm": 0.2625681459903717, + "learning_rate": 2.5870860301865944e-05, + "loss": 0.4537, + "num_input_tokens_seen": 1652392, + "step": 5790 + }, + { + "epoch": 10.811567164179104, + "grad_norm": 0.44546982645988464, + "learning_rate": 2.5830181227018833e-05, + "loss": 0.6328, + "num_input_tokens_seen": 1653800, + "step": 5795 + }, + { + "epoch": 10.82089552238806, + "grad_norm": 0.4501853287220001, + "learning_rate": 2.5789499951587575e-05, + "loss": 0.6935, + "num_input_tokens_seen": 1655368, + "step": 5800 + }, + { + "epoch": 10.830223880597014, + "grad_norm": 0.6749563813209534, + "learning_rate": 2.5748816583407163e-05, + "loss": 0.615, + "num_input_tokens_seen": 1656712, + "step": 5805 + }, + { + "epoch": 10.83955223880597, + "grad_norm": 0.47277528047561646, + "learning_rate": 2.570813123031811e-05, + "loss": 0.8365, + "num_input_tokens_seen": 1658056, + "step": 5810 + }, + { + "epoch": 10.848880597014926, + "grad_norm": 0.5621316432952881, + "learning_rate": 2.5667444000166203e-05, + "loss": 0.9027, + "num_input_tokens_seen": 1659400, + "step": 5815 + }, + { + "epoch": 10.85820895522388, + "grad_norm": 0.38922998309135437, + "learning_rate": 2.5626755000802188e-05, + "loss": 0.8515, + "num_input_tokens_seen": 1660840, + "step": 5820 + }, + { + "epoch": 10.867537313432836, + "grad_norm": 0.575376033782959, + "learning_rate": 2.5586064340081516e-05, + "loss": 0.5486, + "num_input_tokens_seen": 1662248, + "step": 5825 + }, + { + "epoch": 10.876865671641792, + "grad_norm": 0.5097258687019348, + "learning_rate": 2.5545372125864032e-05, + "loss": 0.7214, + "num_input_tokens_seen": 1663592, + "step": 5830 + }, + { + "epoch": 10.886194029850746, + "grad_norm": 0.3734528124332428, + "learning_rate": 2.5504678466013705e-05, + "loss": 0.6188, + "num_input_tokens_seen": 1665032, + "step": 5835 + }, + { + "epoch": 10.895522388059701, + "grad_norm": 0.3539853096008301, + "learning_rate": 2.546398346839834e-05, + "loss": 0.6212, + "num_input_tokens_seen": 1666696, + "step": 5840 + }, + { + "epoch": 10.904850746268657, + "grad_norm": 0.354053258895874, + "learning_rate": 2.5423287240889277e-05, + "loss": 0.7334, + "num_input_tokens_seen": 1668168, + "step": 5845 + }, + { + "epoch": 10.914179104477611, + "grad_norm": 0.4186680018901825, + "learning_rate": 2.5382589891361125e-05, + "loss": 0.8075, + "num_input_tokens_seen": 1669640, + "step": 5850 + }, + { + "epoch": 10.923507462686567, + "grad_norm": 0.3145810663700104, + "learning_rate": 2.5341891527691457e-05, + "loss": 0.598, + "num_input_tokens_seen": 1671016, + "step": 5855 + }, + { + "epoch": 10.932835820895523, + "grad_norm": 0.592525839805603, + "learning_rate": 2.5301192257760555e-05, + "loss": 0.6583, + "num_input_tokens_seen": 1672392, + "step": 5860 + }, + { + "epoch": 10.942164179104477, + "grad_norm": 0.4756569266319275, + "learning_rate": 2.5260492189451073e-05, + "loss": 0.7808, + "num_input_tokens_seen": 1673864, + "step": 5865 + }, + { + "epoch": 10.951492537313433, + "grad_norm": 0.38150256872177124, + "learning_rate": 2.521979143064781e-05, + "loss": 0.7377, + "num_input_tokens_seen": 1675208, + "step": 5870 + }, + { + "epoch": 10.960820895522389, + "grad_norm": 0.4519204795360565, + "learning_rate": 2.5179090089237378e-05, + "loss": 0.9045, + "num_input_tokens_seen": 1676616, + "step": 5875 + }, + { + "epoch": 10.970149253731343, + "grad_norm": 0.41227924823760986, + "learning_rate": 2.5138388273107932e-05, + "loss": 0.7074, + "num_input_tokens_seen": 1677992, + "step": 5880 + }, + { + "epoch": 10.979477611940299, + "grad_norm": 0.31640151143074036, + "learning_rate": 2.5097686090148904e-05, + "loss": 0.7883, + "num_input_tokens_seen": 1679464, + "step": 5885 + }, + { + "epoch": 10.988805970149254, + "grad_norm": 0.4898882508277893, + "learning_rate": 2.5056983648250677e-05, + "loss": 0.6696, + "num_input_tokens_seen": 1680968, + "step": 5890 + }, + { + "epoch": 10.998134328358208, + "grad_norm": 0.529682457447052, + "learning_rate": 2.501628105530433e-05, + "loss": 0.8136, + "num_input_tokens_seen": 1682280, + "step": 5895 + }, + { + "epoch": 11.0, + "eval_loss": 0.7153291702270508, + "eval_runtime": 2.452, + "eval_samples_per_second": 97.064, + "eval_steps_per_second": 24.47, + "num_input_tokens_seen": 1682336, + "step": 5896 + }, + { + "epoch": 11.007462686567164, + "grad_norm": 0.3709951639175415, + "learning_rate": 2.4975578419201347e-05, + "loss": 0.711, + "num_input_tokens_seen": 1683488, + "step": 5900 + }, + { + "epoch": 11.01679104477612, + "grad_norm": 0.5264549255371094, + "learning_rate": 2.4934875847833308e-05, + "loss": 0.7659, + "num_input_tokens_seen": 1684896, + "step": 5905 + }, + { + "epoch": 11.026119402985074, + "grad_norm": 0.4328784942626953, + "learning_rate": 2.489417344909166e-05, + "loss": 0.7947, + "num_input_tokens_seen": 1686240, + "step": 5910 + }, + { + "epoch": 11.03544776119403, + "grad_norm": 0.7111842036247253, + "learning_rate": 2.4853471330867335e-05, + "loss": 0.7677, + "num_input_tokens_seen": 1687552, + "step": 5915 + }, + { + "epoch": 11.044776119402986, + "grad_norm": 0.4643982946872711, + "learning_rate": 2.4812769601050587e-05, + "loss": 0.6495, + "num_input_tokens_seen": 1688896, + "step": 5920 + }, + { + "epoch": 11.05410447761194, + "grad_norm": 0.47614234685897827, + "learning_rate": 2.477206836753057e-05, + "loss": 0.8168, + "num_input_tokens_seen": 1690336, + "step": 5925 + }, + { + "epoch": 11.063432835820896, + "grad_norm": 0.5013654232025146, + "learning_rate": 2.4731367738195195e-05, + "loss": 0.6477, + "num_input_tokens_seen": 1691872, + "step": 5930 + }, + { + "epoch": 11.072761194029852, + "grad_norm": 0.43805474042892456, + "learning_rate": 2.4690667820930706e-05, + "loss": 0.5216, + "num_input_tokens_seen": 1693248, + "step": 5935 + }, + { + "epoch": 11.082089552238806, + "grad_norm": 0.43377432227134705, + "learning_rate": 2.4649968723621502e-05, + "loss": 0.6628, + "num_input_tokens_seen": 1694752, + "step": 5940 + }, + { + "epoch": 11.091417910447761, + "grad_norm": 0.4986671805381775, + "learning_rate": 2.460927055414981e-05, + "loss": 0.773, + "num_input_tokens_seen": 1696000, + "step": 5945 + }, + { + "epoch": 11.100746268656716, + "grad_norm": 0.7377024292945862, + "learning_rate": 2.4568573420395354e-05, + "loss": 0.6457, + "num_input_tokens_seen": 1697568, + "step": 5950 + }, + { + "epoch": 11.110074626865671, + "grad_norm": 0.530397355556488, + "learning_rate": 2.452787743023517e-05, + "loss": 0.506, + "num_input_tokens_seen": 1699200, + "step": 5955 + }, + { + "epoch": 11.119402985074627, + "grad_norm": 0.7094915509223938, + "learning_rate": 2.4487182691543206e-05, + "loss": 0.7309, + "num_input_tokens_seen": 1700512, + "step": 5960 + }, + { + "epoch": 11.128731343283581, + "grad_norm": 0.396619975566864, + "learning_rate": 2.4446489312190144e-05, + "loss": 0.7621, + "num_input_tokens_seen": 1702016, + "step": 5965 + }, + { + "epoch": 11.138059701492537, + "grad_norm": 0.4564957022666931, + "learning_rate": 2.4405797400043034e-05, + "loss": 0.6048, + "num_input_tokens_seen": 1703456, + "step": 5970 + }, + { + "epoch": 11.147388059701493, + "grad_norm": 0.5083081126213074, + "learning_rate": 2.436510706296504e-05, + "loss": 0.7432, + "num_input_tokens_seen": 1704768, + "step": 5975 + }, + { + "epoch": 11.156716417910447, + "grad_norm": 0.5034011006355286, + "learning_rate": 2.432441840881516e-05, + "loss": 0.6479, + "num_input_tokens_seen": 1706176, + "step": 5980 + }, + { + "epoch": 11.166044776119403, + "grad_norm": 0.39043909311294556, + "learning_rate": 2.428373154544791e-05, + "loss": 0.818, + "num_input_tokens_seen": 1707552, + "step": 5985 + }, + { + "epoch": 11.175373134328359, + "grad_norm": 0.4669814109802246, + "learning_rate": 2.424304658071309e-05, + "loss": 0.8447, + "num_input_tokens_seen": 1708960, + "step": 5990 + }, + { + "epoch": 11.184701492537313, + "grad_norm": 0.26795756816864014, + "learning_rate": 2.4202363622455442e-05, + "loss": 0.7119, + "num_input_tokens_seen": 1710368, + "step": 5995 + }, + { + "epoch": 11.194029850746269, + "grad_norm": 0.5347198843955994, + "learning_rate": 2.4161682778514403e-05, + "loss": 0.7169, + "num_input_tokens_seen": 1711776, + "step": 6000 + }, + { + "epoch": 11.203358208955224, + "grad_norm": 0.38599684834480286, + "learning_rate": 2.4121004156723802e-05, + "loss": 0.8335, + "num_input_tokens_seen": 1713120, + "step": 6005 + }, + { + "epoch": 11.212686567164178, + "grad_norm": 0.5371614694595337, + "learning_rate": 2.4080327864911567e-05, + "loss": 0.7013, + "num_input_tokens_seen": 1714464, + "step": 6010 + }, + { + "epoch": 11.222014925373134, + "grad_norm": 0.3844442069530487, + "learning_rate": 2.403965401089947e-05, + "loss": 0.69, + "num_input_tokens_seen": 1715808, + "step": 6015 + }, + { + "epoch": 11.23134328358209, + "grad_norm": 0.3364402949810028, + "learning_rate": 2.3998982702502807e-05, + "loss": 0.6439, + "num_input_tokens_seen": 1717536, + "step": 6020 + }, + { + "epoch": 11.240671641791044, + "grad_norm": 0.39194613695144653, + "learning_rate": 2.3958314047530125e-05, + "loss": 0.5591, + "num_input_tokens_seen": 1718816, + "step": 6025 + }, + { + "epoch": 11.25, + "grad_norm": 0.41498756408691406, + "learning_rate": 2.3917648153782956e-05, + "loss": 0.5663, + "num_input_tokens_seen": 1720224, + "step": 6030 + }, + { + "epoch": 11.259328358208956, + "grad_norm": 0.46931102871894836, + "learning_rate": 2.3876985129055486e-05, + "loss": 0.9218, + "num_input_tokens_seen": 1721568, + "step": 6035 + }, + { + "epoch": 11.26865671641791, + "grad_norm": 0.4897795617580414, + "learning_rate": 2.3836325081134314e-05, + "loss": 0.6896, + "num_input_tokens_seen": 1723040, + "step": 6040 + }, + { + "epoch": 11.277985074626866, + "grad_norm": 0.5477071404457092, + "learning_rate": 2.3795668117798138e-05, + "loss": 0.7458, + "num_input_tokens_seen": 1724544, + "step": 6045 + }, + { + "epoch": 11.287313432835822, + "grad_norm": 0.7365218997001648, + "learning_rate": 2.37550143468175e-05, + "loss": 0.5121, + "num_input_tokens_seen": 1726016, + "step": 6050 + }, + { + "epoch": 11.296641791044776, + "grad_norm": 0.48599621653556824, + "learning_rate": 2.3714363875954447e-05, + "loss": 0.6824, + "num_input_tokens_seen": 1727488, + "step": 6055 + }, + { + "epoch": 11.305970149253731, + "grad_norm": 0.3254573345184326, + "learning_rate": 2.36737168129623e-05, + "loss": 0.9167, + "num_input_tokens_seen": 1728800, + "step": 6060 + }, + { + "epoch": 11.315298507462687, + "grad_norm": 0.3482373356819153, + "learning_rate": 2.3633073265585356e-05, + "loss": 0.5388, + "num_input_tokens_seen": 1730176, + "step": 6065 + }, + { + "epoch": 11.324626865671641, + "grad_norm": 0.4062928259372711, + "learning_rate": 2.3592433341558563e-05, + "loss": 0.6622, + "num_input_tokens_seen": 1731616, + "step": 6070 + }, + { + "epoch": 11.333955223880597, + "grad_norm": 0.4049259126186371, + "learning_rate": 2.3551797148607298e-05, + "loss": 0.8275, + "num_input_tokens_seen": 1732800, + "step": 6075 + }, + { + "epoch": 11.343283582089553, + "grad_norm": 0.37475481629371643, + "learning_rate": 2.3511164794447015e-05, + "loss": 0.7594, + "num_input_tokens_seen": 1734176, + "step": 6080 + }, + { + "epoch": 11.352611940298507, + "grad_norm": 0.35297706723213196, + "learning_rate": 2.347053638678302e-05, + "loss": 0.6507, + "num_input_tokens_seen": 1735744, + "step": 6085 + }, + { + "epoch": 11.361940298507463, + "grad_norm": 0.5052271485328674, + "learning_rate": 2.3429912033310143e-05, + "loss": 0.8036, + "num_input_tokens_seen": 1737408, + "step": 6090 + }, + { + "epoch": 11.371268656716419, + "grad_norm": 0.40577614307403564, + "learning_rate": 2.338929184171247e-05, + "loss": 0.7839, + "num_input_tokens_seen": 1738912, + "step": 6095 + }, + { + "epoch": 11.380597014925373, + "grad_norm": 0.4513566792011261, + "learning_rate": 2.3348675919663065e-05, + "loss": 0.8246, + "num_input_tokens_seen": 1740256, + "step": 6100 + }, + { + "epoch": 11.389925373134329, + "grad_norm": 0.444598525762558, + "learning_rate": 2.330806437482365e-05, + "loss": 0.6385, + "num_input_tokens_seen": 1741728, + "step": 6105 + }, + { + "epoch": 11.399253731343283, + "grad_norm": 0.6836233735084534, + "learning_rate": 2.3267457314844372e-05, + "loss": 0.8039, + "num_input_tokens_seen": 1743264, + "step": 6110 + }, + { + "epoch": 11.408582089552239, + "grad_norm": 0.5955387949943542, + "learning_rate": 2.3226854847363473e-05, + "loss": 0.8057, + "num_input_tokens_seen": 1744608, + "step": 6115 + }, + { + "epoch": 11.417910447761194, + "grad_norm": 0.7112632393836975, + "learning_rate": 2.3186257080007016e-05, + "loss": 0.8301, + "num_input_tokens_seen": 1745952, + "step": 6120 + }, + { + "epoch": 11.427238805970148, + "grad_norm": 0.40500038862228394, + "learning_rate": 2.314566412038865e-05, + "loss": 0.7569, + "num_input_tokens_seen": 1747392, + "step": 6125 + }, + { + "epoch": 11.436567164179104, + "grad_norm": 0.28113678097724915, + "learning_rate": 2.31050760761092e-05, + "loss": 0.4642, + "num_input_tokens_seen": 1748928, + "step": 6130 + }, + { + "epoch": 11.44589552238806, + "grad_norm": 0.5434813499450684, + "learning_rate": 2.306449305475655e-05, + "loss": 0.7334, + "num_input_tokens_seen": 1750624, + "step": 6135 + }, + { + "epoch": 11.455223880597014, + "grad_norm": 0.8741121292114258, + "learning_rate": 2.3023915163905198e-05, + "loss": 0.6358, + "num_input_tokens_seen": 1752064, + "step": 6140 + }, + { + "epoch": 11.46455223880597, + "grad_norm": 0.705390214920044, + "learning_rate": 2.298334251111607e-05, + "loss": 0.6578, + "num_input_tokens_seen": 1753376, + "step": 6145 + }, + { + "epoch": 11.473880597014926, + "grad_norm": 0.5385904312133789, + "learning_rate": 2.2942775203936238e-05, + "loss": 0.8614, + "num_input_tokens_seen": 1754784, + "step": 6150 + }, + { + "epoch": 11.48320895522388, + "grad_norm": 0.20257875323295593, + "learning_rate": 2.290221334989854e-05, + "loss": 0.5883, + "num_input_tokens_seen": 1756544, + "step": 6155 + }, + { + "epoch": 11.492537313432836, + "grad_norm": 0.5672345757484436, + "learning_rate": 2.286165705652143e-05, + "loss": 0.804, + "num_input_tokens_seen": 1757952, + "step": 6160 + }, + { + "epoch": 11.501865671641792, + "grad_norm": 0.4762251675128937, + "learning_rate": 2.2821106431308544e-05, + "loss": 0.7736, + "num_input_tokens_seen": 1759360, + "step": 6165 + }, + { + "epoch": 11.511194029850746, + "grad_norm": 0.44462209939956665, + "learning_rate": 2.278056158174858e-05, + "loss": 0.7083, + "num_input_tokens_seen": 1760928, + "step": 6170 + }, + { + "epoch": 11.520522388059701, + "grad_norm": 0.688033938407898, + "learning_rate": 2.274002261531484e-05, + "loss": 1.0177, + "num_input_tokens_seen": 1762208, + "step": 6175 + }, + { + "epoch": 11.529850746268657, + "grad_norm": 0.5024594664573669, + "learning_rate": 2.2699489639465103e-05, + "loss": 0.8101, + "num_input_tokens_seen": 1763808, + "step": 6180 + }, + { + "epoch": 11.539179104477611, + "grad_norm": 0.4428049325942993, + "learning_rate": 2.2658962761641232e-05, + "loss": 0.7706, + "num_input_tokens_seen": 1765248, + "step": 6185 + }, + { + "epoch": 11.548507462686567, + "grad_norm": 0.44100823998451233, + "learning_rate": 2.2618442089268926e-05, + "loss": 0.6807, + "num_input_tokens_seen": 1766656, + "step": 6190 + }, + { + "epoch": 11.557835820895523, + "grad_norm": 0.5056848526000977, + "learning_rate": 2.2577927729757458e-05, + "loss": 0.6217, + "num_input_tokens_seen": 1767936, + "step": 6195 + }, + { + "epoch": 11.567164179104477, + "grad_norm": 0.5112699270248413, + "learning_rate": 2.2537419790499323e-05, + "loss": 0.6691, + "num_input_tokens_seen": 1769216, + "step": 6200 + }, + { + "epoch": 11.576492537313433, + "grad_norm": 0.49766305088996887, + "learning_rate": 2.249691837887005e-05, + "loss": 0.693, + "num_input_tokens_seen": 1770656, + "step": 6205 + }, + { + "epoch": 11.585820895522389, + "grad_norm": 0.4704808294773102, + "learning_rate": 2.2456423602227835e-05, + "loss": 0.5865, + "num_input_tokens_seen": 1772096, + "step": 6210 + }, + { + "epoch": 11.595149253731343, + "grad_norm": 0.39703282713890076, + "learning_rate": 2.2415935567913286e-05, + "loss": 0.6224, + "num_input_tokens_seen": 1773472, + "step": 6215 + }, + { + "epoch": 11.604477611940299, + "grad_norm": 0.31121140718460083, + "learning_rate": 2.2375454383249154e-05, + "loss": 0.6797, + "num_input_tokens_seen": 1775104, + "step": 6220 + }, + { + "epoch": 11.613805970149254, + "grad_norm": 0.5745762586593628, + "learning_rate": 2.233498015554002e-05, + "loss": 0.8775, + "num_input_tokens_seen": 1776512, + "step": 6225 + }, + { + "epoch": 11.623134328358208, + "grad_norm": 0.4752705693244934, + "learning_rate": 2.229451299207203e-05, + "loss": 0.7079, + "num_input_tokens_seen": 1777920, + "step": 6230 + }, + { + "epoch": 11.632462686567164, + "grad_norm": 0.3659050166606903, + "learning_rate": 2.2254053000112597e-05, + "loss": 0.6841, + "num_input_tokens_seen": 1779488, + "step": 6235 + }, + { + "epoch": 11.64179104477612, + "grad_norm": 0.34801799058914185, + "learning_rate": 2.2213600286910134e-05, + "loss": 0.7833, + "num_input_tokens_seen": 1781120, + "step": 6240 + }, + { + "epoch": 11.651119402985074, + "grad_norm": 0.47012534737586975, + "learning_rate": 2.217315495969377e-05, + "loss": 0.7583, + "num_input_tokens_seen": 1782496, + "step": 6245 + }, + { + "epoch": 11.66044776119403, + "grad_norm": 0.39586329460144043, + "learning_rate": 2.2132717125673024e-05, + "loss": 0.4666, + "num_input_tokens_seen": 1784096, + "step": 6250 + }, + { + "epoch": 11.669776119402986, + "grad_norm": 0.699928879737854, + "learning_rate": 2.209228689203758e-05, + "loss": 0.9824, + "num_input_tokens_seen": 1785504, + "step": 6255 + }, + { + "epoch": 11.67910447761194, + "grad_norm": 0.541426956653595, + "learning_rate": 2.205186436595696e-05, + "loss": 0.608, + "num_input_tokens_seen": 1786784, + "step": 6260 + }, + { + "epoch": 11.688432835820896, + "grad_norm": 0.2842217683792114, + "learning_rate": 2.2011449654580266e-05, + "loss": 0.6628, + "num_input_tokens_seen": 1788160, + "step": 6265 + }, + { + "epoch": 11.697761194029852, + "grad_norm": 0.4538499712944031, + "learning_rate": 2.197104286503589e-05, + "loss": 0.7821, + "num_input_tokens_seen": 1789504, + "step": 6270 + }, + { + "epoch": 11.707089552238806, + "grad_norm": 0.6322186589241028, + "learning_rate": 2.1930644104431197e-05, + "loss": 0.5668, + "num_input_tokens_seen": 1790912, + "step": 6275 + }, + { + "epoch": 11.716417910447761, + "grad_norm": 0.7879039645195007, + "learning_rate": 2.1890253479852307e-05, + "loss": 0.9782, + "num_input_tokens_seen": 1792320, + "step": 6280 + }, + { + "epoch": 11.725746268656717, + "grad_norm": 0.5750066041946411, + "learning_rate": 2.184987109836374e-05, + "loss": 0.7463, + "num_input_tokens_seen": 1793568, + "step": 6285 + }, + { + "epoch": 11.735074626865671, + "grad_norm": 0.7576110363006592, + "learning_rate": 2.18094970670082e-05, + "loss": 0.7479, + "num_input_tokens_seen": 1794912, + "step": 6290 + }, + { + "epoch": 11.744402985074627, + "grad_norm": 0.4067259430885315, + "learning_rate": 2.176913149280622e-05, + "loss": 0.8237, + "num_input_tokens_seen": 1796256, + "step": 6295 + }, + { + "epoch": 11.753731343283581, + "grad_norm": 0.46755173802375793, + "learning_rate": 2.1728774482755942e-05, + "loss": 0.5091, + "num_input_tokens_seen": 1797856, + "step": 6300 + }, + { + "epoch": 11.763059701492537, + "grad_norm": 0.3662881851196289, + "learning_rate": 2.1688426143832803e-05, + "loss": 0.9356, + "num_input_tokens_seen": 1799232, + "step": 6305 + }, + { + "epoch": 11.772388059701493, + "grad_norm": 0.5010446310043335, + "learning_rate": 2.1648086582989242e-05, + "loss": 0.8543, + "num_input_tokens_seen": 1800640, + "step": 6310 + }, + { + "epoch": 11.781716417910447, + "grad_norm": 0.42489689588546753, + "learning_rate": 2.160775590715445e-05, + "loss": 0.8861, + "num_input_tokens_seen": 1802144, + "step": 6315 + }, + { + "epoch": 11.791044776119403, + "grad_norm": 0.47351089119911194, + "learning_rate": 2.1567434223234038e-05, + "loss": 0.6287, + "num_input_tokens_seen": 1803744, + "step": 6320 + }, + { + "epoch": 11.800373134328359, + "grad_norm": 0.5252626538276672, + "learning_rate": 2.152712163810981e-05, + "loss": 0.7744, + "num_input_tokens_seen": 1805088, + "step": 6325 + }, + { + "epoch": 11.809701492537313, + "grad_norm": 0.3913781940937042, + "learning_rate": 2.1486818258639445e-05, + "loss": 0.7016, + "num_input_tokens_seen": 1806464, + "step": 6330 + }, + { + "epoch": 11.819029850746269, + "grad_norm": 0.4217471778392792, + "learning_rate": 2.1446524191656205e-05, + "loss": 0.4255, + "num_input_tokens_seen": 1807808, + "step": 6335 + }, + { + "epoch": 11.828358208955224, + "grad_norm": 0.5851516723632812, + "learning_rate": 2.1406239543968687e-05, + "loss": 0.8773, + "num_input_tokens_seen": 1809216, + "step": 6340 + }, + { + "epoch": 11.837686567164178, + "grad_norm": 0.4443756341934204, + "learning_rate": 2.1365964422360497e-05, + "loss": 0.5874, + "num_input_tokens_seen": 1810560, + "step": 6345 + }, + { + "epoch": 11.847014925373134, + "grad_norm": 0.43719276785850525, + "learning_rate": 2.132569893359002e-05, + "loss": 0.5357, + "num_input_tokens_seen": 1812064, + "step": 6350 + }, + { + "epoch": 11.85634328358209, + "grad_norm": 0.35505589842796326, + "learning_rate": 2.1285443184390076e-05, + "loss": 0.5922, + "num_input_tokens_seen": 1813664, + "step": 6355 + }, + { + "epoch": 11.865671641791044, + "grad_norm": 0.34281590580940247, + "learning_rate": 2.1245197281467686e-05, + "loss": 0.713, + "num_input_tokens_seen": 1815264, + "step": 6360 + }, + { + "epoch": 11.875, + "grad_norm": 0.40556201338768005, + "learning_rate": 2.1204961331503787e-05, + "loss": 0.5017, + "num_input_tokens_seen": 1816832, + "step": 6365 + }, + { + "epoch": 11.884328358208956, + "grad_norm": 0.46798616647720337, + "learning_rate": 2.1164735441152882e-05, + "loss": 0.672, + "num_input_tokens_seen": 1818336, + "step": 6370 + }, + { + "epoch": 11.89365671641791, + "grad_norm": 0.4588167071342468, + "learning_rate": 2.1124519717042873e-05, + "loss": 0.6245, + "num_input_tokens_seen": 1819712, + "step": 6375 + }, + { + "epoch": 11.902985074626866, + "grad_norm": 0.37605220079421997, + "learning_rate": 2.108431426577466e-05, + "loss": 0.6694, + "num_input_tokens_seen": 1821120, + "step": 6380 + }, + { + "epoch": 11.912313432835822, + "grad_norm": 0.48783600330352783, + "learning_rate": 2.104411919392193e-05, + "loss": 0.6342, + "num_input_tokens_seen": 1822528, + "step": 6385 + }, + { + "epoch": 11.921641791044776, + "grad_norm": 0.4631548523902893, + "learning_rate": 2.1003934608030895e-05, + "loss": 0.4644, + "num_input_tokens_seen": 1824160, + "step": 6390 + }, + { + "epoch": 11.930970149253731, + "grad_norm": 0.6006349921226501, + "learning_rate": 2.0963760614619898e-05, + "loss": 0.5721, + "num_input_tokens_seen": 1825568, + "step": 6395 + }, + { + "epoch": 11.940298507462687, + "grad_norm": 0.335234135389328, + "learning_rate": 2.092359732017928e-05, + "loss": 0.6996, + "num_input_tokens_seen": 1827008, + "step": 6400 + }, + { + "epoch": 11.949626865671641, + "grad_norm": 0.8393757939338684, + "learning_rate": 2.0883444831170952e-05, + "loss": 0.8017, + "num_input_tokens_seen": 1828480, + "step": 6405 + }, + { + "epoch": 11.958955223880597, + "grad_norm": 0.5563313961029053, + "learning_rate": 2.0843303254028253e-05, + "loss": 0.8202, + "num_input_tokens_seen": 1829888, + "step": 6410 + }, + { + "epoch": 11.968283582089553, + "grad_norm": 0.45816513895988464, + "learning_rate": 2.0803172695155526e-05, + "loss": 0.8177, + "num_input_tokens_seen": 1831456, + "step": 6415 + }, + { + "epoch": 11.977611940298507, + "grad_norm": 0.2963179051876068, + "learning_rate": 2.076305326092796e-05, + "loss": 0.6493, + "num_input_tokens_seen": 1832736, + "step": 6420 + }, + { + "epoch": 11.986940298507463, + "grad_norm": 0.44608190655708313, + "learning_rate": 2.0722945057691252e-05, + "loss": 0.6175, + "num_input_tokens_seen": 1834208, + "step": 6425 + }, + { + "epoch": 11.996268656716419, + "grad_norm": 0.3543734848499298, + "learning_rate": 2.0682848191761296e-05, + "loss": 0.5549, + "num_input_tokens_seen": 1835584, + "step": 6430 + }, + { + "epoch": 12.0, + "eval_loss": 0.7097558379173279, + "eval_runtime": 2.4544, + "eval_samples_per_second": 96.97, + "eval_steps_per_second": 24.446, + "num_input_tokens_seen": 1835928, + "step": 6432 + }, + { + "epoch": 12.005597014925373, + "grad_norm": 0.5140202045440674, + "learning_rate": 2.0642762769423968e-05, + "loss": 0.902, + "num_input_tokens_seen": 1836728, + "step": 6435 + }, + { + "epoch": 12.014925373134329, + "grad_norm": 0.3134395182132721, + "learning_rate": 2.060268889693477e-05, + "loss": 0.699, + "num_input_tokens_seen": 1838008, + "step": 6440 + }, + { + "epoch": 12.024253731343284, + "grad_norm": 0.4919174611568451, + "learning_rate": 2.056262668051864e-05, + "loss": 0.9602, + "num_input_tokens_seen": 1839416, + "step": 6445 + }, + { + "epoch": 12.033582089552239, + "grad_norm": 0.610877513885498, + "learning_rate": 2.0522576226369592e-05, + "loss": 1.1926, + "num_input_tokens_seen": 1840760, + "step": 6450 + }, + { + "epoch": 12.042910447761194, + "grad_norm": 0.3318273425102234, + "learning_rate": 2.048253764065045e-05, + "loss": 0.6687, + "num_input_tokens_seen": 1842136, + "step": 6455 + }, + { + "epoch": 12.052238805970148, + "grad_norm": 0.45240479707717896, + "learning_rate": 2.04425110294926e-05, + "loss": 0.5342, + "num_input_tokens_seen": 1843640, + "step": 6460 + }, + { + "epoch": 12.061567164179104, + "grad_norm": 0.4665420353412628, + "learning_rate": 2.0402496498995667e-05, + "loss": 0.8758, + "num_input_tokens_seen": 1845048, + "step": 6465 + }, + { + "epoch": 12.07089552238806, + "grad_norm": 0.587179958820343, + "learning_rate": 2.0362494155227275e-05, + "loss": 0.5689, + "num_input_tokens_seen": 1846584, + "step": 6470 + }, + { + "epoch": 12.080223880597014, + "grad_norm": 0.5795404314994812, + "learning_rate": 2.0322504104222723e-05, + "loss": 0.8022, + "num_input_tokens_seen": 1847928, + "step": 6475 + }, + { + "epoch": 12.08955223880597, + "grad_norm": 0.4778423607349396, + "learning_rate": 2.028252645198474e-05, + "loss": 0.6233, + "num_input_tokens_seen": 1849208, + "step": 6480 + }, + { + "epoch": 12.098880597014926, + "grad_norm": 0.49149301648139954, + "learning_rate": 2.024256130448319e-05, + "loss": 0.5536, + "num_input_tokens_seen": 1850520, + "step": 6485 + }, + { + "epoch": 12.10820895522388, + "grad_norm": 0.5957432389259338, + "learning_rate": 2.0202608767654773e-05, + "loss": 0.5697, + "num_input_tokens_seen": 1852056, + "step": 6490 + }, + { + "epoch": 12.117537313432836, + "grad_norm": 0.5170521140098572, + "learning_rate": 2.016266894740279e-05, + "loss": 0.7984, + "num_input_tokens_seen": 1853496, + "step": 6495 + }, + { + "epoch": 12.126865671641792, + "grad_norm": 0.5206648707389832, + "learning_rate": 2.0122741949596797e-05, + "loss": 0.7266, + "num_input_tokens_seen": 1855096, + "step": 6500 + }, + { + "epoch": 12.136194029850746, + "grad_norm": 0.4649660885334015, + "learning_rate": 2.008282788007239e-05, + "loss": 0.7324, + "num_input_tokens_seen": 1856440, + "step": 6505 + }, + { + "epoch": 12.145522388059701, + "grad_norm": 0.3949867784976959, + "learning_rate": 2.0042926844630896e-05, + "loss": 0.7741, + "num_input_tokens_seen": 1858168, + "step": 6510 + }, + { + "epoch": 12.154850746268657, + "grad_norm": 0.433849573135376, + "learning_rate": 2.000303894903907e-05, + "loss": 0.4925, + "num_input_tokens_seen": 1859512, + "step": 6515 + }, + { + "epoch": 12.164179104477611, + "grad_norm": 0.6308825016021729, + "learning_rate": 1.9963164299028865e-05, + "loss": 0.4725, + "num_input_tokens_seen": 1860888, + "step": 6520 + }, + { + "epoch": 12.173507462686567, + "grad_norm": 0.3378816246986389, + "learning_rate": 1.992330300029709e-05, + "loss": 0.641, + "num_input_tokens_seen": 1862360, + "step": 6525 + }, + { + "epoch": 12.182835820895523, + "grad_norm": 0.3896324038505554, + "learning_rate": 1.988345515850519e-05, + "loss": 0.7959, + "num_input_tokens_seen": 1863672, + "step": 6530 + }, + { + "epoch": 12.192164179104477, + "grad_norm": 0.30465734004974365, + "learning_rate": 1.984362087927894e-05, + "loss": 0.8698, + "num_input_tokens_seen": 1865112, + "step": 6535 + }, + { + "epoch": 12.201492537313433, + "grad_norm": 0.41110512614250183, + "learning_rate": 1.9803800268208146e-05, + "loss": 0.8666, + "num_input_tokens_seen": 1866456, + "step": 6540 + }, + { + "epoch": 12.210820895522389, + "grad_norm": 0.366074800491333, + "learning_rate": 1.9763993430846395e-05, + "loss": 0.7163, + "num_input_tokens_seen": 1868024, + "step": 6545 + }, + { + "epoch": 12.220149253731343, + "grad_norm": 0.4653415381908417, + "learning_rate": 1.972420047271076e-05, + "loss": 0.6579, + "num_input_tokens_seen": 1869464, + "step": 6550 + }, + { + "epoch": 12.229477611940299, + "grad_norm": 0.25266626477241516, + "learning_rate": 1.9684421499281533e-05, + "loss": 0.6693, + "num_input_tokens_seen": 1870968, + "step": 6555 + }, + { + "epoch": 12.238805970149254, + "grad_norm": 0.46460574865341187, + "learning_rate": 1.964465661600192e-05, + "loss": 0.7301, + "num_input_tokens_seen": 1872312, + "step": 6560 + }, + { + "epoch": 12.248134328358208, + "grad_norm": 0.6310614347457886, + "learning_rate": 1.9604905928277782e-05, + "loss": 0.5882, + "num_input_tokens_seen": 1873816, + "step": 6565 + }, + { + "epoch": 12.257462686567164, + "grad_norm": 0.44004449248313904, + "learning_rate": 1.9565169541477388e-05, + "loss": 0.8647, + "num_input_tokens_seen": 1875192, + "step": 6570 + }, + { + "epoch": 12.26679104477612, + "grad_norm": 0.35160258412361145, + "learning_rate": 1.9525447560931028e-05, + "loss": 0.7601, + "num_input_tokens_seen": 1876536, + "step": 6575 + }, + { + "epoch": 12.276119402985074, + "grad_norm": 0.38567399978637695, + "learning_rate": 1.948574009193087e-05, + "loss": 0.6754, + "num_input_tokens_seen": 1878104, + "step": 6580 + }, + { + "epoch": 12.28544776119403, + "grad_norm": 0.48192933201789856, + "learning_rate": 1.944604723973058e-05, + "loss": 0.5403, + "num_input_tokens_seen": 1879352, + "step": 6585 + }, + { + "epoch": 12.294776119402986, + "grad_norm": 0.5226728916168213, + "learning_rate": 1.940636910954508e-05, + "loss": 0.6068, + "num_input_tokens_seen": 1880792, + "step": 6590 + }, + { + "epoch": 12.30410447761194, + "grad_norm": 0.329391211271286, + "learning_rate": 1.9366705806550308e-05, + "loss": 0.7867, + "num_input_tokens_seen": 1882264, + "step": 6595 + }, + { + "epoch": 12.313432835820896, + "grad_norm": 0.34140923619270325, + "learning_rate": 1.9327057435882835e-05, + "loss": 0.8209, + "num_input_tokens_seen": 1883544, + "step": 6600 + }, + { + "epoch": 12.322761194029852, + "grad_norm": 0.2912420928478241, + "learning_rate": 1.928742410263971e-05, + "loss": 0.7437, + "num_input_tokens_seen": 1885080, + "step": 6605 + }, + { + "epoch": 12.332089552238806, + "grad_norm": 0.7227346301078796, + "learning_rate": 1.9247805911878065e-05, + "loss": 0.6582, + "num_input_tokens_seen": 1886392, + "step": 6610 + }, + { + "epoch": 12.341417910447761, + "grad_norm": 0.4550590217113495, + "learning_rate": 1.920820296861496e-05, + "loss": 0.7116, + "num_input_tokens_seen": 1887800, + "step": 6615 + }, + { + "epoch": 12.350746268656717, + "grad_norm": 0.43829527497291565, + "learning_rate": 1.916861537782697e-05, + "loss": 0.7128, + "num_input_tokens_seen": 1889240, + "step": 6620 + }, + { + "epoch": 12.360074626865671, + "grad_norm": 0.5074513554573059, + "learning_rate": 1.9129043244450026e-05, + "loss": 0.8254, + "num_input_tokens_seen": 1890680, + "step": 6625 + }, + { + "epoch": 12.369402985074627, + "grad_norm": 0.3276227116584778, + "learning_rate": 1.9089486673379073e-05, + "loss": 0.6829, + "num_input_tokens_seen": 1892056, + "step": 6630 + }, + { + "epoch": 12.378731343283581, + "grad_norm": 0.44065117835998535, + "learning_rate": 1.904994576946777e-05, + "loss": 0.6337, + "num_input_tokens_seen": 1893496, + "step": 6635 + }, + { + "epoch": 12.388059701492537, + "grad_norm": 0.43937233090400696, + "learning_rate": 1.9010420637528308e-05, + "loss": 0.6692, + "num_input_tokens_seen": 1894904, + "step": 6640 + }, + { + "epoch": 12.397388059701493, + "grad_norm": 0.5081276893615723, + "learning_rate": 1.8970911382331006e-05, + "loss": 0.6629, + "num_input_tokens_seen": 1896376, + "step": 6645 + }, + { + "epoch": 12.406716417910447, + "grad_norm": 0.3650656044483185, + "learning_rate": 1.8931418108604148e-05, + "loss": 0.6651, + "num_input_tokens_seen": 1898104, + "step": 6650 + }, + { + "epoch": 12.416044776119403, + "grad_norm": 0.27774691581726074, + "learning_rate": 1.889194092103364e-05, + "loss": 0.5099, + "num_input_tokens_seen": 1899768, + "step": 6655 + }, + { + "epoch": 12.425373134328359, + "grad_norm": 0.3988930284976959, + "learning_rate": 1.8852479924262733e-05, + "loss": 0.6364, + "num_input_tokens_seen": 1901112, + "step": 6660 + }, + { + "epoch": 12.434701492537313, + "grad_norm": 0.6572892665863037, + "learning_rate": 1.8813035222891784e-05, + "loss": 0.7926, + "num_input_tokens_seen": 1902648, + "step": 6665 + }, + { + "epoch": 12.444029850746269, + "grad_norm": 0.4455818831920624, + "learning_rate": 1.8773606921477933e-05, + "loss": 0.8115, + "num_input_tokens_seen": 1904248, + "step": 6670 + }, + { + "epoch": 12.453358208955224, + "grad_norm": 0.3937157988548279, + "learning_rate": 1.873419512453487e-05, + "loss": 0.6087, + "num_input_tokens_seen": 1905752, + "step": 6675 + }, + { + "epoch": 12.462686567164178, + "grad_norm": 0.533672034740448, + "learning_rate": 1.869479993653252e-05, + "loss": 0.9004, + "num_input_tokens_seen": 1907160, + "step": 6680 + }, + { + "epoch": 12.472014925373134, + "grad_norm": 0.36333566904067993, + "learning_rate": 1.865542146189678e-05, + "loss": 0.7601, + "num_input_tokens_seen": 1908728, + "step": 6685 + }, + { + "epoch": 12.48134328358209, + "grad_norm": 0.3752221465110779, + "learning_rate": 1.861605980500927e-05, + "loss": 0.6766, + "num_input_tokens_seen": 1910264, + "step": 6690 + }, + { + "epoch": 12.490671641791044, + "grad_norm": 0.2916198968887329, + "learning_rate": 1.8576715070206992e-05, + "loss": 0.5588, + "num_input_tokens_seen": 1911768, + "step": 6695 + }, + { + "epoch": 12.5, + "grad_norm": 0.764884889125824, + "learning_rate": 1.853738736178213e-05, + "loss": 0.5978, + "num_input_tokens_seen": 1913304, + "step": 6700 + }, + { + "epoch": 12.509328358208956, + "grad_norm": 0.5469979643821716, + "learning_rate": 1.849807678398171e-05, + "loss": 0.6361, + "num_input_tokens_seen": 1914584, + "step": 6705 + }, + { + "epoch": 12.51865671641791, + "grad_norm": 0.5045479536056519, + "learning_rate": 1.845878344100736e-05, + "loss": 0.6449, + "num_input_tokens_seen": 1915896, + "step": 6710 + }, + { + "epoch": 12.527985074626866, + "grad_norm": 0.32327505946159363, + "learning_rate": 1.8419507437015022e-05, + "loss": 0.8722, + "num_input_tokens_seen": 1917272, + "step": 6715 + }, + { + "epoch": 12.537313432835822, + "grad_norm": 0.5720032453536987, + "learning_rate": 1.838024887611467e-05, + "loss": 0.7896, + "num_input_tokens_seen": 1919032, + "step": 6720 + }, + { + "epoch": 12.546641791044776, + "grad_norm": 0.41109827160835266, + "learning_rate": 1.8341007862370056e-05, + "loss": 0.4534, + "num_input_tokens_seen": 1920632, + "step": 6725 + }, + { + "epoch": 12.555970149253731, + "grad_norm": 0.38978853821754456, + "learning_rate": 1.830178449979841e-05, + "loss": 0.5518, + "num_input_tokens_seen": 1922072, + "step": 6730 + }, + { + "epoch": 12.565298507462687, + "grad_norm": 0.5245189070701599, + "learning_rate": 1.826257889237017e-05, + "loss": 0.8693, + "num_input_tokens_seen": 1923480, + "step": 6735 + }, + { + "epoch": 12.574626865671641, + "grad_norm": 0.706494927406311, + "learning_rate": 1.822339114400871e-05, + "loss": 0.8331, + "num_input_tokens_seen": 1924824, + "step": 6740 + }, + { + "epoch": 12.583955223880597, + "grad_norm": 0.5491077303886414, + "learning_rate": 1.818422135859008e-05, + "loss": 0.7846, + "num_input_tokens_seen": 1926296, + "step": 6745 + }, + { + "epoch": 12.593283582089553, + "grad_norm": 0.7595633268356323, + "learning_rate": 1.8145069639942697e-05, + "loss": 0.9618, + "num_input_tokens_seen": 1927864, + "step": 6750 + }, + { + "epoch": 12.602611940298507, + "grad_norm": 0.33538517355918884, + "learning_rate": 1.8105936091847093e-05, + "loss": 0.689, + "num_input_tokens_seen": 1929304, + "step": 6755 + }, + { + "epoch": 12.611940298507463, + "grad_norm": 0.3504796028137207, + "learning_rate": 1.806682081803564e-05, + "loss": 0.6998, + "num_input_tokens_seen": 1930840, + "step": 6760 + }, + { + "epoch": 12.621268656716419, + "grad_norm": 0.41527459025382996, + "learning_rate": 1.8027723922192264e-05, + "loss": 0.6664, + "num_input_tokens_seen": 1932216, + "step": 6765 + }, + { + "epoch": 12.630597014925373, + "grad_norm": 0.4545871913433075, + "learning_rate": 1.798864550795218e-05, + "loss": 0.5262, + "num_input_tokens_seen": 1933752, + "step": 6770 + }, + { + "epoch": 12.639925373134329, + "grad_norm": 0.43095430731773376, + "learning_rate": 1.7949585678901614e-05, + "loss": 0.6603, + "num_input_tokens_seen": 1935256, + "step": 6775 + }, + { + "epoch": 12.649253731343283, + "grad_norm": 0.27440059185028076, + "learning_rate": 1.7910544538577522e-05, + "loss": 0.5656, + "num_input_tokens_seen": 1936568, + "step": 6780 + }, + { + "epoch": 12.658582089552239, + "grad_norm": 0.5431836843490601, + "learning_rate": 1.787152219046733e-05, + "loss": 0.7971, + "num_input_tokens_seen": 1937848, + "step": 6785 + }, + { + "epoch": 12.667910447761194, + "grad_norm": 0.4308947026729584, + "learning_rate": 1.783251873800863e-05, + "loss": 0.6134, + "num_input_tokens_seen": 1939160, + "step": 6790 + }, + { + "epoch": 12.677238805970148, + "grad_norm": 0.261123925447464, + "learning_rate": 1.779353428458896e-05, + "loss": 0.4944, + "num_input_tokens_seen": 1940568, + "step": 6795 + }, + { + "epoch": 12.686567164179104, + "grad_norm": 0.5042296051979065, + "learning_rate": 1.775456893354547e-05, + "loss": 0.7118, + "num_input_tokens_seen": 1941848, + "step": 6800 + }, + { + "epoch": 12.69589552238806, + "grad_norm": 0.35568058490753174, + "learning_rate": 1.7715622788164683e-05, + "loss": 0.5791, + "num_input_tokens_seen": 1943448, + "step": 6805 + }, + { + "epoch": 12.705223880597014, + "grad_norm": 0.44231343269348145, + "learning_rate": 1.767669595168223e-05, + "loss": 0.7661, + "num_input_tokens_seen": 1944824, + "step": 6810 + }, + { + "epoch": 12.71455223880597, + "grad_norm": 0.5539527535438538, + "learning_rate": 1.7637788527282522e-05, + "loss": 0.686, + "num_input_tokens_seen": 1946136, + "step": 6815 + }, + { + "epoch": 12.723880597014926, + "grad_norm": 0.5302578210830688, + "learning_rate": 1.7598900618098548e-05, + "loss": 0.9863, + "num_input_tokens_seen": 1947416, + "step": 6820 + }, + { + "epoch": 12.73320895522388, + "grad_norm": 0.48223942518234253, + "learning_rate": 1.7560032327211544e-05, + "loss": 0.7998, + "num_input_tokens_seen": 1948856, + "step": 6825 + }, + { + "epoch": 12.742537313432836, + "grad_norm": 0.4571443200111389, + "learning_rate": 1.7521183757650762e-05, + "loss": 0.6504, + "num_input_tokens_seen": 1950392, + "step": 6830 + }, + { + "epoch": 12.751865671641792, + "grad_norm": 0.5913133025169373, + "learning_rate": 1.7482355012393177e-05, + "loss": 0.7161, + "num_input_tokens_seen": 1951800, + "step": 6835 + }, + { + "epoch": 12.761194029850746, + "grad_norm": 0.7472345232963562, + "learning_rate": 1.7443546194363188e-05, + "loss": 0.5924, + "num_input_tokens_seen": 1953144, + "step": 6840 + }, + { + "epoch": 12.770522388059701, + "grad_norm": 0.5724983215332031, + "learning_rate": 1.740475740643242e-05, + "loss": 0.739, + "num_input_tokens_seen": 1954456, + "step": 6845 + }, + { + "epoch": 12.779850746268657, + "grad_norm": 0.4338374435901642, + "learning_rate": 1.7365988751419343e-05, + "loss": 0.6591, + "num_input_tokens_seen": 1955864, + "step": 6850 + }, + { + "epoch": 12.789179104477611, + "grad_norm": 0.4927447736263275, + "learning_rate": 1.7327240332089128e-05, + "loss": 0.7264, + "num_input_tokens_seen": 1957272, + "step": 6855 + }, + { + "epoch": 12.798507462686567, + "grad_norm": 0.4093066453933716, + "learning_rate": 1.7288512251153243e-05, + "loss": 0.7766, + "num_input_tokens_seen": 1958776, + "step": 6860 + }, + { + "epoch": 12.807835820895523, + "grad_norm": 0.32650360465049744, + "learning_rate": 1.7249804611269288e-05, + "loss": 0.5859, + "num_input_tokens_seen": 1960216, + "step": 6865 + }, + { + "epoch": 12.817164179104477, + "grad_norm": 0.5671585202217102, + "learning_rate": 1.7211117515040676e-05, + "loss": 0.6694, + "num_input_tokens_seen": 1961592, + "step": 6870 + }, + { + "epoch": 12.826492537313433, + "grad_norm": 0.6301414370536804, + "learning_rate": 1.717245106501632e-05, + "loss": 0.7363, + "num_input_tokens_seen": 1962904, + "step": 6875 + }, + { + "epoch": 12.835820895522389, + "grad_norm": 0.5519711375236511, + "learning_rate": 1.7133805363690478e-05, + "loss": 0.6853, + "num_input_tokens_seen": 1964248, + "step": 6880 + }, + { + "epoch": 12.845149253731343, + "grad_norm": 0.42616549134254456, + "learning_rate": 1.7095180513502334e-05, + "loss": 0.6738, + "num_input_tokens_seen": 1965592, + "step": 6885 + }, + { + "epoch": 12.854477611940299, + "grad_norm": 0.4124564826488495, + "learning_rate": 1.705657661683586e-05, + "loss": 0.5402, + "num_input_tokens_seen": 1966904, + "step": 6890 + }, + { + "epoch": 12.863805970149254, + "grad_norm": 0.5619946718215942, + "learning_rate": 1.701799377601946e-05, + "loss": 0.547, + "num_input_tokens_seen": 1968280, + "step": 6895 + }, + { + "epoch": 12.873134328358208, + "grad_norm": 0.4494348168373108, + "learning_rate": 1.697943209332572e-05, + "loss": 0.7172, + "num_input_tokens_seen": 1969816, + "step": 6900 + }, + { + "epoch": 12.882462686567164, + "grad_norm": 0.5434309840202332, + "learning_rate": 1.694089167097116e-05, + "loss": 0.6153, + "num_input_tokens_seen": 1971320, + "step": 6905 + }, + { + "epoch": 12.89179104477612, + "grad_norm": 0.4547886252403259, + "learning_rate": 1.6902372611115917e-05, + "loss": 0.6032, + "num_input_tokens_seen": 1972824, + "step": 6910 + }, + { + "epoch": 12.901119402985074, + "grad_norm": 0.4960781931877136, + "learning_rate": 1.686387501586354e-05, + "loss": 0.8126, + "num_input_tokens_seen": 1974392, + "step": 6915 + }, + { + "epoch": 12.91044776119403, + "grad_norm": 0.4702379107475281, + "learning_rate": 1.6825398987260642e-05, + "loss": 0.7643, + "num_input_tokens_seen": 1975736, + "step": 6920 + }, + { + "epoch": 12.919776119402986, + "grad_norm": 0.6422269940376282, + "learning_rate": 1.678694462729669e-05, + "loss": 0.5751, + "num_input_tokens_seen": 1977240, + "step": 6925 + }, + { + "epoch": 12.92910447761194, + "grad_norm": 0.3540300130844116, + "learning_rate": 1.6748512037903725e-05, + "loss": 0.5676, + "num_input_tokens_seen": 1978808, + "step": 6930 + }, + { + "epoch": 12.938432835820896, + "grad_norm": 0.5464089512825012, + "learning_rate": 1.671010132095604e-05, + "loss": 0.8857, + "num_input_tokens_seen": 1980120, + "step": 6935 + }, + { + "epoch": 12.947761194029852, + "grad_norm": 0.709486722946167, + "learning_rate": 1.6671712578269997e-05, + "loss": 0.6987, + "num_input_tokens_seen": 1981496, + "step": 6940 + }, + { + "epoch": 12.957089552238806, + "grad_norm": 0.6778075695037842, + "learning_rate": 1.663334591160368e-05, + "loss": 0.6405, + "num_input_tokens_seen": 1982936, + "step": 6945 + }, + { + "epoch": 12.966417910447761, + "grad_norm": 0.7117565870285034, + "learning_rate": 1.659500142265666e-05, + "loss": 0.5897, + "num_input_tokens_seen": 1984472, + "step": 6950 + }, + { + "epoch": 12.975746268656717, + "grad_norm": 0.8322007060050964, + "learning_rate": 1.655667921306973e-05, + "loss": 0.9419, + "num_input_tokens_seen": 1985752, + "step": 6955 + }, + { + "epoch": 12.985074626865671, + "grad_norm": 0.5429271459579468, + "learning_rate": 1.6518379384424616e-05, + "loss": 0.5908, + "num_input_tokens_seen": 1987128, + "step": 6960 + }, + { + "epoch": 12.994402985074627, + "grad_norm": 0.7510144710540771, + "learning_rate": 1.6480102038243735e-05, + "loss": 0.7974, + "num_input_tokens_seen": 1988504, + "step": 6965 + }, + { + "epoch": 13.0, + "eval_loss": 0.7042965292930603, + "eval_runtime": 2.5485, + "eval_samples_per_second": 93.387, + "eval_steps_per_second": 23.543, + "num_input_tokens_seen": 1989136, + "step": 6968 + }, + { + "epoch": 13.003731343283581, + "grad_norm": 0.35301828384399414, + "learning_rate": 1.644184727598988e-05, + "loss": 0.5069, + "num_input_tokens_seen": 1989712, + "step": 6970 + }, + { + "epoch": 13.013059701492537, + "grad_norm": 0.4964786767959595, + "learning_rate": 1.640361519906602e-05, + "loss": 1.0385, + "num_input_tokens_seen": 1991120, + "step": 6975 + }, + { + "epoch": 13.022388059701493, + "grad_norm": 0.21495074033737183, + "learning_rate": 1.6365405908814946e-05, + "loss": 0.7186, + "num_input_tokens_seen": 1992688, + "step": 6980 + }, + { + "epoch": 13.031716417910447, + "grad_norm": 0.6090225577354431, + "learning_rate": 1.6327219506519083e-05, + "loss": 0.8841, + "num_input_tokens_seen": 1994032, + "step": 6985 + }, + { + "epoch": 13.041044776119403, + "grad_norm": 0.5328956842422485, + "learning_rate": 1.6289056093400178e-05, + "loss": 0.8753, + "num_input_tokens_seen": 1995472, + "step": 6990 + }, + { + "epoch": 13.050373134328359, + "grad_norm": 0.5221139192581177, + "learning_rate": 1.6250915770619023e-05, + "loss": 0.5648, + "num_input_tokens_seen": 1996848, + "step": 6995 + }, + { + "epoch": 13.059701492537313, + "grad_norm": 0.5423232913017273, + "learning_rate": 1.6212798639275233e-05, + "loss": 1.0481, + "num_input_tokens_seen": 1998128, + "step": 7000 + }, + { + "epoch": 13.069029850746269, + "grad_norm": 0.3626515865325928, + "learning_rate": 1.617470480040692e-05, + "loss": 0.669, + "num_input_tokens_seen": 1999472, + "step": 7005 + }, + { + "epoch": 13.078358208955224, + "grad_norm": 0.34185829758644104, + "learning_rate": 1.613663435499047e-05, + "loss": 0.6796, + "num_input_tokens_seen": 2000944, + "step": 7010 + }, + { + "epoch": 13.087686567164178, + "grad_norm": 0.3611834943294525, + "learning_rate": 1.609858740394026e-05, + "loss": 0.9465, + "num_input_tokens_seen": 2002448, + "step": 7015 + }, + { + "epoch": 13.097014925373134, + "grad_norm": 0.4902392625808716, + "learning_rate": 1.6060564048108383e-05, + "loss": 0.67, + "num_input_tokens_seen": 2003888, + "step": 7020 + }, + { + "epoch": 13.10634328358209, + "grad_norm": 0.46479332447052, + "learning_rate": 1.6022564388284392e-05, + "loss": 0.6812, + "num_input_tokens_seen": 2005328, + "step": 7025 + }, + { + "epoch": 13.115671641791044, + "grad_norm": 0.6588858366012573, + "learning_rate": 1.598458852519502e-05, + "loss": 0.6754, + "num_input_tokens_seen": 2006608, + "step": 7030 + }, + { + "epoch": 13.125, + "grad_norm": 0.6324683427810669, + "learning_rate": 1.594663655950394e-05, + "loss": 0.637, + "num_input_tokens_seen": 2007984, + "step": 7035 + }, + { + "epoch": 13.134328358208956, + "grad_norm": 0.6209911108016968, + "learning_rate": 1.590870859181146e-05, + "loss": 0.6033, + "num_input_tokens_seen": 2009392, + "step": 7040 + }, + { + "epoch": 13.14365671641791, + "grad_norm": 0.3238638937473297, + "learning_rate": 1.5870804722654275e-05, + "loss": 0.7395, + "num_input_tokens_seen": 2010672, + "step": 7045 + }, + { + "epoch": 13.152985074626866, + "grad_norm": 0.45025041699409485, + "learning_rate": 1.5832925052505235e-05, + "loss": 0.5959, + "num_input_tokens_seen": 2012176, + "step": 7050 + }, + { + "epoch": 13.162313432835822, + "grad_norm": 0.42980897426605225, + "learning_rate": 1.579506968177299e-05, + "loss": 0.5887, + "num_input_tokens_seen": 2013584, + "step": 7055 + }, + { + "epoch": 13.171641791044776, + "grad_norm": 0.45898208022117615, + "learning_rate": 1.575723871080184e-05, + "loss": 0.5068, + "num_input_tokens_seen": 2014896, + "step": 7060 + }, + { + "epoch": 13.180970149253731, + "grad_norm": 0.4348423182964325, + "learning_rate": 1.5719432239871347e-05, + "loss": 0.592, + "num_input_tokens_seen": 2016304, + "step": 7065 + }, + { + "epoch": 13.190298507462687, + "grad_norm": 0.3794540762901306, + "learning_rate": 1.5681650369196165e-05, + "loss": 0.6415, + "num_input_tokens_seen": 2018032, + "step": 7070 + }, + { + "epoch": 13.199626865671641, + "grad_norm": 0.3234213590621948, + "learning_rate": 1.5643893198925764e-05, + "loss": 0.6437, + "num_input_tokens_seen": 2019536, + "step": 7075 + }, + { + "epoch": 13.208955223880597, + "grad_norm": 0.5665496587753296, + "learning_rate": 1.5606160829144066e-05, + "loss": 0.7662, + "num_input_tokens_seen": 2020784, + "step": 7080 + }, + { + "epoch": 13.218283582089553, + "grad_norm": 0.44309860467910767, + "learning_rate": 1.5568453359869334e-05, + "loss": 0.5562, + "num_input_tokens_seen": 2022288, + "step": 7085 + }, + { + "epoch": 13.227611940298507, + "grad_norm": 0.2326476275920868, + "learning_rate": 1.5530770891053763e-05, + "loss": 0.473, + "num_input_tokens_seen": 2023952, + "step": 7090 + }, + { + "epoch": 13.236940298507463, + "grad_norm": 0.5203782320022583, + "learning_rate": 1.5493113522583334e-05, + "loss": 0.8687, + "num_input_tokens_seen": 2025392, + "step": 7095 + }, + { + "epoch": 13.246268656716419, + "grad_norm": 0.5410526990890503, + "learning_rate": 1.5455481354277435e-05, + "loss": 0.5748, + "num_input_tokens_seen": 2026992, + "step": 7100 + }, + { + "epoch": 13.255597014925373, + "grad_norm": 0.6034708023071289, + "learning_rate": 1.5417874485888706e-05, + "loss": 0.8927, + "num_input_tokens_seen": 2028400, + "step": 7105 + }, + { + "epoch": 13.264925373134329, + "grad_norm": 0.26713141798973083, + "learning_rate": 1.53802930171027e-05, + "loss": 0.7319, + "num_input_tokens_seen": 2029872, + "step": 7110 + }, + { + "epoch": 13.274253731343283, + "grad_norm": 0.56208336353302, + "learning_rate": 1.5342737047537643e-05, + "loss": 0.8414, + "num_input_tokens_seen": 2031408, + "step": 7115 + }, + { + "epoch": 13.283582089552239, + "grad_norm": 0.46734827756881714, + "learning_rate": 1.5305206676744187e-05, + "loss": 0.7203, + "num_input_tokens_seen": 2032688, + "step": 7120 + }, + { + "epoch": 13.292910447761194, + "grad_norm": 0.40145453810691833, + "learning_rate": 1.5267702004205083e-05, + "loss": 0.867, + "num_input_tokens_seen": 2034192, + "step": 7125 + }, + { + "epoch": 13.302238805970148, + "grad_norm": 0.32588744163513184, + "learning_rate": 1.5230223129335019e-05, + "loss": 0.6854, + "num_input_tokens_seen": 2035504, + "step": 7130 + }, + { + "epoch": 13.311567164179104, + "grad_norm": 0.5111632347106934, + "learning_rate": 1.5192770151480278e-05, + "loss": 0.6535, + "num_input_tokens_seen": 2036944, + "step": 7135 + }, + { + "epoch": 13.32089552238806, + "grad_norm": 0.48555824160575867, + "learning_rate": 1.5155343169918485e-05, + "loss": 0.8008, + "num_input_tokens_seen": 2038384, + "step": 7140 + }, + { + "epoch": 13.330223880597014, + "grad_norm": 0.756938099861145, + "learning_rate": 1.5117942283858369e-05, + "loss": 0.7971, + "num_input_tokens_seen": 2039696, + "step": 7145 + }, + { + "epoch": 13.33955223880597, + "grad_norm": 0.8027671575546265, + "learning_rate": 1.5080567592439479e-05, + "loss": 0.8899, + "num_input_tokens_seen": 2041104, + "step": 7150 + }, + { + "epoch": 13.348880597014926, + "grad_norm": 0.26233938336372375, + "learning_rate": 1.5043219194731945e-05, + "loss": 0.7341, + "num_input_tokens_seen": 2042576, + "step": 7155 + }, + { + "epoch": 13.35820895522388, + "grad_norm": 0.5465114116668701, + "learning_rate": 1.5005897189736173e-05, + "loss": 0.5678, + "num_input_tokens_seen": 2044176, + "step": 7160 + }, + { + "epoch": 13.367537313432836, + "grad_norm": 0.422993928194046, + "learning_rate": 1.4968601676382634e-05, + "loss": 0.6941, + "num_input_tokens_seen": 2045552, + "step": 7165 + }, + { + "epoch": 13.376865671641792, + "grad_norm": 0.6368779540061951, + "learning_rate": 1.4931332753531574e-05, + "loss": 0.7818, + "num_input_tokens_seen": 2047024, + "step": 7170 + }, + { + "epoch": 13.386194029850746, + "grad_norm": 0.524102509021759, + "learning_rate": 1.4894090519972731e-05, + "loss": 0.7619, + "num_input_tokens_seen": 2048432, + "step": 7175 + }, + { + "epoch": 13.395522388059701, + "grad_norm": 0.4211618900299072, + "learning_rate": 1.4856875074425131e-05, + "loss": 0.6177, + "num_input_tokens_seen": 2049712, + "step": 7180 + }, + { + "epoch": 13.404850746268657, + "grad_norm": 0.3790084719657898, + "learning_rate": 1.4819686515536762e-05, + "loss": 0.6033, + "num_input_tokens_seen": 2051088, + "step": 7185 + }, + { + "epoch": 13.414179104477611, + "grad_norm": 0.27184632420539856, + "learning_rate": 1.4782524941884366e-05, + "loss": 0.9284, + "num_input_tokens_seen": 2052560, + "step": 7190 + }, + { + "epoch": 13.423507462686567, + "grad_norm": 0.4439396560192108, + "learning_rate": 1.4745390451973146e-05, + "loss": 0.7874, + "num_input_tokens_seen": 2054032, + "step": 7195 + }, + { + "epoch": 13.432835820895523, + "grad_norm": 0.6156230568885803, + "learning_rate": 1.4708283144236498e-05, + "loss": 0.7782, + "num_input_tokens_seen": 2055440, + "step": 7200 + }, + { + "epoch": 13.442164179104477, + "grad_norm": 0.4678042531013489, + "learning_rate": 1.4671203117035799e-05, + "loss": 0.5654, + "num_input_tokens_seen": 2056688, + "step": 7205 + }, + { + "epoch": 13.451492537313433, + "grad_norm": 0.3958576023578644, + "learning_rate": 1.4634150468660073e-05, + "loss": 0.4222, + "num_input_tokens_seen": 2058096, + "step": 7210 + }, + { + "epoch": 13.460820895522389, + "grad_norm": 0.4917243421077728, + "learning_rate": 1.4597125297325811e-05, + "loss": 0.6162, + "num_input_tokens_seen": 2059504, + "step": 7215 + }, + { + "epoch": 13.470149253731343, + "grad_norm": 0.2976827025413513, + "learning_rate": 1.4560127701176635e-05, + "loss": 0.6077, + "num_input_tokens_seen": 2060848, + "step": 7220 + }, + { + "epoch": 13.479477611940299, + "grad_norm": 0.4831414818763733, + "learning_rate": 1.452315777828308e-05, + "loss": 0.6284, + "num_input_tokens_seen": 2062288, + "step": 7225 + }, + { + "epoch": 13.488805970149254, + "grad_norm": 0.7448445558547974, + "learning_rate": 1.4486215626642364e-05, + "loss": 0.6042, + "num_input_tokens_seen": 2063696, + "step": 7230 + }, + { + "epoch": 13.498134328358208, + "grad_norm": 0.3919883072376251, + "learning_rate": 1.4449301344178037e-05, + "loss": 0.6723, + "num_input_tokens_seen": 2065168, + "step": 7235 + }, + { + "epoch": 13.507462686567164, + "grad_norm": 0.6802276968955994, + "learning_rate": 1.4412415028739806e-05, + "loss": 0.735, + "num_input_tokens_seen": 2066608, + "step": 7240 + }, + { + "epoch": 13.51679104477612, + "grad_norm": 0.30533722043037415, + "learning_rate": 1.4375556778103227e-05, + "loss": 0.6191, + "num_input_tokens_seen": 2068016, + "step": 7245 + }, + { + "epoch": 13.526119402985074, + "grad_norm": 0.5054107308387756, + "learning_rate": 1.4338726689969495e-05, + "loss": 0.7354, + "num_input_tokens_seen": 2069488, + "step": 7250 + }, + { + "epoch": 13.53544776119403, + "grad_norm": 0.42571523785591125, + "learning_rate": 1.4301924861965122e-05, + "loss": 0.6436, + "num_input_tokens_seen": 2070896, + "step": 7255 + }, + { + "epoch": 13.544776119402986, + "grad_norm": 0.4726997911930084, + "learning_rate": 1.4265151391641719e-05, + "loss": 0.7374, + "num_input_tokens_seen": 2072464, + "step": 7260 + }, + { + "epoch": 13.55410447761194, + "grad_norm": 0.5709852576255798, + "learning_rate": 1.4228406376475742e-05, + "loss": 0.7379, + "num_input_tokens_seen": 2074032, + "step": 7265 + }, + { + "epoch": 13.563432835820896, + "grad_norm": 0.5252785086631775, + "learning_rate": 1.4191689913868206e-05, + "loss": 0.7267, + "num_input_tokens_seen": 2075440, + "step": 7270 + }, + { + "epoch": 13.572761194029852, + "grad_norm": 0.4794166386127472, + "learning_rate": 1.4155002101144443e-05, + "loss": 0.8959, + "num_input_tokens_seen": 2076720, + "step": 7275 + }, + { + "epoch": 13.582089552238806, + "grad_norm": 0.7892388701438904, + "learning_rate": 1.4118343035553836e-05, + "loss": 0.646, + "num_input_tokens_seen": 2078192, + "step": 7280 + }, + { + "epoch": 13.591417910447761, + "grad_norm": 0.33416256308555603, + "learning_rate": 1.4081712814269593e-05, + "loss": 0.6261, + "num_input_tokens_seen": 2079536, + "step": 7285 + }, + { + "epoch": 13.600746268656717, + "grad_norm": 0.5855578184127808, + "learning_rate": 1.4045111534388433e-05, + "loss": 0.7993, + "num_input_tokens_seen": 2080944, + "step": 7290 + }, + { + "epoch": 13.610074626865671, + "grad_norm": 0.49772247672080994, + "learning_rate": 1.4008539292930368e-05, + "loss": 0.7716, + "num_input_tokens_seen": 2082288, + "step": 7295 + }, + { + "epoch": 13.619402985074627, + "grad_norm": 0.44283321499824524, + "learning_rate": 1.3971996186838451e-05, + "loss": 0.6178, + "num_input_tokens_seen": 2083632, + "step": 7300 + }, + { + "epoch": 13.628731343283581, + "grad_norm": 0.40576690435409546, + "learning_rate": 1.3935482312978492e-05, + "loss": 0.6197, + "num_input_tokens_seen": 2085168, + "step": 7305 + }, + { + "epoch": 13.638059701492537, + "grad_norm": 0.6459487676620483, + "learning_rate": 1.389899776813881e-05, + "loss": 0.8838, + "num_input_tokens_seen": 2086640, + "step": 7310 + }, + { + "epoch": 13.647388059701493, + "grad_norm": 0.39349034428596497, + "learning_rate": 1.3862542649030002e-05, + "loss": 0.6301, + "num_input_tokens_seen": 2088240, + "step": 7315 + }, + { + "epoch": 13.656716417910447, + "grad_norm": 0.4369026720523834, + "learning_rate": 1.3826117052284646e-05, + "loss": 0.6683, + "num_input_tokens_seen": 2089712, + "step": 7320 + }, + { + "epoch": 13.666044776119403, + "grad_norm": 0.410055935382843, + "learning_rate": 1.3789721074457063e-05, + "loss": 0.5782, + "num_input_tokens_seen": 2091056, + "step": 7325 + }, + { + "epoch": 13.675373134328359, + "grad_norm": 0.40338942408561707, + "learning_rate": 1.3753354812023067e-05, + "loss": 0.6513, + "num_input_tokens_seen": 2092624, + "step": 7330 + }, + { + "epoch": 13.684701492537313, + "grad_norm": 0.6772986054420471, + "learning_rate": 1.3717018361379719e-05, + "loss": 0.7331, + "num_input_tokens_seen": 2094256, + "step": 7335 + }, + { + "epoch": 13.694029850746269, + "grad_norm": 0.42802876234054565, + "learning_rate": 1.3680711818845039e-05, + "loss": 0.7159, + "num_input_tokens_seen": 2095664, + "step": 7340 + }, + { + "epoch": 13.703358208955224, + "grad_norm": 0.4825426936149597, + "learning_rate": 1.3644435280657764e-05, + "loss": 0.5797, + "num_input_tokens_seen": 2096976, + "step": 7345 + }, + { + "epoch": 13.712686567164178, + "grad_norm": 0.32418757677078247, + "learning_rate": 1.3608188842977127e-05, + "loss": 0.3717, + "num_input_tokens_seen": 2098416, + "step": 7350 + }, + { + "epoch": 13.722014925373134, + "grad_norm": 0.40364202857017517, + "learning_rate": 1.3571972601882544e-05, + "loss": 0.5511, + "num_input_tokens_seen": 2099888, + "step": 7355 + }, + { + "epoch": 13.73134328358209, + "grad_norm": 0.5198777914047241, + "learning_rate": 1.3535786653373397e-05, + "loss": 0.7494, + "num_input_tokens_seen": 2101200, + "step": 7360 + }, + { + "epoch": 13.740671641791044, + "grad_norm": 0.5332244038581848, + "learning_rate": 1.3499631093368764e-05, + "loss": 0.7062, + "num_input_tokens_seen": 2102448, + "step": 7365 + }, + { + "epoch": 13.75, + "grad_norm": 0.45164352655410767, + "learning_rate": 1.3463506017707197e-05, + "loss": 0.6917, + "num_input_tokens_seen": 2104048, + "step": 7370 + }, + { + "epoch": 13.759328358208956, + "grad_norm": 0.4053598940372467, + "learning_rate": 1.3427411522146416e-05, + "loss": 0.5581, + "num_input_tokens_seen": 2105648, + "step": 7375 + }, + { + "epoch": 13.76865671641791, + "grad_norm": 0.43489593267440796, + "learning_rate": 1.3391347702363078e-05, + "loss": 0.6633, + "num_input_tokens_seen": 2107152, + "step": 7380 + }, + { + "epoch": 13.777985074626866, + "grad_norm": 0.41336676478385925, + "learning_rate": 1.3355314653952555e-05, + "loss": 0.7548, + "num_input_tokens_seen": 2108688, + "step": 7385 + }, + { + "epoch": 13.787313432835822, + "grad_norm": 0.49324342608451843, + "learning_rate": 1.3319312472428636e-05, + "loss": 0.5988, + "num_input_tokens_seen": 2110320, + "step": 7390 + }, + { + "epoch": 13.796641791044776, + "grad_norm": 0.12578782439231873, + "learning_rate": 1.328334125322328e-05, + "loss": 0.6366, + "num_input_tokens_seen": 2111760, + "step": 7395 + }, + { + "epoch": 13.805970149253731, + "grad_norm": 0.4318375289440155, + "learning_rate": 1.3247401091686379e-05, + "loss": 0.8985, + "num_input_tokens_seen": 2113392, + "step": 7400 + }, + { + "epoch": 13.815298507462687, + "grad_norm": 0.612823486328125, + "learning_rate": 1.3211492083085511e-05, + "loss": 0.7545, + "num_input_tokens_seen": 2114736, + "step": 7405 + }, + { + "epoch": 13.824626865671641, + "grad_norm": 0.5589946508407593, + "learning_rate": 1.317561432260569e-05, + "loss": 0.6401, + "num_input_tokens_seen": 2116368, + "step": 7410 + }, + { + "epoch": 13.833955223880597, + "grad_norm": 0.42385515570640564, + "learning_rate": 1.3139767905349044e-05, + "loss": 0.69, + "num_input_tokens_seen": 2117680, + "step": 7415 + }, + { + "epoch": 13.843283582089553, + "grad_norm": 0.37731748819351196, + "learning_rate": 1.3103952926334678e-05, + "loss": 0.8334, + "num_input_tokens_seen": 2118992, + "step": 7420 + }, + { + "epoch": 13.852611940298507, + "grad_norm": 0.44106295704841614, + "learning_rate": 1.3068169480498333e-05, + "loss": 1.0066, + "num_input_tokens_seen": 2120304, + "step": 7425 + }, + { + "epoch": 13.861940298507463, + "grad_norm": 0.4479944407939911, + "learning_rate": 1.3032417662692153e-05, + "loss": 0.6476, + "num_input_tokens_seen": 2121744, + "step": 7430 + }, + { + "epoch": 13.871268656716419, + "grad_norm": 0.4414093792438507, + "learning_rate": 1.2996697567684485e-05, + "loss": 0.6199, + "num_input_tokens_seen": 2123184, + "step": 7435 + }, + { + "epoch": 13.880597014925373, + "grad_norm": 0.3590048551559448, + "learning_rate": 1.2961009290159549e-05, + "loss": 0.7323, + "num_input_tokens_seen": 2124720, + "step": 7440 + }, + { + "epoch": 13.889925373134329, + "grad_norm": 0.5442506670951843, + "learning_rate": 1.292535292471726e-05, + "loss": 0.5165, + "num_input_tokens_seen": 2126288, + "step": 7445 + }, + { + "epoch": 13.899253731343283, + "grad_norm": 0.4895356297492981, + "learning_rate": 1.2889728565872888e-05, + "loss": 0.6472, + "num_input_tokens_seen": 2127728, + "step": 7450 + }, + { + "epoch": 13.908582089552239, + "grad_norm": 0.3468961715698242, + "learning_rate": 1.2854136308056933e-05, + "loss": 0.5175, + "num_input_tokens_seen": 2129200, + "step": 7455 + }, + { + "epoch": 13.917910447761194, + "grad_norm": 0.406730592250824, + "learning_rate": 1.281857624561474e-05, + "loss": 0.5714, + "num_input_tokens_seen": 2130608, + "step": 7460 + }, + { + "epoch": 13.927238805970148, + "grad_norm": 0.47234898805618286, + "learning_rate": 1.2783048472806363e-05, + "loss": 0.442, + "num_input_tokens_seen": 2132016, + "step": 7465 + }, + { + "epoch": 13.936567164179104, + "grad_norm": 0.6582010388374329, + "learning_rate": 1.274755308380624e-05, + "loss": 0.6739, + "num_input_tokens_seen": 2133488, + "step": 7470 + }, + { + "epoch": 13.94589552238806, + "grad_norm": 0.5281171202659607, + "learning_rate": 1.2712090172702951e-05, + "loss": 0.6493, + "num_input_tokens_seen": 2134704, + "step": 7475 + }, + { + "epoch": 13.955223880597014, + "grad_norm": 0.4359031617641449, + "learning_rate": 1.2676659833499044e-05, + "loss": 0.5984, + "num_input_tokens_seen": 2136240, + "step": 7480 + }, + { + "epoch": 13.96455223880597, + "grad_norm": 0.5118557810783386, + "learning_rate": 1.2641262160110645e-05, + "loss": 0.8923, + "num_input_tokens_seen": 2137648, + "step": 7485 + }, + { + "epoch": 13.973880597014926, + "grad_norm": 0.48048529028892517, + "learning_rate": 1.260589724636736e-05, + "loss": 0.711, + "num_input_tokens_seen": 2138992, + "step": 7490 + }, + { + "epoch": 13.98320895522388, + "grad_norm": 0.535887598991394, + "learning_rate": 1.2570565186011946e-05, + "loss": 0.7049, + "num_input_tokens_seen": 2140432, + "step": 7495 + }, + { + "epoch": 13.992537313432836, + "grad_norm": 0.4411444365978241, + "learning_rate": 1.2535266072700052e-05, + "loss": 0.5278, + "num_input_tokens_seen": 2141808, + "step": 7500 + }, + { + "epoch": 14.0, + "eval_loss": 0.7004382610321045, + "eval_runtime": 2.4573, + "eval_samples_per_second": 96.856, + "eval_steps_per_second": 24.418, + "num_input_tokens_seen": 2142632, + "step": 7504 + }, + { + "epoch": 14.001865671641792, + "grad_norm": 0.348638117313385, + "learning_rate": 1.2500000000000006e-05, + "loss": 0.8338, + "num_input_tokens_seen": 2142920, + "step": 7505 + }, + { + "epoch": 14.011194029850746, + "grad_norm": 0.807561457157135, + "learning_rate": 1.2464767061392543e-05, + "loss": 0.7153, + "num_input_tokens_seen": 2144136, + "step": 7510 + }, + { + "epoch": 14.020522388059701, + "grad_norm": 0.4515995979309082, + "learning_rate": 1.2429567350270596e-05, + "loss": 0.9013, + "num_input_tokens_seen": 2145544, + "step": 7515 + }, + { + "epoch": 14.029850746268657, + "grad_norm": 0.45097631216049194, + "learning_rate": 1.2394400959939001e-05, + "loss": 0.9691, + "num_input_tokens_seen": 2146792, + "step": 7520 + }, + { + "epoch": 14.039179104477611, + "grad_norm": 0.38524311780929565, + "learning_rate": 1.2359267983614256e-05, + "loss": 0.7586, + "num_input_tokens_seen": 2148264, + "step": 7525 + }, + { + "epoch": 14.048507462686567, + "grad_norm": 0.49703046679496765, + "learning_rate": 1.2324168514424328e-05, + "loss": 0.6523, + "num_input_tokens_seen": 2149672, + "step": 7530 + }, + { + "epoch": 14.057835820895523, + "grad_norm": 0.7228829264640808, + "learning_rate": 1.2289102645408329e-05, + "loss": 0.6841, + "num_input_tokens_seen": 2151112, + "step": 7535 + }, + { + "epoch": 14.067164179104477, + "grad_norm": 0.35217952728271484, + "learning_rate": 1.2254070469516324e-05, + "loss": 0.5628, + "num_input_tokens_seen": 2152552, + "step": 7540 + }, + { + "epoch": 14.076492537313433, + "grad_norm": 0.41561049222946167, + "learning_rate": 1.2219072079609045e-05, + "loss": 0.7227, + "num_input_tokens_seen": 2154024, + "step": 7545 + }, + { + "epoch": 14.085820895522389, + "grad_norm": 0.6066386103630066, + "learning_rate": 1.2184107568457709e-05, + "loss": 0.7981, + "num_input_tokens_seen": 2155208, + "step": 7550 + }, + { + "epoch": 14.095149253731343, + "grad_norm": 0.37525200843811035, + "learning_rate": 1.2149177028743691e-05, + "loss": 0.5362, + "num_input_tokens_seen": 2156648, + "step": 7555 + }, + { + "epoch": 14.104477611940299, + "grad_norm": 0.5538756847381592, + "learning_rate": 1.2114280553058324e-05, + "loss": 1.1871, + "num_input_tokens_seen": 2158088, + "step": 7560 + }, + { + "epoch": 14.113805970149254, + "grad_norm": 0.5264307856559753, + "learning_rate": 1.2079418233902667e-05, + "loss": 0.7186, + "num_input_tokens_seen": 2159496, + "step": 7565 + }, + { + "epoch": 14.123134328358208, + "grad_norm": 0.35347917675971985, + "learning_rate": 1.2044590163687219e-05, + "loss": 0.503, + "num_input_tokens_seen": 2161288, + "step": 7570 + }, + { + "epoch": 14.132462686567164, + "grad_norm": 0.4889480769634247, + "learning_rate": 1.2009796434731688e-05, + "loss": 0.7122, + "num_input_tokens_seen": 2162600, + "step": 7575 + }, + { + "epoch": 14.14179104477612, + "grad_norm": 0.44853460788726807, + "learning_rate": 1.1975037139264782e-05, + "loss": 0.5682, + "num_input_tokens_seen": 2164328, + "step": 7580 + }, + { + "epoch": 14.151119402985074, + "grad_norm": 0.658817708492279, + "learning_rate": 1.1940312369423917e-05, + "loss": 0.6161, + "num_input_tokens_seen": 2165800, + "step": 7585 + }, + { + "epoch": 14.16044776119403, + "grad_norm": 0.43865370750427246, + "learning_rate": 1.1905622217254986e-05, + "loss": 0.9536, + "num_input_tokens_seen": 2167272, + "step": 7590 + }, + { + "epoch": 14.169776119402986, + "grad_norm": 0.35276228189468384, + "learning_rate": 1.1870966774712114e-05, + "loss": 0.5622, + "num_input_tokens_seen": 2168840, + "step": 7595 + }, + { + "epoch": 14.17910447761194, + "grad_norm": 0.4293757677078247, + "learning_rate": 1.1836346133657459e-05, + "loss": 0.7471, + "num_input_tokens_seen": 2170280, + "step": 7600 + }, + { + "epoch": 14.188432835820896, + "grad_norm": 0.4628300070762634, + "learning_rate": 1.1801760385860885e-05, + "loss": 0.5302, + "num_input_tokens_seen": 2171848, + "step": 7605 + }, + { + "epoch": 14.197761194029852, + "grad_norm": 0.44287726283073425, + "learning_rate": 1.1767209622999781e-05, + "loss": 0.7269, + "num_input_tokens_seen": 2173288, + "step": 7610 + }, + { + "epoch": 14.207089552238806, + "grad_norm": 0.5706661939620972, + "learning_rate": 1.1732693936658814e-05, + "loss": 0.8941, + "num_input_tokens_seen": 2174504, + "step": 7615 + }, + { + "epoch": 14.216417910447761, + "grad_norm": 0.4463728666305542, + "learning_rate": 1.169821341832965e-05, + "loss": 0.816, + "num_input_tokens_seen": 2176072, + "step": 7620 + }, + { + "epoch": 14.225746268656717, + "grad_norm": 0.6871224641799927, + "learning_rate": 1.1663768159410748e-05, + "loss": 0.602, + "num_input_tokens_seen": 2177288, + "step": 7625 + }, + { + "epoch": 14.235074626865671, + "grad_norm": 0.3281518220901489, + "learning_rate": 1.162935825120709e-05, + "loss": 0.6072, + "num_input_tokens_seen": 2178728, + "step": 7630 + }, + { + "epoch": 14.244402985074627, + "grad_norm": 0.6984177231788635, + "learning_rate": 1.1594983784929969e-05, + "loss": 0.8344, + "num_input_tokens_seen": 2180072, + "step": 7635 + }, + { + "epoch": 14.253731343283581, + "grad_norm": 0.40176549553871155, + "learning_rate": 1.1560644851696744e-05, + "loss": 0.649, + "num_input_tokens_seen": 2181608, + "step": 7640 + }, + { + "epoch": 14.263059701492537, + "grad_norm": 0.4366752803325653, + "learning_rate": 1.152634154253053e-05, + "loss": 0.5946, + "num_input_tokens_seen": 2183080, + "step": 7645 + }, + { + "epoch": 14.272388059701493, + "grad_norm": 0.3669073283672333, + "learning_rate": 1.149207394836008e-05, + "loss": 0.7453, + "num_input_tokens_seen": 2184680, + "step": 7650 + }, + { + "epoch": 14.281716417910447, + "grad_norm": 0.5296749472618103, + "learning_rate": 1.1457842160019419e-05, + "loss": 0.7204, + "num_input_tokens_seen": 2186120, + "step": 7655 + }, + { + "epoch": 14.291044776119403, + "grad_norm": 0.46801039576530457, + "learning_rate": 1.1423646268247722e-05, + "loss": 0.8708, + "num_input_tokens_seen": 2187528, + "step": 7660 + }, + { + "epoch": 14.300373134328359, + "grad_norm": 0.6651462316513062, + "learning_rate": 1.1389486363688934e-05, + "loss": 0.5588, + "num_input_tokens_seen": 2189032, + "step": 7665 + }, + { + "epoch": 14.309701492537313, + "grad_norm": 0.5248293876647949, + "learning_rate": 1.1355362536891673e-05, + "loss": 0.7261, + "num_input_tokens_seen": 2190408, + "step": 7670 + }, + { + "epoch": 14.319029850746269, + "grad_norm": 0.4290376901626587, + "learning_rate": 1.1321274878308919e-05, + "loss": 0.8362, + "num_input_tokens_seen": 2191816, + "step": 7675 + }, + { + "epoch": 14.328358208955224, + "grad_norm": 0.43097221851348877, + "learning_rate": 1.1287223478297724e-05, + "loss": 0.477, + "num_input_tokens_seen": 2193480, + "step": 7680 + }, + { + "epoch": 14.337686567164178, + "grad_norm": 0.31797534227371216, + "learning_rate": 1.1253208427119094e-05, + "loss": 0.5078, + "num_input_tokens_seen": 2194824, + "step": 7685 + }, + { + "epoch": 14.347014925373134, + "grad_norm": 0.45103785395622253, + "learning_rate": 1.1219229814937646e-05, + "loss": 0.4358, + "num_input_tokens_seen": 2196136, + "step": 7690 + }, + { + "epoch": 14.35634328358209, + "grad_norm": 0.4524008333683014, + "learning_rate": 1.1185287731821429e-05, + "loss": 0.6211, + "num_input_tokens_seen": 2197544, + "step": 7695 + }, + { + "epoch": 14.365671641791044, + "grad_norm": 0.43386319279670715, + "learning_rate": 1.1151382267741647e-05, + "loss": 0.5184, + "num_input_tokens_seen": 2199080, + "step": 7700 + }, + { + "epoch": 14.375, + "grad_norm": 0.5687544941902161, + "learning_rate": 1.1117513512572436e-05, + "loss": 0.6875, + "num_input_tokens_seen": 2200552, + "step": 7705 + }, + { + "epoch": 14.384328358208956, + "grad_norm": 0.4616321325302124, + "learning_rate": 1.108368155609065e-05, + "loss": 0.5411, + "num_input_tokens_seen": 2202152, + "step": 7710 + }, + { + "epoch": 14.39365671641791, + "grad_norm": 0.5329272150993347, + "learning_rate": 1.1049886487975572e-05, + "loss": 0.8335, + "num_input_tokens_seen": 2203592, + "step": 7715 + }, + { + "epoch": 14.402985074626866, + "grad_norm": 0.5026233792304993, + "learning_rate": 1.1016128397808716e-05, + "loss": 0.874, + "num_input_tokens_seen": 2205160, + "step": 7720 + }, + { + "epoch": 14.412313432835822, + "grad_norm": 0.5041499137878418, + "learning_rate": 1.0982407375073573e-05, + "loss": 0.6811, + "num_input_tokens_seen": 2206472, + "step": 7725 + }, + { + "epoch": 14.421641791044776, + "grad_norm": 0.6593303084373474, + "learning_rate": 1.0948723509155393e-05, + "loss": 0.5681, + "num_input_tokens_seen": 2207880, + "step": 7730 + }, + { + "epoch": 14.430970149253731, + "grad_norm": 0.6740328073501587, + "learning_rate": 1.0915076889340917e-05, + "loss": 0.7697, + "num_input_tokens_seen": 2209256, + "step": 7735 + }, + { + "epoch": 14.440298507462687, + "grad_norm": 0.3586782217025757, + "learning_rate": 1.0881467604818154e-05, + "loss": 0.9258, + "num_input_tokens_seen": 2210696, + "step": 7740 + }, + { + "epoch": 14.449626865671641, + "grad_norm": 0.29311618208885193, + "learning_rate": 1.0847895744676173e-05, + "loss": 0.514, + "num_input_tokens_seen": 2212168, + "step": 7745 + }, + { + "epoch": 14.458955223880597, + "grad_norm": 0.21502862870693207, + "learning_rate": 1.0814361397904812e-05, + "loss": 0.4064, + "num_input_tokens_seen": 2213800, + "step": 7750 + }, + { + "epoch": 14.468283582089553, + "grad_norm": 0.3098459839820862, + "learning_rate": 1.0780864653394478e-05, + "loss": 0.5757, + "num_input_tokens_seen": 2215368, + "step": 7755 + }, + { + "epoch": 14.477611940298507, + "grad_norm": 0.385659396648407, + "learning_rate": 1.0747405599935929e-05, + "loss": 0.8805, + "num_input_tokens_seen": 2216712, + "step": 7760 + }, + { + "epoch": 14.486940298507463, + "grad_norm": 0.5838318467140198, + "learning_rate": 1.0713984326219978e-05, + "loss": 0.7383, + "num_input_tokens_seen": 2218248, + "step": 7765 + }, + { + "epoch": 14.496268656716419, + "grad_norm": 0.4556827247142792, + "learning_rate": 1.0680600920837319e-05, + "loss": 0.5928, + "num_input_tokens_seen": 2219656, + "step": 7770 + }, + { + "epoch": 14.505597014925373, + "grad_norm": 0.5657076239585876, + "learning_rate": 1.064725547227825e-05, + "loss": 0.544, + "num_input_tokens_seen": 2221032, + "step": 7775 + }, + { + "epoch": 14.514925373134329, + "grad_norm": 0.4977869987487793, + "learning_rate": 1.0613948068932481e-05, + "loss": 0.7754, + "num_input_tokens_seen": 2222408, + "step": 7780 + }, + { + "epoch": 14.524253731343283, + "grad_norm": 0.6496837735176086, + "learning_rate": 1.0580678799088848e-05, + "loss": 0.7315, + "num_input_tokens_seen": 2223720, + "step": 7785 + }, + { + "epoch": 14.533582089552239, + "grad_norm": 0.6204323768615723, + "learning_rate": 1.054744775093511e-05, + "loss": 0.5427, + "num_input_tokens_seen": 2225000, + "step": 7790 + }, + { + "epoch": 14.542910447761194, + "grad_norm": 0.5654621124267578, + "learning_rate": 1.0514255012557738e-05, + "loss": 0.7828, + "num_input_tokens_seen": 2226536, + "step": 7795 + }, + { + "epoch": 14.552238805970148, + "grad_norm": 0.41968101263046265, + "learning_rate": 1.0481100671941618e-05, + "loss": 0.7369, + "num_input_tokens_seen": 2228040, + "step": 7800 + }, + { + "epoch": 14.561567164179104, + "grad_norm": 0.4859428405761719, + "learning_rate": 1.0447984816969874e-05, + "loss": 0.9318, + "num_input_tokens_seen": 2229320, + "step": 7805 + }, + { + "epoch": 14.57089552238806, + "grad_norm": 0.3616129457950592, + "learning_rate": 1.0414907535423602e-05, + "loss": 0.6861, + "num_input_tokens_seen": 2230824, + "step": 7810 + }, + { + "epoch": 14.580223880597014, + "grad_norm": 0.598272979259491, + "learning_rate": 1.0381868914981673e-05, + "loss": 0.8091, + "num_input_tokens_seen": 2232136, + "step": 7815 + }, + { + "epoch": 14.58955223880597, + "grad_norm": 0.40962934494018555, + "learning_rate": 1.0348869043220458e-05, + "loss": 0.6448, + "num_input_tokens_seen": 2233608, + "step": 7820 + }, + { + "epoch": 14.598880597014926, + "grad_norm": 0.47600996494293213, + "learning_rate": 1.0315908007613609e-05, + "loss": 0.7522, + "num_input_tokens_seen": 2235080, + "step": 7825 + }, + { + "epoch": 14.60820895522388, + "grad_norm": 0.60047447681427, + "learning_rate": 1.0282985895531865e-05, + "loss": 0.6735, + "num_input_tokens_seen": 2236488, + "step": 7830 + }, + { + "epoch": 14.617537313432836, + "grad_norm": 0.43118199706077576, + "learning_rate": 1.0250102794242767e-05, + "loss": 0.6169, + "num_input_tokens_seen": 2237896, + "step": 7835 + }, + { + "epoch": 14.626865671641792, + "grad_norm": 0.6091872453689575, + "learning_rate": 1.0217258790910448e-05, + "loss": 0.7797, + "num_input_tokens_seen": 2239240, + "step": 7840 + }, + { + "epoch": 14.636194029850746, + "grad_norm": 0.6718793511390686, + "learning_rate": 1.01844539725954e-05, + "loss": 0.7631, + "num_input_tokens_seen": 2240648, + "step": 7845 + }, + { + "epoch": 14.645522388059701, + "grad_norm": 0.42696183919906616, + "learning_rate": 1.0151688426254264e-05, + "loss": 0.8028, + "num_input_tokens_seen": 2242024, + "step": 7850 + }, + { + "epoch": 14.654850746268657, + "grad_norm": 0.49519696831703186, + "learning_rate": 1.0118962238739586e-05, + "loss": 0.4735, + "num_input_tokens_seen": 2243528, + "step": 7855 + }, + { + "epoch": 14.664179104477611, + "grad_norm": 0.302437961101532, + "learning_rate": 1.0086275496799535e-05, + "loss": 0.594, + "num_input_tokens_seen": 2244808, + "step": 7860 + }, + { + "epoch": 14.673507462686567, + "grad_norm": 0.521212637424469, + "learning_rate": 1.0053628287077782e-05, + "loss": 0.4419, + "num_input_tokens_seen": 2246088, + "step": 7865 + }, + { + "epoch": 14.682835820895523, + "grad_norm": 0.4451550543308258, + "learning_rate": 1.002102069611317e-05, + "loss": 0.7658, + "num_input_tokens_seen": 2247560, + "step": 7870 + }, + { + "epoch": 14.692164179104477, + "grad_norm": 0.3655129373073578, + "learning_rate": 9.988452810339527e-06, + "loss": 0.624, + "num_input_tokens_seen": 2248904, + "step": 7875 + }, + { + "epoch": 14.701492537313433, + "grad_norm": 0.7198426723480225, + "learning_rate": 9.955924716085455e-06, + "loss": 0.6424, + "num_input_tokens_seen": 2250248, + "step": 7880 + }, + { + "epoch": 14.710820895522389, + "grad_norm": 0.7240375280380249, + "learning_rate": 9.923436499574046e-06, + "loss": 0.6495, + "num_input_tokens_seen": 2251624, + "step": 7885 + }, + { + "epoch": 14.720149253731343, + "grad_norm": 0.33628085255622864, + "learning_rate": 9.89098824692274e-06, + "loss": 0.8104, + "num_input_tokens_seen": 2253192, + "step": 7890 + }, + { + "epoch": 14.729477611940299, + "grad_norm": 0.4623497426509857, + "learning_rate": 9.858580044142966e-06, + "loss": 0.8592, + "num_input_tokens_seen": 2254600, + "step": 7895 + }, + { + "epoch": 14.738805970149254, + "grad_norm": 0.30861082673072815, + "learning_rate": 9.826211977140065e-06, + "loss": 0.5353, + "num_input_tokens_seen": 2256008, + "step": 7900 + }, + { + "epoch": 14.748134328358208, + "grad_norm": 0.5111231803894043, + "learning_rate": 9.793884131712943e-06, + "loss": 0.6335, + "num_input_tokens_seen": 2257512, + "step": 7905 + }, + { + "epoch": 14.757462686567164, + "grad_norm": 0.4546292722225189, + "learning_rate": 9.761596593553924e-06, + "loss": 0.5276, + "num_input_tokens_seen": 2258760, + "step": 7910 + }, + { + "epoch": 14.76679104477612, + "grad_norm": 0.6024869680404663, + "learning_rate": 9.72934944824846e-06, + "loss": 0.7042, + "num_input_tokens_seen": 2259976, + "step": 7915 + }, + { + "epoch": 14.776119402985074, + "grad_norm": 0.36558103561401367, + "learning_rate": 9.69714278127493e-06, + "loss": 0.8714, + "num_input_tokens_seen": 2261320, + "step": 7920 + }, + { + "epoch": 14.78544776119403, + "grad_norm": 0.4374063014984131, + "learning_rate": 9.664976678004464e-06, + "loss": 0.7313, + "num_input_tokens_seen": 2262728, + "step": 7925 + }, + { + "epoch": 14.794776119402986, + "grad_norm": 0.27362290024757385, + "learning_rate": 9.632851223700593e-06, + "loss": 0.8989, + "num_input_tokens_seen": 2264232, + "step": 7930 + }, + { + "epoch": 14.80410447761194, + "grad_norm": 0.6172934770584106, + "learning_rate": 9.600766503519154e-06, + "loss": 0.608, + "num_input_tokens_seen": 2265608, + "step": 7935 + }, + { + "epoch": 14.813432835820896, + "grad_norm": 0.3909018933773041, + "learning_rate": 9.568722602508009e-06, + "loss": 0.4461, + "num_input_tokens_seen": 2266920, + "step": 7940 + }, + { + "epoch": 14.822761194029852, + "grad_norm": 0.4254691004753113, + "learning_rate": 9.536719605606795e-06, + "loss": 0.9562, + "num_input_tokens_seen": 2268200, + "step": 7945 + }, + { + "epoch": 14.832089552238806, + "grad_norm": 0.5555359721183777, + "learning_rate": 9.50475759764673e-06, + "loss": 1.0724, + "num_input_tokens_seen": 2269608, + "step": 7950 + }, + { + "epoch": 14.841417910447761, + "grad_norm": 0.4892066717147827, + "learning_rate": 9.472836663350377e-06, + "loss": 0.7326, + "num_input_tokens_seen": 2270920, + "step": 7955 + }, + { + "epoch": 14.850746268656717, + "grad_norm": 0.3675912618637085, + "learning_rate": 9.440956887331446e-06, + "loss": 0.4615, + "num_input_tokens_seen": 2272232, + "step": 7960 + }, + { + "epoch": 14.860074626865671, + "grad_norm": 0.29648450016975403, + "learning_rate": 9.40911835409453e-06, + "loss": 0.5132, + "num_input_tokens_seen": 2273576, + "step": 7965 + }, + { + "epoch": 14.869402985074627, + "grad_norm": 0.42020854353904724, + "learning_rate": 9.37732114803489e-06, + "loss": 0.8047, + "num_input_tokens_seen": 2274824, + "step": 7970 + }, + { + "epoch": 14.878731343283581, + "grad_norm": 0.5597383379936218, + "learning_rate": 9.345565353438268e-06, + "loss": 0.5939, + "num_input_tokens_seen": 2276200, + "step": 7975 + }, + { + "epoch": 14.888059701492537, + "grad_norm": 0.2963702976703644, + "learning_rate": 9.313851054480614e-06, + "loss": 0.6092, + "num_input_tokens_seen": 2277864, + "step": 7980 + }, + { + "epoch": 14.897388059701493, + "grad_norm": 0.3192145824432373, + "learning_rate": 9.282178335227884e-06, + "loss": 0.5526, + "num_input_tokens_seen": 2279432, + "step": 7985 + }, + { + "epoch": 14.906716417910447, + "grad_norm": 0.2655926048755646, + "learning_rate": 9.250547279635818e-06, + "loss": 0.5339, + "num_input_tokens_seen": 2281000, + "step": 7990 + }, + { + "epoch": 14.916044776119403, + "grad_norm": 0.41343241930007935, + "learning_rate": 9.218957971549742e-06, + "loss": 0.8317, + "num_input_tokens_seen": 2282408, + "step": 7995 + }, + { + "epoch": 14.925373134328359, + "grad_norm": 0.3386465013027191, + "learning_rate": 9.187410494704286e-06, + "loss": 0.6827, + "num_input_tokens_seen": 2284136, + "step": 8000 + }, + { + "epoch": 14.934701492537313, + "grad_norm": 0.4952506721019745, + "learning_rate": 9.155904932723202e-06, + "loss": 0.6157, + "num_input_tokens_seen": 2285544, + "step": 8005 + }, + { + "epoch": 14.944029850746269, + "grad_norm": 0.6702660918235779, + "learning_rate": 9.124441369119171e-06, + "loss": 0.75, + "num_input_tokens_seen": 2286952, + "step": 8010 + }, + { + "epoch": 14.953358208955224, + "grad_norm": 0.4703066051006317, + "learning_rate": 9.093019887293514e-06, + "loss": 0.6578, + "num_input_tokens_seen": 2288360, + "step": 8015 + }, + { + "epoch": 14.962686567164178, + "grad_norm": 0.4472179710865021, + "learning_rate": 9.061640570536007e-06, + "loss": 0.6491, + "num_input_tokens_seen": 2289704, + "step": 8020 + }, + { + "epoch": 14.972014925373134, + "grad_norm": 0.4690456986427307, + "learning_rate": 9.030303502024661e-06, + "loss": 0.6448, + "num_input_tokens_seen": 2291208, + "step": 8025 + }, + { + "epoch": 14.98134328358209, + "grad_norm": 0.4778496026992798, + "learning_rate": 8.99900876482552e-06, + "loss": 0.6586, + "num_input_tokens_seen": 2292552, + "step": 8030 + }, + { + "epoch": 14.990671641791044, + "grad_norm": 0.6813544034957886, + "learning_rate": 8.967756441892395e-06, + "loss": 0.6354, + "num_input_tokens_seen": 2293928, + "step": 8035 + }, + { + "epoch": 15.0, + "grad_norm": 0.5840533375740051, + "learning_rate": 8.93654661606666e-06, + "loss": 0.7199, + "num_input_tokens_seen": 2295280, + "step": 8040 + }, + { + "epoch": 15.0, + "eval_loss": 0.6973978877067566, + "eval_runtime": 2.4668, + "eval_samples_per_second": 96.483, + "eval_steps_per_second": 24.323, + "num_input_tokens_seen": 2295280, + "step": 8040 + }, + { + "epoch": 15.009328358208956, + "grad_norm": 0.4994085431098938, + "learning_rate": 8.905379370077077e-06, + "loss": 1.0967, + "num_input_tokens_seen": 2296752, + "step": 8045 + }, + { + "epoch": 15.01865671641791, + "grad_norm": 0.44239410758018494, + "learning_rate": 8.87425478653951e-06, + "loss": 0.7532, + "num_input_tokens_seen": 2298192, + "step": 8050 + }, + { + "epoch": 15.027985074626866, + "grad_norm": 0.42527177929878235, + "learning_rate": 8.843172947956737e-06, + "loss": 0.6167, + "num_input_tokens_seen": 2299696, + "step": 8055 + }, + { + "epoch": 15.037313432835822, + "grad_norm": 0.6217042803764343, + "learning_rate": 8.81213393671826e-06, + "loss": 0.5724, + "num_input_tokens_seen": 2301232, + "step": 8060 + }, + { + "epoch": 15.046641791044776, + "grad_norm": 0.4398249685764313, + "learning_rate": 8.78113783510002e-06, + "loss": 0.6978, + "num_input_tokens_seen": 2302576, + "step": 8065 + }, + { + "epoch": 15.055970149253731, + "grad_norm": 0.41197600960731506, + "learning_rate": 8.75018472526424e-06, + "loss": 0.5801, + "num_input_tokens_seen": 2303920, + "step": 8070 + }, + { + "epoch": 15.065298507462687, + "grad_norm": 0.4890790581703186, + "learning_rate": 8.719274689259166e-06, + "loss": 0.674, + "num_input_tokens_seen": 2305296, + "step": 8075 + }, + { + "epoch": 15.074626865671641, + "grad_norm": 0.5646170377731323, + "learning_rate": 8.688407809018895e-06, + "loss": 0.6452, + "num_input_tokens_seen": 2306672, + "step": 8080 + }, + { + "epoch": 15.083955223880597, + "grad_norm": 0.5388579964637756, + "learning_rate": 8.657584166363103e-06, + "loss": 0.5604, + "num_input_tokens_seen": 2308016, + "step": 8085 + }, + { + "epoch": 15.093283582089553, + "grad_norm": 0.3548063337802887, + "learning_rate": 8.626803842996856e-06, + "loss": 0.7138, + "num_input_tokens_seen": 2309392, + "step": 8090 + }, + { + "epoch": 15.102611940298507, + "grad_norm": 0.5212141871452332, + "learning_rate": 8.596066920510417e-06, + "loss": 0.7351, + "num_input_tokens_seen": 2310992, + "step": 8095 + }, + { + "epoch": 15.111940298507463, + "grad_norm": 0.4723356068134308, + "learning_rate": 8.565373480378977e-06, + "loss": 0.82, + "num_input_tokens_seen": 2312432, + "step": 8100 + }, + { + "epoch": 15.121268656716419, + "grad_norm": 0.6648253202438354, + "learning_rate": 8.534723603962497e-06, + "loss": 0.9291, + "num_input_tokens_seen": 2313840, + "step": 8105 + }, + { + "epoch": 15.130597014925373, + "grad_norm": 0.4424423277378082, + "learning_rate": 8.504117372505416e-06, + "loss": 0.7597, + "num_input_tokens_seen": 2315216, + "step": 8110 + }, + { + "epoch": 15.139925373134329, + "grad_norm": 0.4375438392162323, + "learning_rate": 8.473554867136524e-06, + "loss": 0.8602, + "num_input_tokens_seen": 2316464, + "step": 8115 + }, + { + "epoch": 15.149253731343283, + "grad_norm": 0.38585683703422546, + "learning_rate": 8.443036168868709e-06, + "loss": 0.5528, + "num_input_tokens_seen": 2318000, + "step": 8120 + }, + { + "epoch": 15.158582089552239, + "grad_norm": 0.3575310707092285, + "learning_rate": 8.412561358598693e-06, + "loss": 0.531, + "num_input_tokens_seen": 2319504, + "step": 8125 + }, + { + "epoch": 15.167910447761194, + "grad_norm": 0.20812393724918365, + "learning_rate": 8.382130517106907e-06, + "loss": 0.63, + "num_input_tokens_seen": 2321008, + "step": 8130 + }, + { + "epoch": 15.177238805970148, + "grad_norm": 0.4260925352573395, + "learning_rate": 8.351743725057204e-06, + "loss": 0.5554, + "num_input_tokens_seen": 2322448, + "step": 8135 + }, + { + "epoch": 15.186567164179104, + "grad_norm": 0.3125847578048706, + "learning_rate": 8.321401062996714e-06, + "loss": 0.5592, + "num_input_tokens_seen": 2324272, + "step": 8140 + }, + { + "epoch": 15.19589552238806, + "grad_norm": 0.31381407380104065, + "learning_rate": 8.291102611355525e-06, + "loss": 0.6385, + "num_input_tokens_seen": 2325776, + "step": 8145 + }, + { + "epoch": 15.205223880597014, + "grad_norm": 0.47557154297828674, + "learning_rate": 8.260848450446596e-06, + "loss": 0.7206, + "num_input_tokens_seen": 2327056, + "step": 8150 + }, + { + "epoch": 15.21455223880597, + "grad_norm": 0.37800341844558716, + "learning_rate": 8.230638660465461e-06, + "loss": 0.4897, + "num_input_tokens_seen": 2328496, + "step": 8155 + }, + { + "epoch": 15.223880597014926, + "grad_norm": 0.34270042181015015, + "learning_rate": 8.200473321490035e-06, + "loss": 0.6865, + "num_input_tokens_seen": 2330064, + "step": 8160 + }, + { + "epoch": 15.23320895522388, + "grad_norm": 0.5298364162445068, + "learning_rate": 8.170352513480408e-06, + "loss": 0.7219, + "num_input_tokens_seen": 2331536, + "step": 8165 + }, + { + "epoch": 15.242537313432836, + "grad_norm": 0.49227920174598694, + "learning_rate": 8.140276316278623e-06, + "loss": 0.6468, + "num_input_tokens_seen": 2332848, + "step": 8170 + }, + { + "epoch": 15.251865671641792, + "grad_norm": 0.42142006754875183, + "learning_rate": 8.110244809608495e-06, + "loss": 0.6995, + "num_input_tokens_seen": 2334448, + "step": 8175 + }, + { + "epoch": 15.261194029850746, + "grad_norm": 0.2300989329814911, + "learning_rate": 8.080258073075357e-06, + "loss": 0.733, + "num_input_tokens_seen": 2335856, + "step": 8180 + }, + { + "epoch": 15.270522388059701, + "grad_norm": 0.4269353747367859, + "learning_rate": 8.050316186165863e-06, + "loss": 0.8391, + "num_input_tokens_seen": 2337392, + "step": 8185 + }, + { + "epoch": 15.279850746268657, + "grad_norm": 0.43494758009910583, + "learning_rate": 8.020419228247807e-06, + "loss": 0.3202, + "num_input_tokens_seen": 2339056, + "step": 8190 + }, + { + "epoch": 15.289179104477611, + "grad_norm": 0.23484297096729279, + "learning_rate": 7.990567278569872e-06, + "loss": 0.6101, + "num_input_tokens_seen": 2340624, + "step": 8195 + }, + { + "epoch": 15.298507462686567, + "grad_norm": 0.3341365158557892, + "learning_rate": 7.960760416261437e-06, + "loss": 0.506, + "num_input_tokens_seen": 2341936, + "step": 8200 + }, + { + "epoch": 15.307835820895523, + "grad_norm": 0.4534667134284973, + "learning_rate": 7.930998720332358e-06, + "loss": 0.6198, + "num_input_tokens_seen": 2343280, + "step": 8205 + }, + { + "epoch": 15.317164179104477, + "grad_norm": 0.5420719385147095, + "learning_rate": 7.901282269672799e-06, + "loss": 0.7116, + "num_input_tokens_seen": 2344624, + "step": 8210 + }, + { + "epoch": 15.326492537313433, + "grad_norm": 0.33378538489341736, + "learning_rate": 7.87161114305296e-06, + "loss": 0.5818, + "num_input_tokens_seen": 2346192, + "step": 8215 + }, + { + "epoch": 15.335820895522389, + "grad_norm": 0.6411337852478027, + "learning_rate": 7.8419854191229e-06, + "loss": 0.6275, + "num_input_tokens_seen": 2347568, + "step": 8220 + }, + { + "epoch": 15.345149253731343, + "grad_norm": 0.4301784038543701, + "learning_rate": 7.812405176412355e-06, + "loss": 0.6469, + "num_input_tokens_seen": 2348816, + "step": 8225 + }, + { + "epoch": 15.354477611940299, + "grad_norm": 0.4400712847709656, + "learning_rate": 7.782870493330473e-06, + "loss": 0.5948, + "num_input_tokens_seen": 2350384, + "step": 8230 + }, + { + "epoch": 15.363805970149254, + "grad_norm": 0.39551377296447754, + "learning_rate": 7.75338144816565e-06, + "loss": 0.7528, + "num_input_tokens_seen": 2351696, + "step": 8235 + }, + { + "epoch": 15.373134328358208, + "grad_norm": 0.21130073070526123, + "learning_rate": 7.723938119085314e-06, + "loss": 0.614, + "num_input_tokens_seen": 2353328, + "step": 8240 + }, + { + "epoch": 15.382462686567164, + "grad_norm": 0.35370466113090515, + "learning_rate": 7.694540584135696e-06, + "loss": 0.7092, + "num_input_tokens_seen": 2354608, + "step": 8245 + }, + { + "epoch": 15.39179104477612, + "grad_norm": 0.465166836977005, + "learning_rate": 7.665188921241654e-06, + "loss": 0.6503, + "num_input_tokens_seen": 2356144, + "step": 8250 + }, + { + "epoch": 15.401119402985074, + "grad_norm": 0.37843093276023865, + "learning_rate": 7.635883208206429e-06, + "loss": 0.6161, + "num_input_tokens_seen": 2357680, + "step": 8255 + }, + { + "epoch": 15.41044776119403, + "grad_norm": 0.6032745242118835, + "learning_rate": 7.606623522711498e-06, + "loss": 0.6728, + "num_input_tokens_seen": 2359120, + "step": 8260 + }, + { + "epoch": 15.419776119402986, + "grad_norm": 0.6538360714912415, + "learning_rate": 7.5774099423163045e-06, + "loss": 0.9303, + "num_input_tokens_seen": 2360496, + "step": 8265 + }, + { + "epoch": 15.42910447761194, + "grad_norm": 0.5246914625167847, + "learning_rate": 7.548242544458078e-06, + "loss": 0.6379, + "num_input_tokens_seen": 2361936, + "step": 8270 + }, + { + "epoch": 15.438432835820896, + "grad_norm": 0.5487461090087891, + "learning_rate": 7.519121406451657e-06, + "loss": 0.8631, + "num_input_tokens_seen": 2363280, + "step": 8275 + }, + { + "epoch": 15.447761194029852, + "grad_norm": 0.6487524509429932, + "learning_rate": 7.490046605489226e-06, + "loss": 0.7219, + "num_input_tokens_seen": 2364720, + "step": 8280 + }, + { + "epoch": 15.457089552238806, + "grad_norm": 0.34058287739753723, + "learning_rate": 7.461018218640162e-06, + "loss": 0.7246, + "num_input_tokens_seen": 2366160, + "step": 8285 + }, + { + "epoch": 15.466417910447761, + "grad_norm": 0.4407007396221161, + "learning_rate": 7.432036322850797e-06, + "loss": 0.9202, + "num_input_tokens_seen": 2367664, + "step": 8290 + }, + { + "epoch": 15.475746268656717, + "grad_norm": 0.4392961263656616, + "learning_rate": 7.403100994944251e-06, + "loss": 0.6032, + "num_input_tokens_seen": 2369136, + "step": 8295 + }, + { + "epoch": 15.485074626865671, + "grad_norm": 0.6045616269111633, + "learning_rate": 7.374212311620182e-06, + "loss": 0.4354, + "num_input_tokens_seen": 2370608, + "step": 8300 + }, + { + "epoch": 15.494402985074627, + "grad_norm": 0.44402018189430237, + "learning_rate": 7.345370349454611e-06, + "loss": 0.6678, + "num_input_tokens_seen": 2371920, + "step": 8305 + }, + { + "epoch": 15.503731343283581, + "grad_norm": 0.36985260248184204, + "learning_rate": 7.3165751848997296e-06, + "loss": 0.6275, + "num_input_tokens_seen": 2373328, + "step": 8310 + }, + { + "epoch": 15.513059701492537, + "grad_norm": 0.3841911852359772, + "learning_rate": 7.287826894283664e-06, + "loss": 0.7171, + "num_input_tokens_seen": 2374640, + "step": 8315 + }, + { + "epoch": 15.522388059701493, + "grad_norm": 0.3771035373210907, + "learning_rate": 7.259125553810295e-06, + "loss": 0.8254, + "num_input_tokens_seen": 2376016, + "step": 8320 + }, + { + "epoch": 15.531716417910447, + "grad_norm": 0.44147777557373047, + "learning_rate": 7.230471239559042e-06, + "loss": 0.5453, + "num_input_tokens_seen": 2377360, + "step": 8325 + }, + { + "epoch": 15.541044776119403, + "grad_norm": 0.8368029594421387, + "learning_rate": 7.201864027484695e-06, + "loss": 0.9809, + "num_input_tokens_seen": 2378960, + "step": 8330 + }, + { + "epoch": 15.550373134328359, + "grad_norm": 0.8279643058776855, + "learning_rate": 7.173303993417185e-06, + "loss": 0.6957, + "num_input_tokens_seen": 2380208, + "step": 8335 + }, + { + "epoch": 15.559701492537313, + "grad_norm": 0.4760940968990326, + "learning_rate": 7.144791213061347e-06, + "loss": 0.8688, + "num_input_tokens_seen": 2381552, + "step": 8340 + }, + { + "epoch": 15.569029850746269, + "grad_norm": 0.4211680293083191, + "learning_rate": 7.116325761996817e-06, + "loss": 0.7907, + "num_input_tokens_seen": 2382960, + "step": 8345 + }, + { + "epoch": 15.578358208955224, + "grad_norm": 0.6073769927024841, + "learning_rate": 7.087907715677733e-06, + "loss": 0.8631, + "num_input_tokens_seen": 2384368, + "step": 8350 + }, + { + "epoch": 15.587686567164178, + "grad_norm": 0.861519455909729, + "learning_rate": 7.059537149432582e-06, + "loss": 0.8371, + "num_input_tokens_seen": 2385552, + "step": 8355 + }, + { + "epoch": 15.597014925373134, + "grad_norm": 0.41607704758644104, + "learning_rate": 7.031214138464023e-06, + "loss": 0.6886, + "num_input_tokens_seen": 2387152, + "step": 8360 + }, + { + "epoch": 15.60634328358209, + "grad_norm": 0.4704280197620392, + "learning_rate": 7.0029387578486146e-06, + "loss": 0.7251, + "num_input_tokens_seen": 2388816, + "step": 8365 + }, + { + "epoch": 15.615671641791044, + "grad_norm": 0.43118754029273987, + "learning_rate": 6.974711082536711e-06, + "loss": 0.6045, + "num_input_tokens_seen": 2390480, + "step": 8370 + }, + { + "epoch": 15.625, + "grad_norm": 0.32585641741752625, + "learning_rate": 6.946531187352154e-06, + "loss": 0.5628, + "num_input_tokens_seen": 2391728, + "step": 8375 + }, + { + "epoch": 15.634328358208956, + "grad_norm": 0.30416399240493774, + "learning_rate": 6.918399146992183e-06, + "loss": 0.5676, + "num_input_tokens_seen": 2393232, + "step": 8380 + }, + { + "epoch": 15.64365671641791, + "grad_norm": 0.39825281500816345, + "learning_rate": 6.8903150360271565e-06, + "loss": 0.4584, + "num_input_tokens_seen": 2394672, + "step": 8385 + }, + { + "epoch": 15.652985074626866, + "grad_norm": 0.4085557460784912, + "learning_rate": 6.862278928900412e-06, + "loss": 0.6913, + "num_input_tokens_seen": 2396048, + "step": 8390 + }, + { + "epoch": 15.662313432835822, + "grad_norm": 0.27569982409477234, + "learning_rate": 6.8342908999280195e-06, + "loss": 0.6222, + "num_input_tokens_seen": 2397488, + "step": 8395 + }, + { + "epoch": 15.671641791044776, + "grad_norm": 0.5066648721694946, + "learning_rate": 6.806351023298604e-06, + "loss": 0.6885, + "num_input_tokens_seen": 2398768, + "step": 8400 + }, + { + "epoch": 15.680970149253731, + "grad_norm": 0.46355175971984863, + "learning_rate": 6.77845937307319e-06, + "loss": 0.7205, + "num_input_tokens_seen": 2400336, + "step": 8405 + }, + { + "epoch": 15.690298507462687, + "grad_norm": 0.42873358726501465, + "learning_rate": 6.750616023184905e-06, + "loss": 0.5871, + "num_input_tokens_seen": 2401616, + "step": 8410 + }, + { + "epoch": 15.699626865671641, + "grad_norm": 0.4757349193096161, + "learning_rate": 6.722821047438896e-06, + "loss": 1.0211, + "num_input_tokens_seen": 2402992, + "step": 8415 + }, + { + "epoch": 15.708955223880597, + "grad_norm": 0.4782407283782959, + "learning_rate": 6.695074519512076e-06, + "loss": 0.5903, + "num_input_tokens_seen": 2404304, + "step": 8420 + }, + { + "epoch": 15.718283582089553, + "grad_norm": 0.4857976734638214, + "learning_rate": 6.667376512952919e-06, + "loss": 0.7331, + "num_input_tokens_seen": 2405584, + "step": 8425 + }, + { + "epoch": 15.727611940298507, + "grad_norm": 0.45438507199287415, + "learning_rate": 6.639727101181286e-06, + "loss": 0.7305, + "num_input_tokens_seen": 2407056, + "step": 8430 + }, + { + "epoch": 15.736940298507463, + "grad_norm": 0.7237128019332886, + "learning_rate": 6.612126357488229e-06, + "loss": 0.7003, + "num_input_tokens_seen": 2408464, + "step": 8435 + }, + { + "epoch": 15.746268656716419, + "grad_norm": 0.4146021008491516, + "learning_rate": 6.584574355035808e-06, + "loss": 0.8157, + "num_input_tokens_seen": 2409872, + "step": 8440 + }, + { + "epoch": 15.755597014925373, + "grad_norm": 0.7049070596694946, + "learning_rate": 6.557071166856862e-06, + "loss": 0.8039, + "num_input_tokens_seen": 2411248, + "step": 8445 + }, + { + "epoch": 15.764925373134329, + "grad_norm": 0.2909722626209259, + "learning_rate": 6.529616865854843e-06, + "loss": 0.8162, + "num_input_tokens_seen": 2412592, + "step": 8450 + }, + { + "epoch": 15.774253731343283, + "grad_norm": 0.3592783212661743, + "learning_rate": 6.502211524803628e-06, + "loss": 0.6587, + "num_input_tokens_seen": 2413904, + "step": 8455 + }, + { + "epoch": 15.783582089552239, + "grad_norm": 0.4258784055709839, + "learning_rate": 6.4748552163473e-06, + "loss": 0.5774, + "num_input_tokens_seen": 2415248, + "step": 8460 + }, + { + "epoch": 15.792910447761194, + "grad_norm": 0.33167654275894165, + "learning_rate": 6.44754801299998e-06, + "loss": 0.5283, + "num_input_tokens_seen": 2416624, + "step": 8465 + }, + { + "epoch": 15.802238805970148, + "grad_norm": 0.3931379020214081, + "learning_rate": 6.420289987145609e-06, + "loss": 1.0523, + "num_input_tokens_seen": 2417968, + "step": 8470 + }, + { + "epoch": 15.811567164179104, + "grad_norm": 0.35015708208084106, + "learning_rate": 6.393081211037799e-06, + "loss": 0.8472, + "num_input_tokens_seen": 2419408, + "step": 8475 + }, + { + "epoch": 15.82089552238806, + "grad_norm": 0.42351722717285156, + "learning_rate": 6.365921756799589e-06, + "loss": 0.407, + "num_input_tokens_seen": 2420848, + "step": 8480 + }, + { + "epoch": 15.830223880597014, + "grad_norm": 0.3022542893886566, + "learning_rate": 6.338811696423283e-06, + "loss": 0.8248, + "num_input_tokens_seen": 2422160, + "step": 8485 + }, + { + "epoch": 15.83955223880597, + "grad_norm": 0.3654246926307678, + "learning_rate": 6.311751101770277e-06, + "loss": 0.9985, + "num_input_tokens_seen": 2423664, + "step": 8490 + }, + { + "epoch": 15.848880597014926, + "grad_norm": 0.6828741431236267, + "learning_rate": 6.284740044570825e-06, + "loss": 0.5824, + "num_input_tokens_seen": 2425200, + "step": 8495 + }, + { + "epoch": 15.85820895522388, + "grad_norm": 0.39014312624931335, + "learning_rate": 6.257778596423869e-06, + "loss": 0.6884, + "num_input_tokens_seen": 2426480, + "step": 8500 + }, + { + "epoch": 15.867537313432836, + "grad_norm": 0.49156954884529114, + "learning_rate": 6.230866828796861e-06, + "loss": 0.679, + "num_input_tokens_seen": 2427792, + "step": 8505 + }, + { + "epoch": 15.876865671641792, + "grad_norm": 0.39299091696739197, + "learning_rate": 6.204004813025568e-06, + "loss": 0.612, + "num_input_tokens_seen": 2429168, + "step": 8510 + }, + { + "epoch": 15.886194029850746, + "grad_norm": 0.6146236658096313, + "learning_rate": 6.177192620313868e-06, + "loss": 0.667, + "num_input_tokens_seen": 2430576, + "step": 8515 + }, + { + "epoch": 15.895522388059701, + "grad_norm": 0.5168152451515198, + "learning_rate": 6.150430321733566e-06, + "loss": 0.902, + "num_input_tokens_seen": 2431760, + "step": 8520 + }, + { + "epoch": 15.904850746268657, + "grad_norm": 0.46214205026626587, + "learning_rate": 6.123717988224237e-06, + "loss": 0.5681, + "num_input_tokens_seen": 2433232, + "step": 8525 + }, + { + "epoch": 15.914179104477611, + "grad_norm": 0.4695645868778229, + "learning_rate": 6.097055690592987e-06, + "loss": 0.5131, + "num_input_tokens_seen": 2434768, + "step": 8530 + }, + { + "epoch": 15.923507462686567, + "grad_norm": 0.38776537775993347, + "learning_rate": 6.070443499514292e-06, + "loss": 0.6904, + "num_input_tokens_seen": 2436144, + "step": 8535 + }, + { + "epoch": 15.932835820895523, + "grad_norm": 0.4838119447231293, + "learning_rate": 6.043881485529831e-06, + "loss": 0.5897, + "num_input_tokens_seen": 2437840, + "step": 8540 + }, + { + "epoch": 15.942164179104477, + "grad_norm": 0.313752681016922, + "learning_rate": 6.017369719048255e-06, + "loss": 0.413, + "num_input_tokens_seen": 2439280, + "step": 8545 + }, + { + "epoch": 15.951492537313433, + "grad_norm": 0.6786274313926697, + "learning_rate": 5.990908270345031e-06, + "loss": 0.8438, + "num_input_tokens_seen": 2440560, + "step": 8550 + }, + { + "epoch": 15.960820895522389, + "grad_norm": 0.26828068494796753, + "learning_rate": 5.964497209562234e-06, + "loss": 0.6767, + "num_input_tokens_seen": 2442128, + "step": 8555 + }, + { + "epoch": 15.970149253731343, + "grad_norm": 0.6297327876091003, + "learning_rate": 5.9381366067084e-06, + "loss": 0.794, + "num_input_tokens_seen": 2443472, + "step": 8560 + }, + { + "epoch": 15.979477611940299, + "grad_norm": 0.4175945222377777, + "learning_rate": 5.911826531658315e-06, + "loss": 0.6523, + "num_input_tokens_seen": 2444816, + "step": 8565 + }, + { + "epoch": 15.988805970149254, + "grad_norm": 0.4706919193267822, + "learning_rate": 5.885567054152785e-06, + "loss": 0.5527, + "num_input_tokens_seen": 2446416, + "step": 8570 + }, + { + "epoch": 15.998134328358208, + "grad_norm": 0.3357122242450714, + "learning_rate": 5.859358243798549e-06, + "loss": 0.5653, + "num_input_tokens_seen": 2447856, + "step": 8575 + }, + { + "epoch": 16.0, + "eval_loss": 0.6962363123893738, + "eval_runtime": 2.455, + "eval_samples_per_second": 96.947, + "eval_steps_per_second": 24.44, + "num_input_tokens_seen": 2447904, + "step": 8576 + }, + { + "epoch": 16.007462686567163, + "grad_norm": 0.44945383071899414, + "learning_rate": 5.8332001700680065e-06, + "loss": 0.7492, + "num_input_tokens_seen": 2448992, + "step": 8580 + }, + { + "epoch": 16.01679104477612, + "grad_norm": 0.3696384131908417, + "learning_rate": 5.8070929022991e-06, + "loss": 0.4226, + "num_input_tokens_seen": 2450560, + "step": 8585 + }, + { + "epoch": 16.026119402985074, + "grad_norm": 0.5610561966896057, + "learning_rate": 5.781036509695048e-06, + "loss": 0.9657, + "num_input_tokens_seen": 2452000, + "step": 8590 + }, + { + "epoch": 16.03544776119403, + "grad_norm": 0.5500913858413696, + "learning_rate": 5.755031061324264e-06, + "loss": 0.6881, + "num_input_tokens_seen": 2453408, + "step": 8595 + }, + { + "epoch": 16.044776119402986, + "grad_norm": 0.49463680386543274, + "learning_rate": 5.7290766261201165e-06, + "loss": 0.6584, + "num_input_tokens_seen": 2454880, + "step": 8600 + }, + { + "epoch": 16.05410447761194, + "grad_norm": 0.5242257714271545, + "learning_rate": 5.703173272880707e-06, + "loss": 0.7067, + "num_input_tokens_seen": 2456384, + "step": 8605 + }, + { + "epoch": 16.063432835820894, + "grad_norm": 0.5985044240951538, + "learning_rate": 5.67732107026879e-06, + "loss": 0.6575, + "num_input_tokens_seen": 2457664, + "step": 8610 + }, + { + "epoch": 16.07276119402985, + "grad_norm": 0.696968674659729, + "learning_rate": 5.6515200868114875e-06, + "loss": 0.8095, + "num_input_tokens_seen": 2459136, + "step": 8615 + }, + { + "epoch": 16.082089552238806, + "grad_norm": 0.3009604811668396, + "learning_rate": 5.625770390900189e-06, + "loss": 0.5055, + "num_input_tokens_seen": 2460576, + "step": 8620 + }, + { + "epoch": 16.09141791044776, + "grad_norm": 0.4586948752403259, + "learning_rate": 5.600072050790317e-06, + "loss": 0.6479, + "num_input_tokens_seen": 2461888, + "step": 8625 + }, + { + "epoch": 16.100746268656717, + "grad_norm": 0.4410480260848999, + "learning_rate": 5.574425134601152e-06, + "loss": 0.7645, + "num_input_tokens_seen": 2463200, + "step": 8630 + }, + { + "epoch": 16.11007462686567, + "grad_norm": 0.2329646795988083, + "learning_rate": 5.548829710315695e-06, + "loss": 0.6559, + "num_input_tokens_seen": 2464768, + "step": 8635 + }, + { + "epoch": 16.119402985074625, + "grad_norm": 0.6402687430381775, + "learning_rate": 5.523285845780432e-06, + "loss": 0.7409, + "num_input_tokens_seen": 2466208, + "step": 8640 + }, + { + "epoch": 16.128731343283583, + "grad_norm": 0.5215405821800232, + "learning_rate": 5.497793608705184e-06, + "loss": 0.6176, + "num_input_tokens_seen": 2467872, + "step": 8645 + }, + { + "epoch": 16.138059701492537, + "grad_norm": 0.36361077427864075, + "learning_rate": 5.472353066662916e-06, + "loss": 0.6753, + "num_input_tokens_seen": 2469280, + "step": 8650 + }, + { + "epoch": 16.14738805970149, + "grad_norm": 0.4101004898548126, + "learning_rate": 5.446964287089587e-06, + "loss": 0.4638, + "num_input_tokens_seen": 2470784, + "step": 8655 + }, + { + "epoch": 16.15671641791045, + "grad_norm": 0.4665655493736267, + "learning_rate": 5.421627337283916e-06, + "loss": 0.9052, + "num_input_tokens_seen": 2472032, + "step": 8660 + }, + { + "epoch": 16.166044776119403, + "grad_norm": 0.6178445219993591, + "learning_rate": 5.396342284407252e-06, + "loss": 0.8055, + "num_input_tokens_seen": 2473280, + "step": 8665 + }, + { + "epoch": 16.175373134328357, + "grad_norm": 0.4862574636936188, + "learning_rate": 5.3711091954833845e-06, + "loss": 0.5883, + "num_input_tokens_seen": 2474624, + "step": 8670 + }, + { + "epoch": 16.184701492537314, + "grad_norm": 0.7042259573936462, + "learning_rate": 5.345928137398351e-06, + "loss": 0.8962, + "num_input_tokens_seen": 2476064, + "step": 8675 + }, + { + "epoch": 16.19402985074627, + "grad_norm": 0.6566451787948608, + "learning_rate": 5.320799176900265e-06, + "loss": 0.924, + "num_input_tokens_seen": 2477440, + "step": 8680 + }, + { + "epoch": 16.203358208955223, + "grad_norm": 0.3997533321380615, + "learning_rate": 5.295722380599166e-06, + "loss": 0.9015, + "num_input_tokens_seen": 2478752, + "step": 8685 + }, + { + "epoch": 16.21268656716418, + "grad_norm": 0.3074372112751007, + "learning_rate": 5.270697814966793e-06, + "loss": 0.6064, + "num_input_tokens_seen": 2480288, + "step": 8690 + }, + { + "epoch": 16.222014925373134, + "grad_norm": 0.5059594511985779, + "learning_rate": 5.245725546336452e-06, + "loss": 0.5401, + "num_input_tokens_seen": 2481600, + "step": 8695 + }, + { + "epoch": 16.23134328358209, + "grad_norm": 0.5166656374931335, + "learning_rate": 5.220805640902815e-06, + "loss": 0.7612, + "num_input_tokens_seen": 2483040, + "step": 8700 + }, + { + "epoch": 16.240671641791046, + "grad_norm": 0.477531373500824, + "learning_rate": 5.1959381647217666e-06, + "loss": 0.7008, + "num_input_tokens_seen": 2484416, + "step": 8705 + }, + { + "epoch": 16.25, + "grad_norm": 0.5277060270309448, + "learning_rate": 5.171123183710205e-06, + "loss": 0.7239, + "num_input_tokens_seen": 2485856, + "step": 8710 + }, + { + "epoch": 16.259328358208954, + "grad_norm": 0.41368958353996277, + "learning_rate": 5.1463607636458716e-06, + "loss": 0.7499, + "num_input_tokens_seen": 2487456, + "step": 8715 + }, + { + "epoch": 16.26865671641791, + "grad_norm": 0.4700171649456024, + "learning_rate": 5.121650970167208e-06, + "loss": 0.8782, + "num_input_tokens_seen": 2488768, + "step": 8720 + }, + { + "epoch": 16.277985074626866, + "grad_norm": 0.4476558268070221, + "learning_rate": 5.096993868773131e-06, + "loss": 0.6315, + "num_input_tokens_seen": 2490336, + "step": 8725 + }, + { + "epoch": 16.28731343283582, + "grad_norm": 0.23112262785434723, + "learning_rate": 5.0723895248228955e-06, + "loss": 0.55, + "num_input_tokens_seen": 2492064, + "step": 8730 + }, + { + "epoch": 16.296641791044777, + "grad_norm": 0.36858755350112915, + "learning_rate": 5.047838003535904e-06, + "loss": 0.5514, + "num_input_tokens_seen": 2493600, + "step": 8735 + }, + { + "epoch": 16.30597014925373, + "grad_norm": 0.25203725695610046, + "learning_rate": 5.0233393699915645e-06, + "loss": 0.6067, + "num_input_tokens_seen": 2495040, + "step": 8740 + }, + { + "epoch": 16.315298507462686, + "grad_norm": 0.4163793623447418, + "learning_rate": 4.998893689129061e-06, + "loss": 0.6716, + "num_input_tokens_seen": 2496608, + "step": 8745 + }, + { + "epoch": 16.324626865671643, + "grad_norm": 0.35860389471054077, + "learning_rate": 4.974501025747233e-06, + "loss": 0.6361, + "num_input_tokens_seen": 2498112, + "step": 8750 + }, + { + "epoch": 16.333955223880597, + "grad_norm": 0.522433340549469, + "learning_rate": 4.950161444504386e-06, + "loss": 0.6253, + "num_input_tokens_seen": 2499584, + "step": 8755 + }, + { + "epoch": 16.34328358208955, + "grad_norm": 0.4152059555053711, + "learning_rate": 4.925875009918116e-06, + "loss": 0.6483, + "num_input_tokens_seen": 2501088, + "step": 8760 + }, + { + "epoch": 16.35261194029851, + "grad_norm": 0.564878523349762, + "learning_rate": 4.901641786365135e-06, + "loss": 0.6334, + "num_input_tokens_seen": 2502496, + "step": 8765 + }, + { + "epoch": 16.361940298507463, + "grad_norm": 0.5130676031112671, + "learning_rate": 4.8774618380811095e-06, + "loss": 0.9111, + "num_input_tokens_seen": 2503872, + "step": 8770 + }, + { + "epoch": 16.371268656716417, + "grad_norm": 0.4435340464115143, + "learning_rate": 4.853335229160497e-06, + "loss": 0.7567, + "num_input_tokens_seen": 2505248, + "step": 8775 + }, + { + "epoch": 16.380597014925375, + "grad_norm": 0.7174609899520874, + "learning_rate": 4.829262023556375e-06, + "loss": 0.8728, + "num_input_tokens_seen": 2506656, + "step": 8780 + }, + { + "epoch": 16.38992537313433, + "grad_norm": 0.263144850730896, + "learning_rate": 4.805242285080222e-06, + "loss": 0.6254, + "num_input_tokens_seen": 2508096, + "step": 8785 + }, + { + "epoch": 16.399253731343283, + "grad_norm": 0.40962842106819153, + "learning_rate": 4.7812760774018365e-06, + "loss": 0.8239, + "num_input_tokens_seen": 2509504, + "step": 8790 + }, + { + "epoch": 16.40858208955224, + "grad_norm": 0.5903270840644836, + "learning_rate": 4.757363464049094e-06, + "loss": 1.0347, + "num_input_tokens_seen": 2510880, + "step": 8795 + }, + { + "epoch": 16.417910447761194, + "grad_norm": 0.43245500326156616, + "learning_rate": 4.733504508407813e-06, + "loss": 0.6515, + "num_input_tokens_seen": 2512448, + "step": 8800 + }, + { + "epoch": 16.42723880597015, + "grad_norm": 0.4000587463378906, + "learning_rate": 4.7096992737215876e-06, + "loss": 0.7226, + "num_input_tokens_seen": 2513856, + "step": 8805 + }, + { + "epoch": 16.436567164179106, + "grad_norm": 0.4057605564594269, + "learning_rate": 4.6859478230916e-06, + "loss": 0.6, + "num_input_tokens_seen": 2515200, + "step": 8810 + }, + { + "epoch": 16.44589552238806, + "grad_norm": 0.3900735676288605, + "learning_rate": 4.6622502194764825e-06, + "loss": 0.8068, + "num_input_tokens_seen": 2516640, + "step": 8815 + }, + { + "epoch": 16.455223880597014, + "grad_norm": 0.38048678636550903, + "learning_rate": 4.6386065256921045e-06, + "loss": 0.5991, + "num_input_tokens_seen": 2518048, + "step": 8820 + }, + { + "epoch": 16.46455223880597, + "grad_norm": 0.3935447037220001, + "learning_rate": 4.615016804411465e-06, + "loss": 0.6461, + "num_input_tokens_seen": 2519584, + "step": 8825 + }, + { + "epoch": 16.473880597014926, + "grad_norm": 0.6210752725601196, + "learning_rate": 4.591481118164479e-06, + "loss": 0.574, + "num_input_tokens_seen": 2521056, + "step": 8830 + }, + { + "epoch": 16.48320895522388, + "grad_norm": 0.6848763227462769, + "learning_rate": 4.567999529337844e-06, + "loss": 0.7907, + "num_input_tokens_seen": 2522432, + "step": 8835 + }, + { + "epoch": 16.492537313432837, + "grad_norm": 0.5486647486686707, + "learning_rate": 4.544572100174843e-06, + "loss": 0.5547, + "num_input_tokens_seen": 2524064, + "step": 8840 + }, + { + "epoch": 16.50186567164179, + "grad_norm": 0.43720221519470215, + "learning_rate": 4.521198892775203e-06, + "loss": 0.9513, + "num_input_tokens_seen": 2525472, + "step": 8845 + }, + { + "epoch": 16.511194029850746, + "grad_norm": 0.4161680042743683, + "learning_rate": 4.4978799690949425e-06, + "loss": 0.5959, + "num_input_tokens_seen": 2526976, + "step": 8850 + }, + { + "epoch": 16.520522388059703, + "grad_norm": 0.39364758133888245, + "learning_rate": 4.474615390946144e-06, + "loss": 0.6969, + "num_input_tokens_seen": 2528448, + "step": 8855 + }, + { + "epoch": 16.529850746268657, + "grad_norm": 0.4170249402523041, + "learning_rate": 4.451405219996876e-06, + "loss": 0.6629, + "num_input_tokens_seen": 2530048, + "step": 8860 + }, + { + "epoch": 16.53917910447761, + "grad_norm": 0.3815752863883972, + "learning_rate": 4.428249517770986e-06, + "loss": 0.7137, + "num_input_tokens_seen": 2531520, + "step": 8865 + }, + { + "epoch": 16.548507462686565, + "grad_norm": 0.3457604944705963, + "learning_rate": 4.405148345647914e-06, + "loss": 0.7261, + "num_input_tokens_seen": 2532832, + "step": 8870 + }, + { + "epoch": 16.557835820895523, + "grad_norm": 0.48885202407836914, + "learning_rate": 4.382101764862576e-06, + "loss": 0.6418, + "num_input_tokens_seen": 2534176, + "step": 8875 + }, + { + "epoch": 16.567164179104477, + "grad_norm": 0.5833420157432556, + "learning_rate": 4.359109836505165e-06, + "loss": 0.7719, + "num_input_tokens_seen": 2535488, + "step": 8880 + }, + { + "epoch": 16.576492537313435, + "grad_norm": 0.3836781084537506, + "learning_rate": 4.336172621521034e-06, + "loss": 0.8383, + "num_input_tokens_seen": 2536768, + "step": 8885 + }, + { + "epoch": 16.58582089552239, + "grad_norm": 0.6093276143074036, + "learning_rate": 4.313290180710478e-06, + "loss": 0.7157, + "num_input_tokens_seen": 2538048, + "step": 8890 + }, + { + "epoch": 16.595149253731343, + "grad_norm": 0.4759739339351654, + "learning_rate": 4.290462574728599e-06, + "loss": 0.5675, + "num_input_tokens_seen": 2539552, + "step": 8895 + }, + { + "epoch": 16.604477611940297, + "grad_norm": 0.3585682511329651, + "learning_rate": 4.267689864085178e-06, + "loss": 0.6942, + "num_input_tokens_seen": 2540928, + "step": 8900 + }, + { + "epoch": 16.613805970149254, + "grad_norm": 0.31211256980895996, + "learning_rate": 4.244972109144454e-06, + "loss": 0.6661, + "num_input_tokens_seen": 2542560, + "step": 8905 + }, + { + "epoch": 16.62313432835821, + "grad_norm": 0.47176212072372437, + "learning_rate": 4.222309370124999e-06, + "loss": 0.5065, + "num_input_tokens_seen": 2543904, + "step": 8910 + }, + { + "epoch": 16.632462686567163, + "grad_norm": 0.5383201837539673, + "learning_rate": 4.199701707099557e-06, + "loss": 0.6756, + "num_input_tokens_seen": 2545152, + "step": 8915 + }, + { + "epoch": 16.64179104477612, + "grad_norm": 0.34184563159942627, + "learning_rate": 4.1771491799948885e-06, + "loss": 0.6352, + "num_input_tokens_seen": 2546496, + "step": 8920 + }, + { + "epoch": 16.651119402985074, + "grad_norm": 0.4637297987937927, + "learning_rate": 4.15465184859159e-06, + "loss": 0.6204, + "num_input_tokens_seen": 2547872, + "step": 8925 + }, + { + "epoch": 16.66044776119403, + "grad_norm": 0.6039229035377502, + "learning_rate": 4.132209772523945e-06, + "loss": 0.7709, + "num_input_tokens_seen": 2549088, + "step": 8930 + }, + { + "epoch": 16.669776119402986, + "grad_norm": 0.4813960790634155, + "learning_rate": 4.1098230112797984e-06, + "loss": 0.7717, + "num_input_tokens_seen": 2550528, + "step": 8935 + }, + { + "epoch": 16.67910447761194, + "grad_norm": 0.4195398688316345, + "learning_rate": 4.087491624200337e-06, + "loss": 0.5576, + "num_input_tokens_seen": 2551904, + "step": 8940 + }, + { + "epoch": 16.688432835820894, + "grad_norm": 0.850997269153595, + "learning_rate": 4.065215670479991e-06, + "loss": 0.5668, + "num_input_tokens_seen": 2553312, + "step": 8945 + }, + { + "epoch": 16.69776119402985, + "grad_norm": 0.3626323342323303, + "learning_rate": 4.042995209166225e-06, + "loss": 0.6103, + "num_input_tokens_seen": 2554784, + "step": 8950 + }, + { + "epoch": 16.707089552238806, + "grad_norm": 0.354629784822464, + "learning_rate": 4.020830299159445e-06, + "loss": 0.5605, + "num_input_tokens_seen": 2556352, + "step": 8955 + }, + { + "epoch": 16.71641791044776, + "grad_norm": 0.35588622093200684, + "learning_rate": 3.998720999212776e-06, + "loss": 0.7251, + "num_input_tokens_seen": 2557696, + "step": 8960 + }, + { + "epoch": 16.725746268656717, + "grad_norm": 0.3483979105949402, + "learning_rate": 3.976667367931941e-06, + "loss": 0.6714, + "num_input_tokens_seen": 2558976, + "step": 8965 + }, + { + "epoch": 16.73507462686567, + "grad_norm": 0.40790486335754395, + "learning_rate": 3.9546694637751125e-06, + "loss": 0.4293, + "num_input_tokens_seen": 2560352, + "step": 8970 + }, + { + "epoch": 16.744402985074625, + "grad_norm": 0.25143370032310486, + "learning_rate": 3.932727345052736e-06, + "loss": 0.7501, + "num_input_tokens_seen": 2561792, + "step": 8975 + }, + { + "epoch": 16.753731343283583, + "grad_norm": 0.42747998237609863, + "learning_rate": 3.910841069927379e-06, + "loss": 0.5213, + "num_input_tokens_seen": 2563200, + "step": 8980 + }, + { + "epoch": 16.763059701492537, + "grad_norm": 0.41014257073402405, + "learning_rate": 3.8890106964136056e-06, + "loss": 0.7597, + "num_input_tokens_seen": 2564576, + "step": 8985 + }, + { + "epoch": 16.77238805970149, + "grad_norm": 0.6378925442695618, + "learning_rate": 3.867236282377776e-06, + "loss": 0.5698, + "num_input_tokens_seen": 2566048, + "step": 8990 + }, + { + "epoch": 16.78171641791045, + "grad_norm": 0.5732792019844055, + "learning_rate": 3.845517885537927e-06, + "loss": 0.5675, + "num_input_tokens_seen": 2567456, + "step": 8995 + }, + { + "epoch": 16.791044776119403, + "grad_norm": 0.5211419463157654, + "learning_rate": 3.823855563463605e-06, + "loss": 0.8741, + "num_input_tokens_seen": 2568992, + "step": 9000 + }, + { + "epoch": 16.800373134328357, + "grad_norm": 0.5421302914619446, + "learning_rate": 3.802249373575728e-06, + "loss": 0.8683, + "num_input_tokens_seen": 2570304, + "step": 9005 + }, + { + "epoch": 16.809701492537314, + "grad_norm": 0.8489362001419067, + "learning_rate": 3.7806993731464154e-06, + "loss": 0.602, + "num_input_tokens_seen": 2571744, + "step": 9010 + }, + { + "epoch": 16.81902985074627, + "grad_norm": 0.5185752511024475, + "learning_rate": 3.7592056192988333e-06, + "loss": 0.6274, + "num_input_tokens_seen": 2573152, + "step": 9015 + }, + { + "epoch": 16.828358208955223, + "grad_norm": 0.4019967317581177, + "learning_rate": 3.7377681690070775e-06, + "loss": 0.5114, + "num_input_tokens_seen": 2574752, + "step": 9020 + }, + { + "epoch": 16.83768656716418, + "grad_norm": 0.39204978942871094, + "learning_rate": 3.716387079095973e-06, + "loss": 0.7044, + "num_input_tokens_seen": 2576096, + "step": 9025 + }, + { + "epoch": 16.847014925373134, + "grad_norm": 0.79412841796875, + "learning_rate": 3.695062406240979e-06, + "loss": 0.6232, + "num_input_tokens_seen": 2577472, + "step": 9030 + }, + { + "epoch": 16.85634328358209, + "grad_norm": 0.3637309968471527, + "learning_rate": 3.6737942069679675e-06, + "loss": 0.5703, + "num_input_tokens_seen": 2578944, + "step": 9035 + }, + { + "epoch": 16.865671641791046, + "grad_norm": 0.453798770904541, + "learning_rate": 3.6525825376531484e-06, + "loss": 0.6322, + "num_input_tokens_seen": 2580256, + "step": 9040 + }, + { + "epoch": 16.875, + "grad_norm": 0.3643854260444641, + "learning_rate": 3.631427454522887e-06, + "loss": 0.5837, + "num_input_tokens_seen": 2581696, + "step": 9045 + }, + { + "epoch": 16.884328358208954, + "grad_norm": 0.3290974795818329, + "learning_rate": 3.610329013653518e-06, + "loss": 0.4683, + "num_input_tokens_seen": 2583104, + "step": 9050 + }, + { + "epoch": 16.89365671641791, + "grad_norm": 0.270783931016922, + "learning_rate": 3.5892872709712726e-06, + "loss": 0.4954, + "num_input_tokens_seen": 2584480, + "step": 9055 + }, + { + "epoch": 16.902985074626866, + "grad_norm": 0.42543724179267883, + "learning_rate": 3.5683022822520675e-06, + "loss": 0.5894, + "num_input_tokens_seen": 2585920, + "step": 9060 + }, + { + "epoch": 16.91231343283582, + "grad_norm": 0.43959951400756836, + "learning_rate": 3.547374103121398e-06, + "loss": 0.699, + "num_input_tokens_seen": 2587488, + "step": 9065 + }, + { + "epoch": 16.921641791044777, + "grad_norm": 0.3797275424003601, + "learning_rate": 3.526502789054148e-06, + "loss": 0.5948, + "num_input_tokens_seen": 2588992, + "step": 9070 + }, + { + "epoch": 16.93097014925373, + "grad_norm": 0.7074321508407593, + "learning_rate": 3.5056883953744844e-06, + "loss": 0.7419, + "num_input_tokens_seen": 2590304, + "step": 9075 + }, + { + "epoch": 16.940298507462686, + "grad_norm": 0.5486608743667603, + "learning_rate": 3.4849309772557043e-06, + "loss": 0.5783, + "num_input_tokens_seen": 2591872, + "step": 9080 + }, + { + "epoch": 16.949626865671643, + "grad_norm": 0.41169974207878113, + "learning_rate": 3.4642305897200548e-06, + "loss": 0.7802, + "num_input_tokens_seen": 2593376, + "step": 9085 + }, + { + "epoch": 16.958955223880597, + "grad_norm": 0.34778860211372375, + "learning_rate": 3.4435872876386193e-06, + "loss": 0.5528, + "num_input_tokens_seen": 2594720, + "step": 9090 + }, + { + "epoch": 16.96828358208955, + "grad_norm": 0.41721537709236145, + "learning_rate": 3.4230011257311625e-06, + "loss": 0.6288, + "num_input_tokens_seen": 2596000, + "step": 9095 + }, + { + "epoch": 16.97761194029851, + "grad_norm": 0.502595841884613, + "learning_rate": 3.4024721585659976e-06, + "loss": 0.626, + "num_input_tokens_seen": 2597600, + "step": 9100 + }, + { + "epoch": 16.986940298507463, + "grad_norm": 0.564177393913269, + "learning_rate": 3.3820004405598154e-06, + "loss": 0.8882, + "num_input_tokens_seen": 2598976, + "step": 9105 + }, + { + "epoch": 16.996268656716417, + "grad_norm": 0.49010705947875977, + "learning_rate": 3.3615860259775507e-06, + "loss": 0.6505, + "num_input_tokens_seen": 2600448, + "step": 9110 + }, + { + "epoch": 17.0, + "eval_loss": 0.6949503421783447, + "eval_runtime": 2.4537, + "eval_samples_per_second": 96.995, + "eval_steps_per_second": 24.453, + "num_input_tokens_seen": 2600776, + "step": 9112 + }, + { + "epoch": 17.005597014925375, + "grad_norm": 0.42001962661743164, + "learning_rate": 3.3412289689322694e-06, + "loss": 0.9183, + "num_input_tokens_seen": 2601544, + "step": 9115 + }, + { + "epoch": 17.01492537313433, + "grad_norm": 0.27195996046066284, + "learning_rate": 3.3209293233849693e-06, + "loss": 0.5477, + "num_input_tokens_seen": 2603176, + "step": 9120 + }, + { + "epoch": 17.024253731343283, + "grad_norm": 0.8275244832038879, + "learning_rate": 3.300687143144482e-06, + "loss": 0.7076, + "num_input_tokens_seen": 2604520, + "step": 9125 + }, + { + "epoch": 17.03358208955224, + "grad_norm": 0.5557956695556641, + "learning_rate": 3.280502481867298e-06, + "loss": 0.6337, + "num_input_tokens_seen": 2605960, + "step": 9130 + }, + { + "epoch": 17.042910447761194, + "grad_norm": 0.4554276764392853, + "learning_rate": 3.260375393057469e-06, + "loss": 0.8117, + "num_input_tokens_seen": 2607304, + "step": 9135 + }, + { + "epoch": 17.05223880597015, + "grad_norm": 0.6719672083854675, + "learning_rate": 3.240305930066412e-06, + "loss": 0.5835, + "num_input_tokens_seen": 2608808, + "step": 9140 + }, + { + "epoch": 17.061567164179106, + "grad_norm": 0.5882443189620972, + "learning_rate": 3.2202941460927976e-06, + "loss": 0.8724, + "num_input_tokens_seen": 2610216, + "step": 9145 + }, + { + "epoch": 17.07089552238806, + "grad_norm": 0.5270630717277527, + "learning_rate": 3.2003400941824217e-06, + "loss": 0.9199, + "num_input_tokens_seen": 2611560, + "step": 9150 + }, + { + "epoch": 17.080223880597014, + "grad_norm": 0.3488577604293823, + "learning_rate": 3.180443827228033e-06, + "loss": 0.5101, + "num_input_tokens_seen": 2612904, + "step": 9155 + }, + { + "epoch": 17.08955223880597, + "grad_norm": 0.409368097782135, + "learning_rate": 3.160605397969202e-06, + "loss": 0.7331, + "num_input_tokens_seen": 2614376, + "step": 9160 + }, + { + "epoch": 17.098880597014926, + "grad_norm": 0.8213173747062683, + "learning_rate": 3.1408248589922083e-06, + "loss": 0.752, + "num_input_tokens_seen": 2615656, + "step": 9165 + }, + { + "epoch": 17.10820895522388, + "grad_norm": 0.5484946370124817, + "learning_rate": 3.1211022627298692e-06, + "loss": 0.6818, + "num_input_tokens_seen": 2617064, + "step": 9170 + }, + { + "epoch": 17.117537313432837, + "grad_norm": 0.571631669998169, + "learning_rate": 3.1014376614614036e-06, + "loss": 0.5772, + "num_input_tokens_seen": 2618472, + "step": 9175 + }, + { + "epoch": 17.12686567164179, + "grad_norm": 0.4218016266822815, + "learning_rate": 3.081831107312308e-06, + "loss": 0.5038, + "num_input_tokens_seen": 2619880, + "step": 9180 + }, + { + "epoch": 17.136194029850746, + "grad_norm": 0.46941930055618286, + "learning_rate": 3.0622826522542196e-06, + "loss": 0.7333, + "num_input_tokens_seen": 2621128, + "step": 9185 + }, + { + "epoch": 17.145522388059703, + "grad_norm": 0.46033117175102234, + "learning_rate": 3.0427923481047645e-06, + "loss": 0.6497, + "num_input_tokens_seen": 2622600, + "step": 9190 + }, + { + "epoch": 17.154850746268657, + "grad_norm": 0.41905683279037476, + "learning_rate": 3.023360246527418e-06, + "loss": 0.6129, + "num_input_tokens_seen": 2624136, + "step": 9195 + }, + { + "epoch": 17.16417910447761, + "grad_norm": 0.6310820579528809, + "learning_rate": 3.0039863990313917e-06, + "loss": 0.7486, + "num_input_tokens_seen": 2625544, + "step": 9200 + }, + { + "epoch": 17.17350746268657, + "grad_norm": 0.3800658881664276, + "learning_rate": 2.984670856971475e-06, + "loss": 0.6723, + "num_input_tokens_seen": 2626856, + "step": 9205 + }, + { + "epoch": 17.182835820895523, + "grad_norm": 0.8281236886978149, + "learning_rate": 2.965413671547901e-06, + "loss": 0.8374, + "num_input_tokens_seen": 2628200, + "step": 9210 + }, + { + "epoch": 17.192164179104477, + "grad_norm": 0.6250938773155212, + "learning_rate": 2.9462148938062123e-06, + "loss": 0.6805, + "num_input_tokens_seen": 2629640, + "step": 9215 + }, + { + "epoch": 17.20149253731343, + "grad_norm": 0.40851956605911255, + "learning_rate": 2.927074574637148e-06, + "loss": 0.623, + "num_input_tokens_seen": 2631080, + "step": 9220 + }, + { + "epoch": 17.21082089552239, + "grad_norm": 0.4707795977592468, + "learning_rate": 2.907992764776471e-06, + "loss": 0.9284, + "num_input_tokens_seen": 2632424, + "step": 9225 + }, + { + "epoch": 17.220149253731343, + "grad_norm": 0.4223838746547699, + "learning_rate": 2.888969514804854e-06, + "loss": 1.1096, + "num_input_tokens_seen": 2633736, + "step": 9230 + }, + { + "epoch": 17.229477611940297, + "grad_norm": 0.7155877351760864, + "learning_rate": 2.8700048751477527e-06, + "loss": 0.5835, + "num_input_tokens_seen": 2635368, + "step": 9235 + }, + { + "epoch": 17.238805970149254, + "grad_norm": 0.3016734719276428, + "learning_rate": 2.8510988960752575e-06, + "loss": 0.3853, + "num_input_tokens_seen": 2637160, + "step": 9240 + }, + { + "epoch": 17.24813432835821, + "grad_norm": 0.3952043056488037, + "learning_rate": 2.8322516277019624e-06, + "loss": 0.7172, + "num_input_tokens_seen": 2638536, + "step": 9245 + }, + { + "epoch": 17.257462686567163, + "grad_norm": 0.38378703594207764, + "learning_rate": 2.813463119986834e-06, + "loss": 0.6969, + "num_input_tokens_seen": 2639880, + "step": 9250 + }, + { + "epoch": 17.26679104477612, + "grad_norm": 0.3611210286617279, + "learning_rate": 2.7947334227330897e-06, + "loss": 0.5964, + "num_input_tokens_seen": 2641192, + "step": 9255 + }, + { + "epoch": 17.276119402985074, + "grad_norm": 0.37247994542121887, + "learning_rate": 2.776062585588063e-06, + "loss": 0.8397, + "num_input_tokens_seen": 2642600, + "step": 9260 + }, + { + "epoch": 17.28544776119403, + "grad_norm": 0.5225608944892883, + "learning_rate": 2.7574506580430287e-06, + "loss": 0.546, + "num_input_tokens_seen": 2644040, + "step": 9265 + }, + { + "epoch": 17.294776119402986, + "grad_norm": 0.46428027749061584, + "learning_rate": 2.7388976894331537e-06, + "loss": 0.5591, + "num_input_tokens_seen": 2645576, + "step": 9270 + }, + { + "epoch": 17.30410447761194, + "grad_norm": 0.549088180065155, + "learning_rate": 2.72040372893729e-06, + "loss": 0.711, + "num_input_tokens_seen": 2647144, + "step": 9275 + }, + { + "epoch": 17.313432835820894, + "grad_norm": 0.5152643322944641, + "learning_rate": 2.7019688255778857e-06, + "loss": 0.8925, + "num_input_tokens_seen": 2648488, + "step": 9280 + }, + { + "epoch": 17.32276119402985, + "grad_norm": 0.3638004660606384, + "learning_rate": 2.6835930282208517e-06, + "loss": 0.629, + "num_input_tokens_seen": 2649960, + "step": 9285 + }, + { + "epoch": 17.332089552238806, + "grad_norm": 0.37481188774108887, + "learning_rate": 2.6652763855754106e-06, + "loss": 0.7025, + "num_input_tokens_seen": 2651272, + "step": 9290 + }, + { + "epoch": 17.34141791044776, + "grad_norm": 0.6873860359191895, + "learning_rate": 2.647018946193999e-06, + "loss": 0.5435, + "num_input_tokens_seen": 2652584, + "step": 9295 + }, + { + "epoch": 17.350746268656717, + "grad_norm": 0.4900089502334595, + "learning_rate": 2.628820758472095e-06, + "loss": 0.7609, + "num_input_tokens_seen": 2654056, + "step": 9300 + }, + { + "epoch": 17.36007462686567, + "grad_norm": 0.4961356222629547, + "learning_rate": 2.610681870648149e-06, + "loss": 0.638, + "num_input_tokens_seen": 2655528, + "step": 9305 + }, + { + "epoch": 17.369402985074625, + "grad_norm": 0.4148246943950653, + "learning_rate": 2.5926023308033952e-06, + "loss": 0.6193, + "num_input_tokens_seen": 2656904, + "step": 9310 + }, + { + "epoch": 17.378731343283583, + "grad_norm": 0.33310869336128235, + "learning_rate": 2.5745821868617792e-06, + "loss": 0.4637, + "num_input_tokens_seen": 2658472, + "step": 9315 + }, + { + "epoch": 17.388059701492537, + "grad_norm": 0.5532823801040649, + "learning_rate": 2.556621486589783e-06, + "loss": 0.5945, + "num_input_tokens_seen": 2659784, + "step": 9320 + }, + { + "epoch": 17.39738805970149, + "grad_norm": 0.3730950355529785, + "learning_rate": 2.5387202775963236e-06, + "loss": 0.5395, + "num_input_tokens_seen": 2661448, + "step": 9325 + }, + { + "epoch": 17.40671641791045, + "grad_norm": 0.5891056060791016, + "learning_rate": 2.520878607332641e-06, + "loss": 0.6108, + "num_input_tokens_seen": 2662792, + "step": 9330 + }, + { + "epoch": 17.416044776119403, + "grad_norm": 0.47666287422180176, + "learning_rate": 2.5030965230921186e-06, + "loss": 0.6506, + "num_input_tokens_seen": 2663976, + "step": 9335 + }, + { + "epoch": 17.425373134328357, + "grad_norm": 0.4376184344291687, + "learning_rate": 2.485374072010224e-06, + "loss": 0.6417, + "num_input_tokens_seen": 2665544, + "step": 9340 + }, + { + "epoch": 17.434701492537314, + "grad_norm": 0.6541747450828552, + "learning_rate": 2.4677113010643486e-06, + "loss": 0.8112, + "num_input_tokens_seen": 2666952, + "step": 9345 + }, + { + "epoch": 17.44402985074627, + "grad_norm": 0.5104885101318359, + "learning_rate": 2.450108257073683e-06, + "loss": 0.5644, + "num_input_tokens_seen": 2668392, + "step": 9350 + }, + { + "epoch": 17.453358208955223, + "grad_norm": 0.6691657304763794, + "learning_rate": 2.4325649866990928e-06, + "loss": 0.7331, + "num_input_tokens_seen": 2669736, + "step": 9355 + }, + { + "epoch": 17.46268656716418, + "grad_norm": 0.24752648174762726, + "learning_rate": 2.4150815364430036e-06, + "loss": 0.6641, + "num_input_tokens_seen": 2671272, + "step": 9360 + }, + { + "epoch": 17.472014925373134, + "grad_norm": 0.4992641806602478, + "learning_rate": 2.397657952649285e-06, + "loss": 0.6265, + "num_input_tokens_seen": 2672680, + "step": 9365 + }, + { + "epoch": 17.48134328358209, + "grad_norm": 0.3523702025413513, + "learning_rate": 2.380294281503104e-06, + "loss": 0.7038, + "num_input_tokens_seen": 2674248, + "step": 9370 + }, + { + "epoch": 17.490671641791046, + "grad_norm": 0.3385435938835144, + "learning_rate": 2.3629905690308126e-06, + "loss": 0.6143, + "num_input_tokens_seen": 2675720, + "step": 9375 + }, + { + "epoch": 17.5, + "grad_norm": 0.5946886539459229, + "learning_rate": 2.3457468610998486e-06, + "loss": 0.5688, + "num_input_tokens_seen": 2677096, + "step": 9380 + }, + { + "epoch": 17.509328358208954, + "grad_norm": 0.6879421472549438, + "learning_rate": 2.328563203418574e-06, + "loss": 1.2265, + "num_input_tokens_seen": 2678504, + "step": 9385 + }, + { + "epoch": 17.51865671641791, + "grad_norm": 0.3800496757030487, + "learning_rate": 2.311439641536184e-06, + "loss": 0.803, + "num_input_tokens_seen": 2680008, + "step": 9390 + }, + { + "epoch": 17.527985074626866, + "grad_norm": 0.6054827570915222, + "learning_rate": 2.2943762208425646e-06, + "loss": 0.6276, + "num_input_tokens_seen": 2681480, + "step": 9395 + }, + { + "epoch": 17.53731343283582, + "grad_norm": 0.6662448644638062, + "learning_rate": 2.2773729865682046e-06, + "loss": 0.6515, + "num_input_tokens_seen": 2682824, + "step": 9400 + }, + { + "epoch": 17.546641791044777, + "grad_norm": 0.4717367887496948, + "learning_rate": 2.2604299837840374e-06, + "loss": 0.5848, + "num_input_tokens_seen": 2684232, + "step": 9405 + }, + { + "epoch": 17.55597014925373, + "grad_norm": 0.40102463960647583, + "learning_rate": 2.2435472574013433e-06, + "loss": 0.7093, + "num_input_tokens_seen": 2685800, + "step": 9410 + }, + { + "epoch": 17.565298507462686, + "grad_norm": 0.39639219641685486, + "learning_rate": 2.2267248521716327e-06, + "loss": 0.5712, + "num_input_tokens_seen": 2687144, + "step": 9415 + }, + { + "epoch": 17.574626865671643, + "grad_norm": 0.575713038444519, + "learning_rate": 2.209962812686514e-06, + "loss": 0.7221, + "num_input_tokens_seen": 2688648, + "step": 9420 + }, + { + "epoch": 17.583955223880597, + "grad_norm": 0.4188419282436371, + "learning_rate": 2.1932611833775846e-06, + "loss": 0.7375, + "num_input_tokens_seen": 2690120, + "step": 9425 + }, + { + "epoch": 17.59328358208955, + "grad_norm": 0.520841658115387, + "learning_rate": 2.1766200085163058e-06, + "loss": 0.594, + "num_input_tokens_seen": 2691528, + "step": 9430 + }, + { + "epoch": 17.60261194029851, + "grad_norm": 0.5092044472694397, + "learning_rate": 2.1600393322139034e-06, + "loss": 0.6798, + "num_input_tokens_seen": 2693096, + "step": 9435 + }, + { + "epoch": 17.611940298507463, + "grad_norm": 0.41505342721939087, + "learning_rate": 2.1435191984212315e-06, + "loss": 0.5149, + "num_input_tokens_seen": 2694664, + "step": 9440 + }, + { + "epoch": 17.621268656716417, + "grad_norm": 0.4712902009487152, + "learning_rate": 2.1270596509286504e-06, + "loss": 0.6447, + "num_input_tokens_seen": 2695944, + "step": 9445 + }, + { + "epoch": 17.630597014925375, + "grad_norm": 0.4986782371997833, + "learning_rate": 2.1106607333659463e-06, + "loss": 0.6261, + "num_input_tokens_seen": 2697320, + "step": 9450 + }, + { + "epoch": 17.63992537313433, + "grad_norm": 0.6195346713066101, + "learning_rate": 2.0943224892021746e-06, + "loss": 0.8068, + "num_input_tokens_seen": 2698664, + "step": 9455 + }, + { + "epoch": 17.649253731343283, + "grad_norm": 0.36630505323410034, + "learning_rate": 2.078044961745562e-06, + "loss": 0.6833, + "num_input_tokens_seen": 2700104, + "step": 9460 + }, + { + "epoch": 17.65858208955224, + "grad_norm": 0.49155208468437195, + "learning_rate": 2.061828194143406e-06, + "loss": 0.5058, + "num_input_tokens_seen": 2701640, + "step": 9465 + }, + { + "epoch": 17.667910447761194, + "grad_norm": 0.42041468620300293, + "learning_rate": 2.0456722293819315e-06, + "loss": 0.766, + "num_input_tokens_seen": 2703080, + "step": 9470 + }, + { + "epoch": 17.67723880597015, + "grad_norm": 0.5946452021598816, + "learning_rate": 2.0295771102861987e-06, + "loss": 0.6095, + "num_input_tokens_seen": 2704488, + "step": 9475 + }, + { + "epoch": 17.686567164179106, + "grad_norm": 0.369649738073349, + "learning_rate": 2.013542879519975e-06, + "loss": 0.7703, + "num_input_tokens_seen": 2705864, + "step": 9480 + }, + { + "epoch": 17.69589552238806, + "grad_norm": 0.5485965609550476, + "learning_rate": 1.997569579585648e-06, + "loss": 0.6, + "num_input_tokens_seen": 2707368, + "step": 9485 + }, + { + "epoch": 17.705223880597014, + "grad_norm": 0.46224379539489746, + "learning_rate": 1.9816572528240707e-06, + "loss": 0.5134, + "num_input_tokens_seen": 2708808, + "step": 9490 + }, + { + "epoch": 17.71455223880597, + "grad_norm": 0.35754474997520447, + "learning_rate": 1.9658059414144834e-06, + "loss": 0.9161, + "num_input_tokens_seen": 2710216, + "step": 9495 + }, + { + "epoch": 17.723880597014926, + "grad_norm": 0.4102347791194916, + "learning_rate": 1.9500156873743985e-06, + "loss": 0.4653, + "num_input_tokens_seen": 2711656, + "step": 9500 + }, + { + "epoch": 17.73320895522388, + "grad_norm": 0.42712458968162537, + "learning_rate": 1.934286532559468e-06, + "loss": 0.6761, + "num_input_tokens_seen": 2713160, + "step": 9505 + }, + { + "epoch": 17.742537313432837, + "grad_norm": 0.5572689771652222, + "learning_rate": 1.9186185186634066e-06, + "loss": 0.7755, + "num_input_tokens_seen": 2714568, + "step": 9510 + }, + { + "epoch": 17.75186567164179, + "grad_norm": 0.5185196399688721, + "learning_rate": 1.9030116872178316e-06, + "loss": 0.6768, + "num_input_tokens_seen": 2715976, + "step": 9515 + }, + { + "epoch": 17.761194029850746, + "grad_norm": 0.5089418888092041, + "learning_rate": 1.8874660795922067e-06, + "loss": 0.5356, + "num_input_tokens_seen": 2717320, + "step": 9520 + }, + { + "epoch": 17.770522388059703, + "grad_norm": 0.5636904835700989, + "learning_rate": 1.8719817369937082e-06, + "loss": 0.7166, + "num_input_tokens_seen": 2718664, + "step": 9525 + }, + { + "epoch": 17.779850746268657, + "grad_norm": 0.49879133701324463, + "learning_rate": 1.8565587004670898e-06, + "loss": 0.8246, + "num_input_tokens_seen": 2720104, + "step": 9530 + }, + { + "epoch": 17.78917910447761, + "grad_norm": 0.4227880537509918, + "learning_rate": 1.8411970108946296e-06, + "loss": 0.6914, + "num_input_tokens_seen": 2721608, + "step": 9535 + }, + { + "epoch": 17.798507462686565, + "grad_norm": 0.5307631492614746, + "learning_rate": 1.8258967089959749e-06, + "loss": 0.6322, + "num_input_tokens_seen": 2723016, + "step": 9540 + }, + { + "epoch": 17.807835820895523, + "grad_norm": 0.3892867863178253, + "learning_rate": 1.8106578353280585e-06, + "loss": 0.855, + "num_input_tokens_seen": 2724296, + "step": 9545 + }, + { + "epoch": 17.817164179104477, + "grad_norm": 0.3210139870643616, + "learning_rate": 1.7954804302849793e-06, + "loss": 0.7638, + "num_input_tokens_seen": 2725832, + "step": 9550 + }, + { + "epoch": 17.826492537313435, + "grad_norm": 0.5675643682479858, + "learning_rate": 1.7803645340978948e-06, + "loss": 0.5106, + "num_input_tokens_seen": 2727240, + "step": 9555 + }, + { + "epoch": 17.83582089552239, + "grad_norm": 0.43469202518463135, + "learning_rate": 1.7653101868349343e-06, + "loss": 0.7506, + "num_input_tokens_seen": 2728616, + "step": 9560 + }, + { + "epoch": 17.845149253731343, + "grad_norm": 0.47377434372901917, + "learning_rate": 1.750317428401066e-06, + "loss": 0.7365, + "num_input_tokens_seen": 2729928, + "step": 9565 + }, + { + "epoch": 17.854477611940297, + "grad_norm": 0.4955195188522339, + "learning_rate": 1.7353862985380027e-06, + "loss": 0.8183, + "num_input_tokens_seen": 2731272, + "step": 9570 + }, + { + "epoch": 17.863805970149254, + "grad_norm": 0.45700129866600037, + "learning_rate": 1.7205168368240986e-06, + "loss": 0.7378, + "num_input_tokens_seen": 2732872, + "step": 9575 + }, + { + "epoch": 17.87313432835821, + "grad_norm": 0.42546823620796204, + "learning_rate": 1.7057090826742505e-06, + "loss": 0.5925, + "num_input_tokens_seen": 2734760, + "step": 9580 + }, + { + "epoch": 17.882462686567163, + "grad_norm": 0.49880337715148926, + "learning_rate": 1.6909630753397716e-06, + "loss": 0.8297, + "num_input_tokens_seen": 2736104, + "step": 9585 + }, + { + "epoch": 17.89179104477612, + "grad_norm": 0.49354884028434753, + "learning_rate": 1.6762788539083086e-06, + "loss": 0.7121, + "num_input_tokens_seen": 2737512, + "step": 9590 + }, + { + "epoch": 17.901119402985074, + "grad_norm": 0.44467586278915405, + "learning_rate": 1.6616564573037342e-06, + "loss": 0.732, + "num_input_tokens_seen": 2738888, + "step": 9595 + }, + { + "epoch": 17.91044776119403, + "grad_norm": 0.7358192205429077, + "learning_rate": 1.6470959242860296e-06, + "loss": 0.9011, + "num_input_tokens_seen": 2740040, + "step": 9600 + }, + { + "epoch": 17.919776119402986, + "grad_norm": 0.3459722399711609, + "learning_rate": 1.6325972934512018e-06, + "loss": 0.6195, + "num_input_tokens_seen": 2741416, + "step": 9605 + }, + { + "epoch": 17.92910447761194, + "grad_norm": 0.3905186057090759, + "learning_rate": 1.6181606032311696e-06, + "loss": 0.4596, + "num_input_tokens_seen": 2742760, + "step": 9610 + }, + { + "epoch": 17.938432835820894, + "grad_norm": 0.6517500281333923, + "learning_rate": 1.6037858918936638e-06, + "loss": 0.7076, + "num_input_tokens_seen": 2743912, + "step": 9615 + }, + { + "epoch": 17.94776119402985, + "grad_norm": 0.4441828429698944, + "learning_rate": 1.589473197542124e-06, + "loss": 0.6186, + "num_input_tokens_seen": 2745416, + "step": 9620 + }, + { + "epoch": 17.957089552238806, + "grad_norm": 0.5332238078117371, + "learning_rate": 1.5752225581155993e-06, + "loss": 0.8379, + "num_input_tokens_seen": 2746984, + "step": 9625 + }, + { + "epoch": 17.96641791044776, + "grad_norm": 0.3403754234313965, + "learning_rate": 1.5610340113886568e-06, + "loss": 0.6485, + "num_input_tokens_seen": 2748552, + "step": 9630 + }, + { + "epoch": 17.975746268656717, + "grad_norm": 0.4957941472530365, + "learning_rate": 1.5469075949712613e-06, + "loss": 0.8664, + "num_input_tokens_seen": 2749992, + "step": 9635 + }, + { + "epoch": 17.98507462686567, + "grad_norm": 0.5368757843971252, + "learning_rate": 1.5328433463086904e-06, + "loss": 0.5613, + "num_input_tokens_seen": 2751336, + "step": 9640 + }, + { + "epoch": 17.994402985074625, + "grad_norm": 0.4908386170864105, + "learning_rate": 1.5188413026814396e-06, + "loss": 0.6361, + "num_input_tokens_seen": 2752808, + "step": 9645 + }, + { + "epoch": 18.0, + "eval_loss": 0.6946848630905151, + "eval_runtime": 2.4594, + "eval_samples_per_second": 96.773, + "eval_steps_per_second": 24.396, + "num_input_tokens_seen": 2753536, + "step": 9648 + }, + { + "epoch": 18.003731343283583, + "grad_norm": 0.6813323497772217, + "learning_rate": 1.5049015012051055e-06, + "loss": 0.6194, + "num_input_tokens_seen": 2754048, + "step": 9650 + }, + { + "epoch": 18.013059701492537, + "grad_norm": 0.4079844355583191, + "learning_rate": 1.4910239788303027e-06, + "loss": 0.6043, + "num_input_tokens_seen": 2755424, + "step": 9655 + }, + { + "epoch": 18.02238805970149, + "grad_norm": 0.3573729693889618, + "learning_rate": 1.4772087723425559e-06, + "loss": 0.703, + "num_input_tokens_seen": 2756736, + "step": 9660 + }, + { + "epoch": 18.03171641791045, + "grad_norm": 0.32669562101364136, + "learning_rate": 1.4634559183622193e-06, + "loss": 0.5608, + "num_input_tokens_seen": 2758208, + "step": 9665 + }, + { + "epoch": 18.041044776119403, + "grad_norm": 0.3174927830696106, + "learning_rate": 1.4497654533443538e-06, + "loss": 0.6379, + "num_input_tokens_seen": 2759616, + "step": 9670 + }, + { + "epoch": 18.050373134328357, + "grad_norm": 0.3567613363265991, + "learning_rate": 1.436137413578653e-06, + "loss": 0.5381, + "num_input_tokens_seen": 2760960, + "step": 9675 + }, + { + "epoch": 18.059701492537314, + "grad_norm": 0.5716961622238159, + "learning_rate": 1.4225718351893374e-06, + "loss": 0.756, + "num_input_tokens_seen": 2762240, + "step": 9680 + }, + { + "epoch": 18.06902985074627, + "grad_norm": 0.45947203040122986, + "learning_rate": 1.4090687541350538e-06, + "loss": 0.6621, + "num_input_tokens_seen": 2763712, + "step": 9685 + }, + { + "epoch": 18.078358208955223, + "grad_norm": 0.4383579194545746, + "learning_rate": 1.3956282062087933e-06, + "loss": 0.5612, + "num_input_tokens_seen": 2765056, + "step": 9690 + }, + { + "epoch": 18.08768656716418, + "grad_norm": 0.5668937563896179, + "learning_rate": 1.3822502270377762e-06, + "loss": 0.9437, + "num_input_tokens_seen": 2766528, + "step": 9695 + }, + { + "epoch": 18.097014925373134, + "grad_norm": 0.4082622826099396, + "learning_rate": 1.368934852083384e-06, + "loss": 0.6803, + "num_input_tokens_seen": 2768064, + "step": 9700 + }, + { + "epoch": 18.10634328358209, + "grad_norm": 0.5031373500823975, + "learning_rate": 1.3556821166410522e-06, + "loss": 0.6564, + "num_input_tokens_seen": 2769568, + "step": 9705 + }, + { + "epoch": 18.115671641791046, + "grad_norm": 0.47611498832702637, + "learning_rate": 1.3424920558401611e-06, + "loss": 0.5792, + "num_input_tokens_seen": 2770848, + "step": 9710 + }, + { + "epoch": 18.125, + "grad_norm": 0.4527207314968109, + "learning_rate": 1.3293647046439678e-06, + "loss": 0.5438, + "num_input_tokens_seen": 2772224, + "step": 9715 + }, + { + "epoch": 18.134328358208954, + "grad_norm": 0.6090534925460815, + "learning_rate": 1.3163000978495072e-06, + "loss": 0.5722, + "num_input_tokens_seen": 2773696, + "step": 9720 + }, + { + "epoch": 18.14365671641791, + "grad_norm": 0.4506579637527466, + "learning_rate": 1.3032982700874802e-06, + "loss": 0.5266, + "num_input_tokens_seen": 2775168, + "step": 9725 + }, + { + "epoch": 18.152985074626866, + "grad_norm": 0.31996673345565796, + "learning_rate": 1.2903592558222016e-06, + "loss": 0.695, + "num_input_tokens_seen": 2776832, + "step": 9730 + }, + { + "epoch": 18.16231343283582, + "grad_norm": 0.7853272557258606, + "learning_rate": 1.2774830893514583e-06, + "loss": 0.7484, + "num_input_tokens_seen": 2778144, + "step": 9735 + }, + { + "epoch": 18.171641791044777, + "grad_norm": 0.44254207611083984, + "learning_rate": 1.2646698048064703e-06, + "loss": 0.7066, + "num_input_tokens_seen": 2779680, + "step": 9740 + }, + { + "epoch": 18.18097014925373, + "grad_norm": 0.47875720262527466, + "learning_rate": 1.2519194361517466e-06, + "loss": 0.7921, + "num_input_tokens_seen": 2781248, + "step": 9745 + }, + { + "epoch": 18.190298507462686, + "grad_norm": 0.4199385941028595, + "learning_rate": 1.2392320171850546e-06, + "loss": 0.7151, + "num_input_tokens_seen": 2782624, + "step": 9750 + }, + { + "epoch": 18.199626865671643, + "grad_norm": 0.3117460012435913, + "learning_rate": 1.2266075815372701e-06, + "loss": 0.473, + "num_input_tokens_seen": 2784224, + "step": 9755 + }, + { + "epoch": 18.208955223880597, + "grad_norm": 0.5486754775047302, + "learning_rate": 1.2140461626723414e-06, + "loss": 0.6928, + "num_input_tokens_seen": 2785664, + "step": 9760 + }, + { + "epoch": 18.21828358208955, + "grad_norm": 0.43209728598594666, + "learning_rate": 1.2015477938871617e-06, + "loss": 0.642, + "num_input_tokens_seen": 2787072, + "step": 9765 + }, + { + "epoch": 18.22761194029851, + "grad_norm": 0.35975730419158936, + "learning_rate": 1.1891125083114962e-06, + "loss": 0.6552, + "num_input_tokens_seen": 2788480, + "step": 9770 + }, + { + "epoch": 18.236940298507463, + "grad_norm": 0.5268393754959106, + "learning_rate": 1.1767403389079057e-06, + "loss": 0.8629, + "num_input_tokens_seen": 2789792, + "step": 9775 + }, + { + "epoch": 18.246268656716417, + "grad_norm": 0.3508278727531433, + "learning_rate": 1.164431318471626e-06, + "loss": 0.6859, + "num_input_tokens_seen": 2791200, + "step": 9780 + }, + { + "epoch": 18.255597014925375, + "grad_norm": 0.38433837890625, + "learning_rate": 1.1521854796305242e-06, + "loss": 0.5254, + "num_input_tokens_seen": 2792896, + "step": 9785 + }, + { + "epoch": 18.26492537313433, + "grad_norm": 0.48293325304985046, + "learning_rate": 1.1400028548449821e-06, + "loss": 0.8092, + "num_input_tokens_seen": 2794336, + "step": 9790 + }, + { + "epoch": 18.274253731343283, + "grad_norm": 0.4614972770214081, + "learning_rate": 1.1278834764078123e-06, + "loss": 0.7838, + "num_input_tokens_seen": 2795744, + "step": 9795 + }, + { + "epoch": 18.28358208955224, + "grad_norm": 0.4968569874763489, + "learning_rate": 1.1158273764441868e-06, + "loss": 0.7545, + "num_input_tokens_seen": 2797152, + "step": 9800 + }, + { + "epoch": 18.292910447761194, + "grad_norm": 0.4263819456100464, + "learning_rate": 1.103834586911534e-06, + "loss": 0.7285, + "num_input_tokens_seen": 2798720, + "step": 9805 + }, + { + "epoch": 18.30223880597015, + "grad_norm": 0.4810596704483032, + "learning_rate": 1.0919051395994778e-06, + "loss": 0.7454, + "num_input_tokens_seen": 2800000, + "step": 9810 + }, + { + "epoch": 18.311567164179106, + "grad_norm": 0.4043166935443878, + "learning_rate": 1.0800390661297261e-06, + "loss": 0.5528, + "num_input_tokens_seen": 2801408, + "step": 9815 + }, + { + "epoch": 18.32089552238806, + "grad_norm": 0.4124656617641449, + "learning_rate": 1.0682363979560046e-06, + "loss": 0.9416, + "num_input_tokens_seen": 2802784, + "step": 9820 + }, + { + "epoch": 18.330223880597014, + "grad_norm": 0.5201969146728516, + "learning_rate": 1.056497166363976e-06, + "loss": 0.6448, + "num_input_tokens_seen": 2804096, + "step": 9825 + }, + { + "epoch": 18.33955223880597, + "grad_norm": 0.4906717836856842, + "learning_rate": 1.0448214024711384e-06, + "loss": 0.8478, + "num_input_tokens_seen": 2805504, + "step": 9830 + }, + { + "epoch": 18.348880597014926, + "grad_norm": 0.610000729560852, + "learning_rate": 1.0332091372267566e-06, + "loss": 0.5772, + "num_input_tokens_seen": 2806944, + "step": 9835 + }, + { + "epoch": 18.35820895522388, + "grad_norm": 0.510421872138977, + "learning_rate": 1.0216604014117837e-06, + "loss": 0.616, + "num_input_tokens_seen": 2808384, + "step": 9840 + }, + { + "epoch": 18.367537313432837, + "grad_norm": 0.5854883790016174, + "learning_rate": 1.0101752256387682e-06, + "loss": 0.8747, + "num_input_tokens_seen": 2809888, + "step": 9845 + }, + { + "epoch": 18.37686567164179, + "grad_norm": 0.3870214521884918, + "learning_rate": 9.98753640351785e-07, + "loss": 0.573, + "num_input_tokens_seen": 2811424, + "step": 9850 + }, + { + "epoch": 18.386194029850746, + "grad_norm": 0.3189608156681061, + "learning_rate": 9.873956758263359e-07, + "loss": 0.647, + "num_input_tokens_seen": 2812736, + "step": 9855 + }, + { + "epoch": 18.395522388059703, + "grad_norm": 0.5276429057121277, + "learning_rate": 9.76101362169296e-07, + "loss": 0.5383, + "num_input_tokens_seen": 2814304, + "step": 9860 + }, + { + "epoch": 18.404850746268657, + "grad_norm": 0.47346940636634827, + "learning_rate": 9.64870729318809e-07, + "loss": 0.7062, + "num_input_tokens_seen": 2815744, + "step": 9865 + }, + { + "epoch": 18.41417910447761, + "grad_norm": 0.4616207182407379, + "learning_rate": 9.537038070442206e-07, + "loss": 0.7897, + "num_input_tokens_seen": 2817120, + "step": 9870 + }, + { + "epoch": 18.423507462686565, + "grad_norm": 0.4814740717411041, + "learning_rate": 9.42600624945994e-07, + "loss": 0.518, + "num_input_tokens_seen": 2818560, + "step": 9875 + }, + { + "epoch": 18.432835820895523, + "grad_norm": 0.3993942141532898, + "learning_rate": 9.315612124556477e-07, + "loss": 0.661, + "num_input_tokens_seen": 2820128, + "step": 9880 + }, + { + "epoch": 18.442164179104477, + "grad_norm": 0.5481740236282349, + "learning_rate": 9.205855988356466e-07, + "loss": 0.7235, + "num_input_tokens_seen": 2821664, + "step": 9885 + }, + { + "epoch": 18.451492537313435, + "grad_norm": 0.508634626865387, + "learning_rate": 9.096738131793542e-07, + "loss": 0.7774, + "num_input_tokens_seen": 2823008, + "step": 9890 + }, + { + "epoch": 18.46082089552239, + "grad_norm": 0.49494069814682007, + "learning_rate": 8.988258844109393e-07, + "loss": 0.7428, + "num_input_tokens_seen": 2824448, + "step": 9895 + }, + { + "epoch": 18.470149253731343, + "grad_norm": 0.4728832244873047, + "learning_rate": 8.880418412853059e-07, + "loss": 0.6033, + "num_input_tokens_seen": 2825888, + "step": 9900 + }, + { + "epoch": 18.479477611940297, + "grad_norm": 0.40324246883392334, + "learning_rate": 8.773217123880073e-07, + "loss": 0.5768, + "num_input_tokens_seen": 2827392, + "step": 9905 + }, + { + "epoch": 18.488805970149254, + "grad_norm": 0.4396240711212158, + "learning_rate": 8.666655261351908e-07, + "loss": 0.7755, + "num_input_tokens_seen": 2828736, + "step": 9910 + }, + { + "epoch": 18.49813432835821, + "grad_norm": 0.5001472234725952, + "learning_rate": 8.560733107734947e-07, + "loss": 0.657, + "num_input_tokens_seen": 2830336, + "step": 9915 + }, + { + "epoch": 18.507462686567163, + "grad_norm": 0.7445406913757324, + "learning_rate": 8.455450943799958e-07, + "loss": 0.7193, + "num_input_tokens_seen": 2831616, + "step": 9920 + }, + { + "epoch": 18.51679104477612, + "grad_norm": 0.5164788365364075, + "learning_rate": 8.35080904862126e-07, + "loss": 0.887, + "num_input_tokens_seen": 2832928, + "step": 9925 + }, + { + "epoch": 18.526119402985074, + "grad_norm": 0.4102502763271332, + "learning_rate": 8.246807699576032e-07, + "loss": 0.4844, + "num_input_tokens_seen": 2834560, + "step": 9930 + }, + { + "epoch": 18.53544776119403, + "grad_norm": 0.4892178177833557, + "learning_rate": 8.143447172343471e-07, + "loss": 0.723, + "num_input_tokens_seen": 2835904, + "step": 9935 + }, + { + "epoch": 18.544776119402986, + "grad_norm": 0.44586893916130066, + "learning_rate": 8.040727740904113e-07, + "loss": 0.5678, + "num_input_tokens_seen": 2837536, + "step": 9940 + }, + { + "epoch": 18.55410447761194, + "grad_norm": 0.6013088226318359, + "learning_rate": 7.938649677539267e-07, + "loss": 0.4664, + "num_input_tokens_seen": 2838816, + "step": 9945 + }, + { + "epoch": 18.563432835820894, + "grad_norm": 0.6007289886474609, + "learning_rate": 7.837213252829989e-07, + "loss": 0.5293, + "num_input_tokens_seen": 2840128, + "step": 9950 + }, + { + "epoch": 18.57276119402985, + "grad_norm": 0.30295974016189575, + "learning_rate": 7.736418735656586e-07, + "loss": 0.8258, + "num_input_tokens_seen": 2841632, + "step": 9955 + }, + { + "epoch": 18.582089552238806, + "grad_norm": 0.3543384373188019, + "learning_rate": 7.636266393197866e-07, + "loss": 0.6839, + "num_input_tokens_seen": 2843040, + "step": 9960 + }, + { + "epoch": 18.59141791044776, + "grad_norm": 0.580561637878418, + "learning_rate": 7.536756490930358e-07, + "loss": 0.4837, + "num_input_tokens_seen": 2844512, + "step": 9965 + }, + { + "epoch": 18.600746268656717, + "grad_norm": 0.31025829911231995, + "learning_rate": 7.437889292627787e-07, + "loss": 0.5624, + "num_input_tokens_seen": 2846048, + "step": 9970 + }, + { + "epoch": 18.61007462686567, + "grad_norm": 0.48535633087158203, + "learning_rate": 7.339665060360018e-07, + "loss": 0.7062, + "num_input_tokens_seen": 2847488, + "step": 9975 + }, + { + "epoch": 18.619402985074625, + "grad_norm": 0.36098381876945496, + "learning_rate": 7.24208405449281e-07, + "loss": 0.7294, + "num_input_tokens_seen": 2848928, + "step": 9980 + }, + { + "epoch": 18.628731343283583, + "grad_norm": 1.0840553045272827, + "learning_rate": 7.145146533686725e-07, + "loss": 0.9256, + "num_input_tokens_seen": 2850176, + "step": 9985 + }, + { + "epoch": 18.638059701492537, + "grad_norm": 0.4708701968193054, + "learning_rate": 7.048852754896806e-07, + "loss": 0.5916, + "num_input_tokens_seen": 2851648, + "step": 9990 + }, + { + "epoch": 18.64738805970149, + "grad_norm": 0.2479024976491928, + "learning_rate": 6.953202973371514e-07, + "loss": 0.4239, + "num_input_tokens_seen": 2853248, + "step": 9995 + }, + { + "epoch": 18.65671641791045, + "grad_norm": 0.4259677827358246, + "learning_rate": 6.858197442652369e-07, + "loss": 0.6579, + "num_input_tokens_seen": 2854656, + "step": 10000 + }, + { + "epoch": 18.666044776119403, + "grad_norm": 0.40276631712913513, + "learning_rate": 6.763836414573232e-07, + "loss": 0.7626, + "num_input_tokens_seen": 2856000, + "step": 10005 + }, + { + "epoch": 18.675373134328357, + "grad_norm": 0.6027032136917114, + "learning_rate": 6.670120139259328e-07, + "loss": 0.5963, + "num_input_tokens_seen": 2857408, + "step": 10010 + }, + { + "epoch": 18.684701492537314, + "grad_norm": 0.48139068484306335, + "learning_rate": 6.577048865127028e-07, + "loss": 0.6501, + "num_input_tokens_seen": 2858784, + "step": 10015 + }, + { + "epoch": 18.69402985074627, + "grad_norm": 0.4329470694065094, + "learning_rate": 6.484622838882903e-07, + "loss": 0.6134, + "num_input_tokens_seen": 2860288, + "step": 10020 + }, + { + "epoch": 18.703358208955223, + "grad_norm": 0.368840754032135, + "learning_rate": 6.392842305523172e-07, + "loss": 0.6841, + "num_input_tokens_seen": 2861664, + "step": 10025 + }, + { + "epoch": 18.71268656716418, + "grad_norm": 0.41267189383506775, + "learning_rate": 6.301707508332977e-07, + "loss": 0.7465, + "num_input_tokens_seen": 2863168, + "step": 10030 + }, + { + "epoch": 18.722014925373134, + "grad_norm": 0.5105572938919067, + "learning_rate": 6.2112186888858e-07, + "loss": 1.2308, + "num_input_tokens_seen": 2864576, + "step": 10035 + }, + { + "epoch": 18.73134328358209, + "grad_norm": 0.5357996821403503, + "learning_rate": 6.121376087042913e-07, + "loss": 0.8963, + "num_input_tokens_seen": 2866080, + "step": 10040 + }, + { + "epoch": 18.740671641791046, + "grad_norm": 0.5947056412696838, + "learning_rate": 6.032179940952509e-07, + "loss": 0.7163, + "num_input_tokens_seen": 2867488, + "step": 10045 + }, + { + "epoch": 18.75, + "grad_norm": 0.42745739221572876, + "learning_rate": 5.943630487049295e-07, + "loss": 0.7495, + "num_input_tokens_seen": 2868832, + "step": 10050 + }, + { + "epoch": 18.759328358208954, + "grad_norm": 0.33982053399086, + "learning_rate": 5.855727960053653e-07, + "loss": 0.648, + "num_input_tokens_seen": 2870208, + "step": 10055 + }, + { + "epoch": 18.76865671641791, + "grad_norm": 0.35320088267326355, + "learning_rate": 5.768472592971308e-07, + "loss": 0.6361, + "num_input_tokens_seen": 2871840, + "step": 10060 + }, + { + "epoch": 18.777985074626866, + "grad_norm": 0.4263074994087219, + "learning_rate": 5.681864617092414e-07, + "loss": 0.5297, + "num_input_tokens_seen": 2873088, + "step": 10065 + }, + { + "epoch": 18.78731343283582, + "grad_norm": 0.7227181196212769, + "learning_rate": 5.595904261991109e-07, + "loss": 0.5447, + "num_input_tokens_seen": 2874688, + "step": 10070 + }, + { + "epoch": 18.796641791044777, + "grad_norm": 0.4648990035057068, + "learning_rate": 5.510591755524874e-07, + "loss": 0.5077, + "num_input_tokens_seen": 2876064, + "step": 10075 + }, + { + "epoch": 18.80597014925373, + "grad_norm": 0.4191240072250366, + "learning_rate": 5.425927323833902e-07, + "loss": 0.4801, + "num_input_tokens_seen": 2877472, + "step": 10080 + }, + { + "epoch": 18.815298507462686, + "grad_norm": 0.4042437672615051, + "learning_rate": 5.341911191340504e-07, + "loss": 0.5506, + "num_input_tokens_seen": 2878912, + "step": 10085 + }, + { + "epoch": 18.824626865671643, + "grad_norm": 0.4710797369480133, + "learning_rate": 5.258543580748565e-07, + "loss": 0.6319, + "num_input_tokens_seen": 2880352, + "step": 10090 + }, + { + "epoch": 18.833955223880597, + "grad_norm": 0.4908454120159149, + "learning_rate": 5.175824713042926e-07, + "loss": 0.579, + "num_input_tokens_seen": 2881824, + "step": 10095 + }, + { + "epoch": 18.84328358208955, + "grad_norm": 0.3901980519294739, + "learning_rate": 5.093754807488693e-07, + "loss": 0.6428, + "num_input_tokens_seen": 2883456, + "step": 10100 + }, + { + "epoch": 18.85261194029851, + "grad_norm": 0.36739906668663025, + "learning_rate": 5.01233408163082e-07, + "loss": 0.8422, + "num_input_tokens_seen": 2884800, + "step": 10105 + }, + { + "epoch": 18.861940298507463, + "grad_norm": 0.47179150581359863, + "learning_rate": 4.931562751293528e-07, + "loss": 0.515, + "num_input_tokens_seen": 2886368, + "step": 10110 + }, + { + "epoch": 18.871268656716417, + "grad_norm": 0.5761358141899109, + "learning_rate": 4.851441030579523e-07, + "loss": 0.7592, + "num_input_tokens_seen": 2887744, + "step": 10115 + }, + { + "epoch": 18.880597014925375, + "grad_norm": 0.509718656539917, + "learning_rate": 4.771969131869669e-07, + "loss": 0.64, + "num_input_tokens_seen": 2889184, + "step": 10120 + }, + { + "epoch": 18.88992537313433, + "grad_norm": 0.5018926858901978, + "learning_rate": 4.6931472658223176e-07, + "loss": 0.808, + "num_input_tokens_seen": 2890592, + "step": 10125 + }, + { + "epoch": 18.899253731343283, + "grad_norm": 0.6670432090759277, + "learning_rate": 4.614975641372754e-07, + "loss": 0.7487, + "num_input_tokens_seen": 2891936, + "step": 10130 + }, + { + "epoch": 18.90858208955224, + "grad_norm": 0.49521708488464355, + "learning_rate": 4.5374544657326157e-07, + "loss": 0.7323, + "num_input_tokens_seen": 2893408, + "step": 10135 + }, + { + "epoch": 18.917910447761194, + "grad_norm": 0.5268117189407349, + "learning_rate": 4.460583944389418e-07, + "loss": 0.9823, + "num_input_tokens_seen": 2894784, + "step": 10140 + }, + { + "epoch": 18.92723880597015, + "grad_norm": 0.39197465777397156, + "learning_rate": 4.3843642811059737e-07, + "loss": 0.852, + "num_input_tokens_seen": 2896160, + "step": 10145 + }, + { + "epoch": 18.936567164179106, + "grad_norm": 0.6142540574073792, + "learning_rate": 4.3087956779198356e-07, + "loss": 0.4968, + "num_input_tokens_seen": 2897408, + "step": 10150 + }, + { + "epoch": 18.94589552238806, + "grad_norm": 0.491021990776062, + "learning_rate": 4.2338783351427156e-07, + "loss": 0.6075, + "num_input_tokens_seen": 2899008, + "step": 10155 + }, + { + "epoch": 18.955223880597014, + "grad_norm": 0.3820241391658783, + "learning_rate": 4.159612451360151e-07, + "loss": 0.469, + "num_input_tokens_seen": 2900448, + "step": 10160 + }, + { + "epoch": 18.96455223880597, + "grad_norm": 0.37915003299713135, + "learning_rate": 4.085998223430698e-07, + "loss": 0.569, + "num_input_tokens_seen": 2901984, + "step": 10165 + }, + { + "epoch": 18.973880597014926, + "grad_norm": 0.44635042548179626, + "learning_rate": 4.013035846485658e-07, + "loss": 0.9144, + "num_input_tokens_seen": 2903392, + "step": 10170 + }, + { + "epoch": 18.98320895522388, + "grad_norm": 0.4836762249469757, + "learning_rate": 3.940725513928323e-07, + "loss": 0.638, + "num_input_tokens_seen": 2904736, + "step": 10175 + }, + { + "epoch": 18.992537313432837, + "grad_norm": 0.48978736996650696, + "learning_rate": 3.8690674174337305e-07, + "loss": 0.8247, + "num_input_tokens_seen": 2905984, + "step": 10180 + }, + { + "epoch": 19.0, + "eval_loss": 0.6947132349014282, + "eval_runtime": 2.459, + "eval_samples_per_second": 96.788, + "eval_steps_per_second": 24.4, + "num_input_tokens_seen": 2906984, + "step": 10184 + }, + { + "epoch": 19.00186567164179, + "grad_norm": 0.3499637842178345, + "learning_rate": 3.7980617469479953e-07, + "loss": 0.6705, + "num_input_tokens_seen": 2907240, + "step": 10185 + }, + { + "epoch": 19.011194029850746, + "grad_norm": 0.4423954486846924, + "learning_rate": 3.7277086906877256e-07, + "loss": 0.6323, + "num_input_tokens_seen": 2908680, + "step": 10190 + }, + { + "epoch": 19.020522388059703, + "grad_norm": 0.30681490898132324, + "learning_rate": 3.65800843513972e-07, + "loss": 0.6044, + "num_input_tokens_seen": 2910248, + "step": 10195 + }, + { + "epoch": 19.029850746268657, + "grad_norm": 0.45089399814605713, + "learning_rate": 3.588961165060356e-07, + "loss": 0.7154, + "num_input_tokens_seen": 2911720, + "step": 10200 + }, + { + "epoch": 19.03917910447761, + "grad_norm": 0.5014017224311829, + "learning_rate": 3.5205670634751163e-07, + "loss": 0.7465, + "num_input_tokens_seen": 2913128, + "step": 10205 + }, + { + "epoch": 19.04850746268657, + "grad_norm": 0.5124428272247314, + "learning_rate": 3.452826311678148e-07, + "loss": 0.7233, + "num_input_tokens_seen": 2914440, + "step": 10210 + }, + { + "epoch": 19.057835820895523, + "grad_norm": 0.4995649456977844, + "learning_rate": 3.3857390892316764e-07, + "loss": 0.6746, + "num_input_tokens_seen": 2915880, + "step": 10215 + }, + { + "epoch": 19.067164179104477, + "grad_norm": 0.5378211140632629, + "learning_rate": 3.319305573965703e-07, + "loss": 0.9204, + "num_input_tokens_seen": 2917320, + "step": 10220 + }, + { + "epoch": 19.07649253731343, + "grad_norm": 0.5839022994041443, + "learning_rate": 3.253525941977309e-07, + "loss": 0.6955, + "num_input_tokens_seen": 2918952, + "step": 10225 + }, + { + "epoch": 19.08582089552239, + "grad_norm": 0.3617372512817383, + "learning_rate": 3.1884003676303786e-07, + "loss": 0.6313, + "num_input_tokens_seen": 2920328, + "step": 10230 + }, + { + "epoch": 19.095149253731343, + "grad_norm": 0.3203364908695221, + "learning_rate": 3.1239290235550724e-07, + "loss": 0.7169, + "num_input_tokens_seen": 2921736, + "step": 10235 + }, + { + "epoch": 19.104477611940297, + "grad_norm": 0.46454712748527527, + "learning_rate": 3.0601120806473535e-07, + "loss": 0.7417, + "num_input_tokens_seen": 2923272, + "step": 10240 + }, + { + "epoch": 19.113805970149254, + "grad_norm": 0.5234792828559875, + "learning_rate": 2.9969497080685196e-07, + "loss": 0.8859, + "num_input_tokens_seen": 2924744, + "step": 10245 + }, + { + "epoch": 19.12313432835821, + "grad_norm": 0.34602123498916626, + "learning_rate": 2.934442073244809e-07, + "loss": 0.5709, + "num_input_tokens_seen": 2926216, + "step": 10250 + }, + { + "epoch": 19.132462686567163, + "grad_norm": 0.5095165967941284, + "learning_rate": 2.87258934186696e-07, + "loss": 0.7567, + "num_input_tokens_seen": 2927720, + "step": 10255 + }, + { + "epoch": 19.14179104477612, + "grad_norm": 0.37290799617767334, + "learning_rate": 2.8113916778896575e-07, + "loss": 0.6973, + "num_input_tokens_seen": 2929160, + "step": 10260 + }, + { + "epoch": 19.151119402985074, + "grad_norm": 0.4673911929130554, + "learning_rate": 2.750849243531223e-07, + "loss": 0.6811, + "num_input_tokens_seen": 2930472, + "step": 10265 + }, + { + "epoch": 19.16044776119403, + "grad_norm": 0.33322834968566895, + "learning_rate": 2.6909621992731726e-07, + "loss": 0.6818, + "num_input_tokens_seen": 2931784, + "step": 10270 + }, + { + "epoch": 19.169776119402986, + "grad_norm": 0.4296891391277313, + "learning_rate": 2.6317307038597196e-07, + "loss": 0.7448, + "num_input_tokens_seen": 2933224, + "step": 10275 + }, + { + "epoch": 19.17910447761194, + "grad_norm": 0.5420965552330017, + "learning_rate": 2.573154914297438e-07, + "loss": 0.71, + "num_input_tokens_seen": 2934600, + "step": 10280 + }, + { + "epoch": 19.188432835820894, + "grad_norm": 0.7150471806526184, + "learning_rate": 2.515234985854736e-07, + "loss": 0.7602, + "num_input_tokens_seen": 2936168, + "step": 10285 + }, + { + "epoch": 19.19776119402985, + "grad_norm": 0.5051119923591614, + "learning_rate": 2.45797107206161e-07, + "loss": 0.8304, + "num_input_tokens_seen": 2937864, + "step": 10290 + }, + { + "epoch": 19.207089552238806, + "grad_norm": 0.45255935192108154, + "learning_rate": 2.401363324709055e-07, + "loss": 0.5908, + "num_input_tokens_seen": 2939272, + "step": 10295 + }, + { + "epoch": 19.21641791044776, + "grad_norm": 0.3827902674674988, + "learning_rate": 2.3454118938487367e-07, + "loss": 0.4307, + "num_input_tokens_seen": 2940616, + "step": 10300 + }, + { + "epoch": 19.225746268656717, + "grad_norm": 0.4523323178291321, + "learning_rate": 2.2901169277927127e-07, + "loss": 0.6784, + "num_input_tokens_seen": 2941960, + "step": 10305 + }, + { + "epoch": 19.23507462686567, + "grad_norm": 0.41109564900398254, + "learning_rate": 2.2354785731128482e-07, + "loss": 0.5246, + "num_input_tokens_seen": 2943400, + "step": 10310 + }, + { + "epoch": 19.244402985074625, + "grad_norm": 0.3078072667121887, + "learning_rate": 2.18149697464054e-07, + "loss": 0.4646, + "num_input_tokens_seen": 2944936, + "step": 10315 + }, + { + "epoch": 19.253731343283583, + "grad_norm": 0.5048292279243469, + "learning_rate": 2.128172275466217e-07, + "loss": 0.8452, + "num_input_tokens_seen": 2946312, + "step": 10320 + }, + { + "epoch": 19.263059701492537, + "grad_norm": 0.46120312809944153, + "learning_rate": 2.0755046169392e-07, + "loss": 0.7365, + "num_input_tokens_seen": 2947624, + "step": 10325 + }, + { + "epoch": 19.27238805970149, + "grad_norm": 0.8852548599243164, + "learning_rate": 2.0234941386670925e-07, + "loss": 0.5902, + "num_input_tokens_seen": 2949160, + "step": 10330 + }, + { + "epoch": 19.28171641791045, + "grad_norm": 0.5784152150154114, + "learning_rate": 1.9721409785154466e-07, + "loss": 0.7088, + "num_input_tokens_seen": 2950536, + "step": 10335 + }, + { + "epoch": 19.291044776119403, + "grad_norm": 0.5544753670692444, + "learning_rate": 1.9214452726075137e-07, + "loss": 0.5402, + "num_input_tokens_seen": 2951848, + "step": 10340 + }, + { + "epoch": 19.300373134328357, + "grad_norm": 0.5361091494560242, + "learning_rate": 1.871407155323801e-07, + "loss": 0.5823, + "num_input_tokens_seen": 2953480, + "step": 10345 + }, + { + "epoch": 19.309701492537314, + "grad_norm": 0.33521825075149536, + "learning_rate": 1.8220267593017092e-07, + "loss": 0.5044, + "num_input_tokens_seen": 2955016, + "step": 10350 + }, + { + "epoch": 19.31902985074627, + "grad_norm": 0.6716789603233337, + "learning_rate": 1.7733042154352008e-07, + "loss": 0.6817, + "num_input_tokens_seen": 2956552, + "step": 10355 + }, + { + "epoch": 19.328358208955223, + "grad_norm": 0.3836686313152313, + "learning_rate": 1.7252396528744663e-07, + "loss": 0.5955, + "num_input_tokens_seen": 2957832, + "step": 10360 + }, + { + "epoch": 19.33768656716418, + "grad_norm": 0.3139139711856842, + "learning_rate": 1.6778331990255914e-07, + "loss": 0.6036, + "num_input_tokens_seen": 2959336, + "step": 10365 + }, + { + "epoch": 19.347014925373134, + "grad_norm": 0.4057820737361908, + "learning_rate": 1.6310849795500848e-07, + "loss": 0.8523, + "num_input_tokens_seen": 2960712, + "step": 10370 + }, + { + "epoch": 19.35634328358209, + "grad_norm": 0.581243097782135, + "learning_rate": 1.584995118364796e-07, + "loss": 0.7317, + "num_input_tokens_seen": 2962024, + "step": 10375 + }, + { + "epoch": 19.365671641791046, + "grad_norm": 0.5093084573745728, + "learning_rate": 1.5395637376413585e-07, + "loss": 0.666, + "num_input_tokens_seen": 2963688, + "step": 10380 + }, + { + "epoch": 19.375, + "grad_norm": 0.3599047362804413, + "learning_rate": 1.4947909578059971e-07, + "loss": 0.692, + "num_input_tokens_seen": 2965032, + "step": 10385 + }, + { + "epoch": 19.384328358208954, + "grad_norm": 0.5187100768089294, + "learning_rate": 1.4506768975391382e-07, + "loss": 0.8505, + "num_input_tokens_seen": 2966376, + "step": 10390 + }, + { + "epoch": 19.39365671641791, + "grad_norm": 0.3339211642742157, + "learning_rate": 1.4072216737751055e-07, + "loss": 0.8438, + "num_input_tokens_seen": 2967656, + "step": 10395 + }, + { + "epoch": 19.402985074626866, + "grad_norm": 0.4266071021556854, + "learning_rate": 1.3644254017018964e-07, + "loss": 0.8172, + "num_input_tokens_seen": 2969224, + "step": 10400 + }, + { + "epoch": 19.41231343283582, + "grad_norm": 0.30329084396362305, + "learning_rate": 1.3222881947607123e-07, + "loss": 0.6583, + "num_input_tokens_seen": 2970632, + "step": 10405 + }, + { + "epoch": 19.421641791044777, + "grad_norm": 0.520057201385498, + "learning_rate": 1.280810164645846e-07, + "loss": 0.6038, + "num_input_tokens_seen": 2972104, + "step": 10410 + }, + { + "epoch": 19.43097014925373, + "grad_norm": 0.2685432434082031, + "learning_rate": 1.2399914213042373e-07, + "loss": 0.5608, + "num_input_tokens_seen": 2973608, + "step": 10415 + }, + { + "epoch": 19.440298507462686, + "grad_norm": 0.4520004987716675, + "learning_rate": 1.1998320729352252e-07, + "loss": 0.6067, + "num_input_tokens_seen": 2974984, + "step": 10420 + }, + { + "epoch": 19.449626865671643, + "grad_norm": 0.3218950629234314, + "learning_rate": 1.160332225990296e-07, + "loss": 0.4829, + "num_input_tokens_seen": 2976712, + "step": 10425 + }, + { + "epoch": 19.458955223880597, + "grad_norm": 0.4655058681964874, + "learning_rate": 1.1214919851728068e-07, + "loss": 0.787, + "num_input_tokens_seen": 2978056, + "step": 10430 + }, + { + "epoch": 19.46828358208955, + "grad_norm": 0.38103121519088745, + "learning_rate": 1.0833114534376798e-07, + "loss": 0.6665, + "num_input_tokens_seen": 2979432, + "step": 10435 + }, + { + "epoch": 19.47761194029851, + "grad_norm": 0.41008391976356506, + "learning_rate": 1.0457907319909865e-07, + "loss": 0.7488, + "num_input_tokens_seen": 2980904, + "step": 10440 + }, + { + "epoch": 19.486940298507463, + "grad_norm": 0.3105763792991638, + "learning_rate": 1.0089299202900304e-07, + "loss": 0.6916, + "num_input_tokens_seen": 2982536, + "step": 10445 + }, + { + "epoch": 19.496268656716417, + "grad_norm": 0.7302044034004211, + "learning_rate": 9.727291160427366e-08, + "loss": 0.9073, + "num_input_tokens_seen": 2984040, + "step": 10450 + }, + { + "epoch": 19.505597014925375, + "grad_norm": 0.4220292866230011, + "learning_rate": 9.371884152075683e-08, + "loss": 0.5175, + "num_input_tokens_seen": 2985608, + "step": 10455 + }, + { + "epoch": 19.51492537313433, + "grad_norm": 0.6388876438140869, + "learning_rate": 9.023079119932498e-08, + "loss": 0.5744, + "num_input_tokens_seen": 2987080, + "step": 10460 + }, + { + "epoch": 19.524253731343283, + "grad_norm": 0.6417595744132996, + "learning_rate": 8.680876988584608e-08, + "loss": 0.8774, + "num_input_tokens_seen": 2988360, + "step": 10465 + }, + { + "epoch": 19.53358208955224, + "grad_norm": 0.5853652954101562, + "learning_rate": 8.345278665116974e-08, + "loss": 0.7055, + "num_input_tokens_seen": 2989960, + "step": 10470 + }, + { + "epoch": 19.542910447761194, + "grad_norm": 0.508922815322876, + "learning_rate": 8.01628503910884e-08, + "loss": 0.7468, + "num_input_tokens_seen": 2991304, + "step": 10475 + }, + { + "epoch": 19.55223880597015, + "grad_norm": 0.44163307547569275, + "learning_rate": 7.693896982632898e-08, + "loss": 0.7607, + "num_input_tokens_seen": 2992584, + "step": 10480 + }, + { + "epoch": 19.561567164179106, + "grad_norm": 0.5154439806938171, + "learning_rate": 7.378115350251957e-08, + "loss": 0.6805, + "num_input_tokens_seen": 2994216, + "step": 10485 + }, + { + "epoch": 19.57089552238806, + "grad_norm": 0.47007569670677185, + "learning_rate": 7.068940979017003e-08, + "loss": 0.672, + "num_input_tokens_seen": 2995560, + "step": 10490 + }, + { + "epoch": 19.580223880597014, + "grad_norm": 0.3846695125102997, + "learning_rate": 6.766374688464971e-08, + "loss": 0.609, + "num_input_tokens_seen": 2997032, + "step": 10495 + }, + { + "epoch": 19.58955223880597, + "grad_norm": 0.3383010923862457, + "learning_rate": 6.470417280616814e-08, + "loss": 0.614, + "num_input_tokens_seen": 2998344, + "step": 10500 + }, + { + "epoch": 19.598880597014926, + "grad_norm": 0.4106837511062622, + "learning_rate": 6.181069539974716e-08, + "loss": 1.0406, + "num_input_tokens_seen": 2999592, + "step": 10505 + }, + { + "epoch": 19.60820895522388, + "grad_norm": 0.25111353397369385, + "learning_rate": 5.898332233520987e-08, + "loss": 0.4992, + "num_input_tokens_seen": 3001064, + "step": 10510 + }, + { + "epoch": 19.617537313432837, + "grad_norm": 0.33857032656669617, + "learning_rate": 5.622206110714734e-08, + "loss": 0.5235, + "num_input_tokens_seen": 3002344, + "step": 10515 + }, + { + "epoch": 19.62686567164179, + "grad_norm": 0.733525812625885, + "learning_rate": 5.352691903491303e-08, + "loss": 0.707, + "num_input_tokens_seen": 3003592, + "step": 10520 + }, + { + "epoch": 19.636194029850746, + "grad_norm": 0.3097364008426666, + "learning_rate": 5.089790326259225e-08, + "loss": 0.7635, + "num_input_tokens_seen": 3004840, + "step": 10525 + }, + { + "epoch": 19.645522388059703, + "grad_norm": 0.5334185361862183, + "learning_rate": 4.83350207589911e-08, + "loss": 0.7534, + "num_input_tokens_seen": 3006120, + "step": 10530 + }, + { + "epoch": 19.654850746268657, + "grad_norm": 0.6511064171791077, + "learning_rate": 4.583827831761145e-08, + "loss": 0.7359, + "num_input_tokens_seen": 3007560, + "step": 10535 + }, + { + "epoch": 19.66417910447761, + "grad_norm": 0.5040066838264465, + "learning_rate": 4.340768255663708e-08, + "loss": 0.8041, + "num_input_tokens_seen": 3008872, + "step": 10540 + }, + { + "epoch": 19.673507462686565, + "grad_norm": 0.5337660312652588, + "learning_rate": 4.1043239918914233e-08, + "loss": 0.7634, + "num_input_tokens_seen": 3010184, + "step": 10545 + }, + { + "epoch": 19.682835820895523, + "grad_norm": 0.393597811460495, + "learning_rate": 3.8744956671937784e-08, + "loss": 0.5307, + "num_input_tokens_seen": 3011560, + "step": 10550 + }, + { + "epoch": 19.692164179104477, + "grad_norm": 0.5080323219299316, + "learning_rate": 3.6512838907828974e-08, + "loss": 0.7157, + "num_input_tokens_seen": 3013160, + "step": 10555 + }, + { + "epoch": 19.701492537313435, + "grad_norm": 0.5119672417640686, + "learning_rate": 3.4346892543321576e-08, + "loss": 0.8588, + "num_input_tokens_seen": 3014632, + "step": 10560 + }, + { + "epoch": 19.71082089552239, + "grad_norm": 0.4590912163257599, + "learning_rate": 3.224712331975077e-08, + "loss": 0.5909, + "num_input_tokens_seen": 3015976, + "step": 10565 + }, + { + "epoch": 19.720149253731343, + "grad_norm": 0.5334969162940979, + "learning_rate": 3.021353680303096e-08, + "loss": 0.6945, + "num_input_tokens_seen": 3017384, + "step": 10570 + }, + { + "epoch": 19.729477611940297, + "grad_norm": 0.9189899563789368, + "learning_rate": 2.82461383836502e-08, + "loss": 0.5384, + "num_input_tokens_seen": 3018888, + "step": 10575 + }, + { + "epoch": 19.738805970149254, + "grad_norm": 0.5642399191856384, + "learning_rate": 2.634493327663967e-08, + "loss": 0.8826, + "num_input_tokens_seen": 3020232, + "step": 10580 + }, + { + "epoch": 19.74813432835821, + "grad_norm": 0.49342435598373413, + "learning_rate": 2.450992652157924e-08, + "loss": 0.8419, + "num_input_tokens_seen": 3021640, + "step": 10585 + }, + { + "epoch": 19.757462686567163, + "grad_norm": 0.27629271149635315, + "learning_rate": 2.2741122982569694e-08, + "loss": 0.6776, + "num_input_tokens_seen": 3023016, + "step": 10590 + }, + { + "epoch": 19.76679104477612, + "grad_norm": 0.4026658833026886, + "learning_rate": 2.1038527348229974e-08, + "loss": 0.7002, + "num_input_tokens_seen": 3024456, + "step": 10595 + }, + { + "epoch": 19.776119402985074, + "grad_norm": 0.3196175694465637, + "learning_rate": 1.940214413167496e-08, + "loss": 0.4083, + "num_input_tokens_seen": 3025928, + "step": 10600 + }, + { + "epoch": 19.78544776119403, + "grad_norm": 0.4234158992767334, + "learning_rate": 1.7831977670507148e-08, + "loss": 0.5945, + "num_input_tokens_seen": 3027240, + "step": 10605 + }, + { + "epoch": 19.794776119402986, + "grad_norm": 0.6898514032363892, + "learning_rate": 1.632803212681666e-08, + "loss": 0.6234, + "num_input_tokens_seen": 3028648, + "step": 10610 + }, + { + "epoch": 19.80410447761194, + "grad_norm": 0.518661618232727, + "learning_rate": 1.4890311487150698e-08, + "loss": 0.7224, + "num_input_tokens_seen": 3030056, + "step": 10615 + }, + { + "epoch": 19.813432835820894, + "grad_norm": 0.6441892385482788, + "learning_rate": 1.3518819562510776e-08, + "loss": 0.858, + "num_input_tokens_seen": 3031432, + "step": 10620 + }, + { + "epoch": 19.82276119402985, + "grad_norm": 0.5124002695083618, + "learning_rate": 1.221355998835272e-08, + "loss": 0.6413, + "num_input_tokens_seen": 3032936, + "step": 10625 + }, + { + "epoch": 19.832089552238806, + "grad_norm": 0.43050768971443176, + "learning_rate": 1.0974536224561682e-08, + "loss": 0.5857, + "num_input_tokens_seen": 3034408, + "step": 10630 + }, + { + "epoch": 19.84141791044776, + "grad_norm": 0.6039997339248657, + "learning_rate": 9.801751555452154e-09, + "loss": 0.8918, + "num_input_tokens_seen": 3035784, + "step": 10635 + }, + { + "epoch": 19.850746268656717, + "grad_norm": 0.45753714442253113, + "learning_rate": 8.695209089759626e-09, + "loss": 0.5075, + "num_input_tokens_seen": 3037320, + "step": 10640 + }, + { + "epoch": 19.86007462686567, + "grad_norm": 0.6077073812484741, + "learning_rate": 7.654911760621163e-09, + "loss": 0.6397, + "num_input_tokens_seen": 3038792, + "step": 10645 + }, + { + "epoch": 19.869402985074625, + "grad_norm": 0.4491482079029083, + "learning_rate": 6.680862325583736e-09, + "loss": 0.4452, + "num_input_tokens_seen": 3040424, + "step": 10650 + }, + { + "epoch": 19.878731343283583, + "grad_norm": 0.5239742994308472, + "learning_rate": 5.7730633665903365e-09, + "loss": 0.8456, + "num_input_tokens_seen": 3041736, + "step": 10655 + }, + { + "epoch": 19.888059701492537, + "grad_norm": 0.32511666417121887, + "learning_rate": 4.931517289963328e-09, + "loss": 0.8273, + "num_input_tokens_seen": 3042984, + "step": 10660 + }, + { + "epoch": 19.89738805970149, + "grad_norm": 0.48604151606559753, + "learning_rate": 4.156226326415547e-09, + "loss": 1.012, + "num_input_tokens_seen": 3044264, + "step": 10665 + }, + { + "epoch": 19.90671641791045, + "grad_norm": 0.5732764005661011, + "learning_rate": 3.4471925310280985e-09, + "loss": 0.6836, + "num_input_tokens_seen": 3045704, + "step": 10670 + }, + { + "epoch": 19.916044776119403, + "grad_norm": 0.5311990976333618, + "learning_rate": 2.804417783261459e-09, + "loss": 0.5702, + "num_input_tokens_seen": 3047144, + "step": 10675 + }, + { + "epoch": 19.925373134328357, + "grad_norm": 0.5661872029304504, + "learning_rate": 2.2279037869304964e-09, + "loss": 0.6302, + "num_input_tokens_seen": 3048648, + "step": 10680 + }, + { + "epoch": 19.934701492537314, + "grad_norm": 0.5163823366165161, + "learning_rate": 1.7176520702238964e-09, + "loss": 0.7604, + "num_input_tokens_seen": 3050184, + "step": 10685 + }, + { + "epoch": 19.94402985074627, + "grad_norm": 0.30865833163261414, + "learning_rate": 1.2736639856736344e-09, + "loss": 0.592, + "num_input_tokens_seen": 3051688, + "step": 10690 + }, + { + "epoch": 19.953358208955223, + "grad_norm": 0.40536224842071533, + "learning_rate": 8.959407101716277e-10, + "loss": 0.4236, + "num_input_tokens_seen": 3052968, + "step": 10695 + }, + { + "epoch": 19.96268656716418, + "grad_norm": 0.4638526141643524, + "learning_rate": 5.844832449641846e-10, + "loss": 0.6928, + "num_input_tokens_seen": 3054536, + "step": 10700 + }, + { + "epoch": 19.972014925373134, + "grad_norm": 0.411295622587204, + "learning_rate": 3.3929241563535053e-10, + "loss": 0.4705, + "num_input_tokens_seen": 3056072, + "step": 10705 + }, + { + "epoch": 19.98134328358209, + "grad_norm": 0.393028199672699, + "learning_rate": 1.6036887212078634e-10, + "loss": 0.8328, + "num_input_tokens_seen": 3057608, + "step": 10710 + }, + { + "epoch": 19.990671641791046, + "grad_norm": 0.4050822854042053, + "learning_rate": 4.771308869666591e-11, + "loss": 0.8278, + "num_input_tokens_seen": 3059016, + "step": 10715 + }, + { + "epoch": 20.0, + "grad_norm": 0.6550551652908325, + "learning_rate": 1.3253639852273126e-12, + "loss": 0.5962, + "num_input_tokens_seen": 3060208, + "step": 10720 + }, + { + "epoch": 20.0, + "eval_loss": 0.6943244934082031, + "eval_runtime": 2.467, + "eval_samples_per_second": 96.473, + "eval_steps_per_second": 24.321, + "num_input_tokens_seen": 3060208, + "step": 10720 + }, + { + "epoch": 20.0, + "num_input_tokens_seen": 3060208, + "step": 10720, + "total_flos": 1.3780339309215744e+17, + "train_loss": 0.8890379216911187, + "train_runtime": 1564.4317, + "train_samples_per_second": 27.371, + "train_steps_per_second": 6.852 + } + ], + "logging_steps": 5, + "max_steps": 10720, + "num_input_tokens_seen": 3060208, + "num_train_epochs": 20, + "save_steps": 536, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": true + }, + "attributes": {} + } + }, + "total_flos": 1.3780339309215744e+17, + "train_batch_size": 4, + "trial_name": null, + "trial_params": null +}