| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.0005091196049231866, | |
| "eval_steps": 3, | |
| "global_step": 10, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 5.0911960492318656e-05, | |
| "grad_norm": 1.652549386024475, | |
| "learning_rate": 2e-05, | |
| "loss": 2.4788, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 5.0911960492318656e-05, | |
| "eval_loss": 2.7557685375213623, | |
| "eval_runtime": 192.2208, | |
| "eval_samples_per_second": 43.029, | |
| "eval_steps_per_second": 21.517, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.00010182392098463731, | |
| "grad_norm": 2.056565761566162, | |
| "learning_rate": 4e-05, | |
| "loss": 2.7191, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.00015273588147695597, | |
| "grad_norm": 2.0462677478790283, | |
| "learning_rate": 6e-05, | |
| "loss": 2.4515, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.00015273588147695597, | |
| "eval_loss": 2.7481844425201416, | |
| "eval_runtime": 196.5675, | |
| "eval_samples_per_second": 42.077, | |
| "eval_steps_per_second": 21.041, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.00020364784196927463, | |
| "grad_norm": 1.7687551975250244, | |
| "learning_rate": 8e-05, | |
| "loss": 2.2275, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.0002545598024615933, | |
| "grad_norm": 1.8058793544769287, | |
| "learning_rate": 0.0001, | |
| "loss": 2.5333, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.00030547176295391194, | |
| "grad_norm": 1.953826665878296, | |
| "learning_rate": 0.00012, | |
| "loss": 3.2319, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.00030547176295391194, | |
| "eval_loss": 2.6476478576660156, | |
| "eval_runtime": 196.7209, | |
| "eval_samples_per_second": 42.044, | |
| "eval_steps_per_second": 21.025, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.0003563837234462306, | |
| "grad_norm": 1.8778396844863892, | |
| "learning_rate": 0.00014, | |
| "loss": 2.8905, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 0.00040729568393854925, | |
| "grad_norm": 2.3539109230041504, | |
| "learning_rate": 0.00016, | |
| "loss": 3.0018, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.0004582076444308679, | |
| "grad_norm": 2.2227814197540283, | |
| "learning_rate": 0.00018, | |
| "loss": 2.3831, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.0004582076444308679, | |
| "eval_loss": 2.393559694290161, | |
| "eval_runtime": 195.3959, | |
| "eval_samples_per_second": 42.329, | |
| "eval_steps_per_second": 21.167, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.0005091196049231866, | |
| "grad_norm": 1.7320082187652588, | |
| "learning_rate": 0.0002, | |
| "loss": 2.0209, | |
| "step": 10 | |
| } | |
| ], | |
| "logging_steps": 1, | |
| "max_steps": 10, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 3, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 240545515438080.0, | |
| "train_batch_size": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |