Training in progress, step 1500, checkpoint
Browse files- checkpoint-1500/config.json +40 -0
- checkpoint-1500/generation_config.json +12 -0
- checkpoint-1500/model.safetensors +3 -0
- checkpoint-1500/optimizer.pt +3 -0
- checkpoint-1500/rng_state.pth +3 -0
- checkpoint-1500/scheduler.pt +3 -0
- checkpoint-1500/trainer_state.json +1083 -0
- checkpoint-1500/training_args.bin +3 -0
checkpoint-1500/config.json
ADDED
|
@@ -0,0 +1,40 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "huihui-ai/Llama-3.2-1B-Instruct-abliterated",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"LlamaForCausalLM"
|
| 5 |
+
],
|
| 6 |
+
"attention_bias": false,
|
| 7 |
+
"attention_dropout": 0.0,
|
| 8 |
+
"bos_token_id": 128000,
|
| 9 |
+
"eos_token_id": [
|
| 10 |
+
128001,
|
| 11 |
+
128008,
|
| 12 |
+
128009
|
| 13 |
+
],
|
| 14 |
+
"head_dim": 64,
|
| 15 |
+
"hidden_act": "silu",
|
| 16 |
+
"hidden_size": 2048,
|
| 17 |
+
"initializer_range": 0.02,
|
| 18 |
+
"intermediate_size": 8192,
|
| 19 |
+
"max_position_embeddings": 131072,
|
| 20 |
+
"mlp_bias": false,
|
| 21 |
+
"model_type": "llama",
|
| 22 |
+
"num_attention_heads": 32,
|
| 23 |
+
"num_hidden_layers": 16,
|
| 24 |
+
"num_key_value_heads": 8,
|
| 25 |
+
"pretraining_tp": 1,
|
| 26 |
+
"rms_norm_eps": 1e-05,
|
| 27 |
+
"rope_scaling": {
|
| 28 |
+
"factor": 32.0,
|
| 29 |
+
"high_freq_factor": 4.0,
|
| 30 |
+
"low_freq_factor": 1.0,
|
| 31 |
+
"original_max_position_embeddings": 8192,
|
| 32 |
+
"rope_type": "llama3"
|
| 33 |
+
},
|
| 34 |
+
"rope_theta": 500000.0,
|
| 35 |
+
"tie_word_embeddings": true,
|
| 36 |
+
"torch_dtype": "bfloat16",
|
| 37 |
+
"transformers_version": "4.46.3",
|
| 38 |
+
"use_cache": true,
|
| 39 |
+
"vocab_size": 128256
|
| 40 |
+
}
|
checkpoint-1500/generation_config.json
ADDED
|
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token_id": 128000,
|
| 3 |
+
"do_sample": true,
|
| 4 |
+
"eos_token_id": [
|
| 5 |
+
128001,
|
| 6 |
+
128008,
|
| 7 |
+
128009
|
| 8 |
+
],
|
| 9 |
+
"temperature": 0.6,
|
| 10 |
+
"top_p": 0.9,
|
| 11 |
+
"transformers_version": "4.46.3"
|
| 12 |
+
}
|
checkpoint-1500/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:74d9cb303e8c1a3e762295ebc427957f59cdd4075973a94f621b72d268de494f
|
| 3 |
+
size 2471645608
|
checkpoint-1500/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d42e625a1598dc9d36cd04c6c90f613155c81378110154e422ca010a556083c4
|
| 3 |
+
size 3724602
|
checkpoint-1500/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:beb08987971240a9cec116c03cb626dee2fa851cac068f6bd3afdd37847bf7c2
|
| 3 |
+
size 14244
|
checkpoint-1500/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:50e0fa899b9afe4b9f410a947b2e82b267a415e5e533180013057e8a556bdbd5
|
| 3 |
+
size 1064
|
checkpoint-1500/trainer_state.json
ADDED
|
@@ -0,0 +1,1083 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.49437646768013843,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 1500,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.0032958431178675895,
|
| 13 |
+
"grad_norm": 74.5,
|
| 14 |
+
"learning_rate": 8.000000000000001e-07,
|
| 15 |
+
"loss": 0.7989,
|
| 16 |
+
"step": 10
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.006591686235735179,
|
| 20 |
+
"grad_norm": 212.0,
|
| 21 |
+
"learning_rate": 1.6000000000000001e-06,
|
| 22 |
+
"loss": 0.7703,
|
| 23 |
+
"step": 20
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.009887529353602768,
|
| 27 |
+
"grad_norm": 482.0,
|
| 28 |
+
"learning_rate": 2.4000000000000003e-06,
|
| 29 |
+
"loss": 0.821,
|
| 30 |
+
"step": 30
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.013183372471470358,
|
| 34 |
+
"grad_norm": 396.0,
|
| 35 |
+
"learning_rate": 3.2000000000000003e-06,
|
| 36 |
+
"loss": 0.8979,
|
| 37 |
+
"step": 40
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.01647921558933795,
|
| 41 |
+
"grad_norm": 488.0,
|
| 42 |
+
"learning_rate": 4.000000000000001e-06,
|
| 43 |
+
"loss": 1.17,
|
| 44 |
+
"step": 50
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.019775058707205535,
|
| 48 |
+
"grad_norm": 608.0,
|
| 49 |
+
"learning_rate": 4.800000000000001e-06,
|
| 50 |
+
"loss": 1.4005,
|
| 51 |
+
"step": 60
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.023070901825073126,
|
| 55 |
+
"grad_norm": 452.0,
|
| 56 |
+
"learning_rate": 5.600000000000001e-06,
|
| 57 |
+
"loss": 1.7127,
|
| 58 |
+
"step": 70
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.026366744942940716,
|
| 62 |
+
"grad_norm": 442.0,
|
| 63 |
+
"learning_rate": 6.4000000000000006e-06,
|
| 64 |
+
"loss": 1.4108,
|
| 65 |
+
"step": 80
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.029662588060808306,
|
| 69 |
+
"grad_norm": 111.0,
|
| 70 |
+
"learning_rate": 7.2000000000000005e-06,
|
| 71 |
+
"loss": 2.3574,
|
| 72 |
+
"step": 90
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.0329584311786759,
|
| 76 |
+
"grad_norm": 109.5,
|
| 77 |
+
"learning_rate": 8.000000000000001e-06,
|
| 78 |
+
"loss": 1.8115,
|
| 79 |
+
"step": 100
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.03625427429654349,
|
| 83 |
+
"grad_norm": 41.25,
|
| 84 |
+
"learning_rate": 8.8e-06,
|
| 85 |
+
"loss": 1.5844,
|
| 86 |
+
"step": 110
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.03955011741441107,
|
| 90 |
+
"grad_norm": 50.0,
|
| 91 |
+
"learning_rate": 9.600000000000001e-06,
|
| 92 |
+
"loss": 1.5766,
|
| 93 |
+
"step": 120
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.04284596053227866,
|
| 97 |
+
"grad_norm": 155.0,
|
| 98 |
+
"learning_rate": 1.04e-05,
|
| 99 |
+
"loss": 5.5685,
|
| 100 |
+
"step": 130
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.04614180365014625,
|
| 104 |
+
"grad_norm": 102.0,
|
| 105 |
+
"learning_rate": 1.1200000000000001e-05,
|
| 106 |
+
"loss": 1.4693,
|
| 107 |
+
"step": 140
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.04943764676801384,
|
| 111 |
+
"grad_norm": 256.0,
|
| 112 |
+
"learning_rate": 1.2e-05,
|
| 113 |
+
"loss": 1.4063,
|
| 114 |
+
"step": 150
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.05273348988588143,
|
| 118 |
+
"grad_norm": 81.0,
|
| 119 |
+
"learning_rate": 1.2800000000000001e-05,
|
| 120 |
+
"loss": 1.2139,
|
| 121 |
+
"step": 160
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.05602933300374902,
|
| 125 |
+
"grad_norm": 61.75,
|
| 126 |
+
"learning_rate": 1.3600000000000002e-05,
|
| 127 |
+
"loss": 1.2504,
|
| 128 |
+
"step": 170
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.05932517612161661,
|
| 132 |
+
"grad_norm": 12.0625,
|
| 133 |
+
"learning_rate": 1.4400000000000001e-05,
|
| 134 |
+
"loss": 1.338,
|
| 135 |
+
"step": 180
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.0626210192394842,
|
| 139 |
+
"grad_norm": 36.5,
|
| 140 |
+
"learning_rate": 1.5200000000000002e-05,
|
| 141 |
+
"loss": 1.2007,
|
| 142 |
+
"step": 190
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.0659168623573518,
|
| 146 |
+
"grad_norm": 35.0,
|
| 147 |
+
"learning_rate": 1.6000000000000003e-05,
|
| 148 |
+
"loss": 1.2368,
|
| 149 |
+
"step": 200
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.06921270547521938,
|
| 153 |
+
"grad_norm": 17.125,
|
| 154 |
+
"learning_rate": 1.6800000000000002e-05,
|
| 155 |
+
"loss": 2.1113,
|
| 156 |
+
"step": 210
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.07250854859308697,
|
| 160 |
+
"grad_norm": 29.125,
|
| 161 |
+
"learning_rate": 1.76e-05,
|
| 162 |
+
"loss": 1.2671,
|
| 163 |
+
"step": 220
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.07580439171095456,
|
| 167 |
+
"grad_norm": 43.0,
|
| 168 |
+
"learning_rate": 1.8400000000000003e-05,
|
| 169 |
+
"loss": 1.2318,
|
| 170 |
+
"step": 230
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.07910023482882214,
|
| 174 |
+
"grad_norm": 26.625,
|
| 175 |
+
"learning_rate": 1.9200000000000003e-05,
|
| 176 |
+
"loss": 1.0837,
|
| 177 |
+
"step": 240
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.08239607794668974,
|
| 181 |
+
"grad_norm": 26.5,
|
| 182 |
+
"learning_rate": 2e-05,
|
| 183 |
+
"loss": 1.2709,
|
| 184 |
+
"step": 250
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.08569192106455732,
|
| 188 |
+
"grad_norm": 23.875,
|
| 189 |
+
"learning_rate": 1.9999025240093045e-05,
|
| 190 |
+
"loss": 1.0127,
|
| 191 |
+
"step": 260
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.08898776418242492,
|
| 195 |
+
"grad_norm": 22.875,
|
| 196 |
+
"learning_rate": 1.9996101150403543e-05,
|
| 197 |
+
"loss": 1.0939,
|
| 198 |
+
"step": 270
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.0922836073002925,
|
| 202 |
+
"grad_norm": 15.1875,
|
| 203 |
+
"learning_rate": 1.9991228300988586e-05,
|
| 204 |
+
"loss": 1.1039,
|
| 205 |
+
"step": 280
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.0955794504181601,
|
| 209 |
+
"grad_norm": 44.25,
|
| 210 |
+
"learning_rate": 1.9984407641819812e-05,
|
| 211 |
+
"loss": 1.1438,
|
| 212 |
+
"step": 290
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.09887529353602768,
|
| 216 |
+
"grad_norm": 27.25,
|
| 217 |
+
"learning_rate": 1.9975640502598243e-05,
|
| 218 |
+
"loss": 1.1164,
|
| 219 |
+
"step": 300
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.10217113665389528,
|
| 223 |
+
"grad_norm": 31.25,
|
| 224 |
+
"learning_rate": 1.9964928592495046e-05,
|
| 225 |
+
"loss": 1.051,
|
| 226 |
+
"step": 310
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.10546697977176286,
|
| 230 |
+
"grad_norm": 15.625,
|
| 231 |
+
"learning_rate": 1.9952273999818312e-05,
|
| 232 |
+
"loss": 0.9002,
|
| 233 |
+
"step": 320
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.10876282288963045,
|
| 237 |
+
"grad_norm": 14.75,
|
| 238 |
+
"learning_rate": 1.9937679191605964e-05,
|
| 239 |
+
"loss": 0.9201,
|
| 240 |
+
"step": 330
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.11205866600749805,
|
| 244 |
+
"grad_norm": 41.0,
|
| 245 |
+
"learning_rate": 1.9921147013144782e-05,
|
| 246 |
+
"loss": 1.0844,
|
| 247 |
+
"step": 340
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.11535450912536563,
|
| 251 |
+
"grad_norm": 17.875,
|
| 252 |
+
"learning_rate": 1.9902680687415704e-05,
|
| 253 |
+
"loss": 0.946,
|
| 254 |
+
"step": 350
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.11865035224323323,
|
| 258 |
+
"grad_norm": 13.1875,
|
| 259 |
+
"learning_rate": 1.988228381446553e-05,
|
| 260 |
+
"loss": 0.9666,
|
| 261 |
+
"step": 360
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.12194619536110081,
|
| 265 |
+
"grad_norm": 15.4375,
|
| 266 |
+
"learning_rate": 1.985996037070505e-05,
|
| 267 |
+
"loss": 0.9573,
|
| 268 |
+
"step": 370
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.1252420384789684,
|
| 272 |
+
"grad_norm": 19.125,
|
| 273 |
+
"learning_rate": 1.983571470813386e-05,
|
| 274 |
+
"loss": 1.0149,
|
| 275 |
+
"step": 380
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.128537881596836,
|
| 279 |
+
"grad_norm": 17.125,
|
| 280 |
+
"learning_rate": 1.9809551553491918e-05,
|
| 281 |
+
"loss": 1.0354,
|
| 282 |
+
"step": 390
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.1318337247147036,
|
| 286 |
+
"grad_norm": 21.75,
|
| 287 |
+
"learning_rate": 1.9781476007338058e-05,
|
| 288 |
+
"loss": 1.0042,
|
| 289 |
+
"step": 400
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.13512956783257116,
|
| 293 |
+
"grad_norm": 12.4375,
|
| 294 |
+
"learning_rate": 1.9751493543055634e-05,
|
| 295 |
+
"loss": 0.8969,
|
| 296 |
+
"step": 410
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.13842541095043875,
|
| 300 |
+
"grad_norm": 14.75,
|
| 301 |
+
"learning_rate": 1.9719610005785466e-05,
|
| 302 |
+
"loss": 0.9187,
|
| 303 |
+
"step": 420
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.14172125406830635,
|
| 307 |
+
"grad_norm": 16.875,
|
| 308 |
+
"learning_rate": 1.9685831611286312e-05,
|
| 309 |
+
"loss": 0.918,
|
| 310 |
+
"step": 430
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.14501709718617395,
|
| 314 |
+
"grad_norm": 28.0,
|
| 315 |
+
"learning_rate": 1.9650164944723116e-05,
|
| 316 |
+
"loss": 0.8747,
|
| 317 |
+
"step": 440
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.14831294030404152,
|
| 321 |
+
"grad_norm": 15.375,
|
| 322 |
+
"learning_rate": 1.961261695938319e-05,
|
| 323 |
+
"loss": 0.9015,
|
| 324 |
+
"step": 450
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.15160878342190912,
|
| 328 |
+
"grad_norm": 26.75,
|
| 329 |
+
"learning_rate": 1.9573194975320672e-05,
|
| 330 |
+
"loss": 0.9056,
|
| 331 |
+
"step": 460
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.1549046265397767,
|
| 335 |
+
"grad_norm": 28.75,
|
| 336 |
+
"learning_rate": 1.9531906677929472e-05,
|
| 337 |
+
"loss": 0.859,
|
| 338 |
+
"step": 470
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.15820046965764428,
|
| 342 |
+
"grad_norm": 12.5,
|
| 343 |
+
"learning_rate": 1.9488760116444966e-05,
|
| 344 |
+
"loss": 0.8454,
|
| 345 |
+
"step": 480
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.16149631277551188,
|
| 349 |
+
"grad_norm": 19.75,
|
| 350 |
+
"learning_rate": 1.944376370237481e-05,
|
| 351 |
+
"loss": 0.9058,
|
| 352 |
+
"step": 490
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.16479215589337948,
|
| 356 |
+
"grad_norm": 18.0,
|
| 357 |
+
"learning_rate": 1.9396926207859085e-05,
|
| 358 |
+
"loss": 0.8204,
|
| 359 |
+
"step": 500
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.16808799901124707,
|
| 363 |
+
"grad_norm": 8.4375,
|
| 364 |
+
"learning_rate": 1.9348256763960146e-05,
|
| 365 |
+
"loss": 0.8926,
|
| 366 |
+
"step": 510
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.17138384212911464,
|
| 370 |
+
"grad_norm": 12.0625,
|
| 371 |
+
"learning_rate": 1.9297764858882516e-05,
|
| 372 |
+
"loss": 0.8243,
|
| 373 |
+
"step": 520
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.17467968524698224,
|
| 377 |
+
"grad_norm": 16.125,
|
| 378 |
+
"learning_rate": 1.9245460336123136e-05,
|
| 379 |
+
"loss": 0.8746,
|
| 380 |
+
"step": 530
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.17797552836484984,
|
| 384 |
+
"grad_norm": 20.25,
|
| 385 |
+
"learning_rate": 1.9191353392552346e-05,
|
| 386 |
+
"loss": 0.878,
|
| 387 |
+
"step": 540
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.18127137148271744,
|
| 391 |
+
"grad_norm": 17.75,
|
| 392 |
+
"learning_rate": 1.913545457642601e-05,
|
| 393 |
+
"loss": 0.7852,
|
| 394 |
+
"step": 550
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.184567214600585,
|
| 398 |
+
"grad_norm": 7.625,
|
| 399 |
+
"learning_rate": 1.907777478532909e-05,
|
| 400 |
+
"loss": 0.7749,
|
| 401 |
+
"step": 560
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.1878630577184526,
|
| 405 |
+
"grad_norm": 18.5,
|
| 406 |
+
"learning_rate": 1.901832526405114e-05,
|
| 407 |
+
"loss": 0.9067,
|
| 408 |
+
"step": 570
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.1911589008363202,
|
| 412 |
+
"grad_norm": 15.5625,
|
| 413 |
+
"learning_rate": 1.895711760239413e-05,
|
| 414 |
+
"loss": 0.8603,
|
| 415 |
+
"step": 580
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.19445474395418777,
|
| 419 |
+
"grad_norm": 16.375,
|
| 420 |
+
"learning_rate": 1.889416373291298e-05,
|
| 421 |
+
"loss": 0.8449,
|
| 422 |
+
"step": 590
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.19775058707205537,
|
| 426 |
+
"grad_norm": 12.4375,
|
| 427 |
+
"learning_rate": 1.8829475928589272e-05,
|
| 428 |
+
"loss": 0.8277,
|
| 429 |
+
"step": 600
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.20104643018992296,
|
| 433 |
+
"grad_norm": 16.0,
|
| 434 |
+
"learning_rate": 1.8763066800438638e-05,
|
| 435 |
+
"loss": 0.8477,
|
| 436 |
+
"step": 610
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.20434227330779056,
|
| 440 |
+
"grad_norm": 13.6875,
|
| 441 |
+
"learning_rate": 1.869494929505219e-05,
|
| 442 |
+
"loss": 0.7843,
|
| 443 |
+
"step": 620
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.20763811642565813,
|
| 447 |
+
"grad_norm": 8.0,
|
| 448 |
+
"learning_rate": 1.8625136692072577e-05,
|
| 449 |
+
"loss": 0.7957,
|
| 450 |
+
"step": 630
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.21093395954352573,
|
| 454 |
+
"grad_norm": 11.8125,
|
| 455 |
+
"learning_rate": 1.855364260160507e-05,
|
| 456 |
+
"loss": 0.8254,
|
| 457 |
+
"step": 640
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.21422980266139333,
|
| 461 |
+
"grad_norm": 8.6875,
|
| 462 |
+
"learning_rate": 1.848048096156426e-05,
|
| 463 |
+
"loss": 0.8332,
|
| 464 |
+
"step": 650
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.2175256457792609,
|
| 468 |
+
"grad_norm": 9.1875,
|
| 469 |
+
"learning_rate": 1.8405666034956842e-05,
|
| 470 |
+
"loss": 0.8352,
|
| 471 |
+
"step": 660
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.2208214888971285,
|
| 475 |
+
"grad_norm": 20.125,
|
| 476 |
+
"learning_rate": 1.8329212407100996e-05,
|
| 477 |
+
"loss": 0.8683,
|
| 478 |
+
"step": 670
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.2241173320149961,
|
| 482 |
+
"grad_norm": 14.25,
|
| 483 |
+
"learning_rate": 1.8251134982782952e-05,
|
| 484 |
+
"loss": 0.7979,
|
| 485 |
+
"step": 680
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.2274131751328637,
|
| 489 |
+
"grad_norm": 17.625,
|
| 490 |
+
"learning_rate": 1.8171448983351284e-05,
|
| 491 |
+
"loss": 0.8485,
|
| 492 |
+
"step": 690
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.23070901825073126,
|
| 496 |
+
"grad_norm": 15.0625,
|
| 497 |
+
"learning_rate": 1.8090169943749477e-05,
|
| 498 |
+
"loss": 0.7674,
|
| 499 |
+
"step": 700
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.23400486136859885,
|
| 503 |
+
"grad_norm": 4.96875,
|
| 504 |
+
"learning_rate": 1.8007313709487334e-05,
|
| 505 |
+
"loss": 0.7827,
|
| 506 |
+
"step": 710
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.23730070448646645,
|
| 510 |
+
"grad_norm": 13.8125,
|
| 511 |
+
"learning_rate": 1.792289643355191e-05,
|
| 512 |
+
"loss": 0.7801,
|
| 513 |
+
"step": 720
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.24059654760433402,
|
| 517 |
+
"grad_norm": 8.6875,
|
| 518 |
+
"learning_rate": 1.78369345732584e-05,
|
| 519 |
+
"loss": 0.8158,
|
| 520 |
+
"step": 730
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.24389239072220162,
|
| 524 |
+
"grad_norm": 20.875,
|
| 525 |
+
"learning_rate": 1.7749444887041797e-05,
|
| 526 |
+
"loss": 0.777,
|
| 527 |
+
"step": 740
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.24718823384006922,
|
| 531 |
+
"grad_norm": 11.0,
|
| 532 |
+
"learning_rate": 1.766044443118978e-05,
|
| 533 |
+
"loss": 0.7895,
|
| 534 |
+
"step": 750
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.2504840769579368,
|
| 538 |
+
"grad_norm": 12.5,
|
| 539 |
+
"learning_rate": 1.7569950556517566e-05,
|
| 540 |
+
"loss": 0.7931,
|
| 541 |
+
"step": 760
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.2537799200758044,
|
| 545 |
+
"grad_norm": 13.6875,
|
| 546 |
+
"learning_rate": 1.747798090498532e-05,
|
| 547 |
+
"loss": 0.8357,
|
| 548 |
+
"step": 770
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.257075763193672,
|
| 552 |
+
"grad_norm": 12.0625,
|
| 553 |
+
"learning_rate": 1.7384553406258842e-05,
|
| 554 |
+
"loss": 0.7632,
|
| 555 |
+
"step": 780
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.26037160631153955,
|
| 559 |
+
"grad_norm": 8.625,
|
| 560 |
+
"learning_rate": 1.7289686274214116e-05,
|
| 561 |
+
"loss": 0.7747,
|
| 562 |
+
"step": 790
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.2636674494294072,
|
| 566 |
+
"grad_norm": 18.75,
|
| 567 |
+
"learning_rate": 1.7193398003386514e-05,
|
| 568 |
+
"loss": 0.7421,
|
| 569 |
+
"step": 800
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.26696329254727474,
|
| 573 |
+
"grad_norm": 8.625,
|
| 574 |
+
"learning_rate": 1.709570736536521e-05,
|
| 575 |
+
"loss": 0.7616,
|
| 576 |
+
"step": 810
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 0.2702591356651423,
|
| 580 |
+
"grad_norm": 14.0,
|
| 581 |
+
"learning_rate": 1.6996633405133656e-05,
|
| 582 |
+
"loss": 0.752,
|
| 583 |
+
"step": 820
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.27355497878300994,
|
| 587 |
+
"grad_norm": 14.5625,
|
| 588 |
+
"learning_rate": 1.68961954373567e-05,
|
| 589 |
+
"loss": 0.7501,
|
| 590 |
+
"step": 830
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.2768508219008775,
|
| 594 |
+
"grad_norm": 19.125,
|
| 595 |
+
"learning_rate": 1.6794413042615168e-05,
|
| 596 |
+
"loss": 0.7874,
|
| 597 |
+
"step": 840
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.28014666501874513,
|
| 601 |
+
"grad_norm": 13.6875,
|
| 602 |
+
"learning_rate": 1.6691306063588583e-05,
|
| 603 |
+
"loss": 0.7667,
|
| 604 |
+
"step": 850
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 0.2834425081366127,
|
| 608 |
+
"grad_norm": 11.9375,
|
| 609 |
+
"learning_rate": 1.6586894601186804e-05,
|
| 610 |
+
"loss": 0.7087,
|
| 611 |
+
"step": 860
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.2867383512544803,
|
| 615 |
+
"grad_norm": 8.5,
|
| 616 |
+
"learning_rate": 1.6481199010631312e-05,
|
| 617 |
+
"loss": 0.7752,
|
| 618 |
+
"step": 870
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.2900341943723479,
|
| 622 |
+
"grad_norm": 7.59375,
|
| 623 |
+
"learning_rate": 1.63742398974869e-05,
|
| 624 |
+
"loss": 0.7764,
|
| 625 |
+
"step": 880
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.29333003749021547,
|
| 629 |
+
"grad_norm": 8.875,
|
| 630 |
+
"learning_rate": 1.6266038113644605e-05,
|
| 631 |
+
"loss": 0.7502,
|
| 632 |
+
"step": 890
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.29662588060808304,
|
| 636 |
+
"grad_norm": 9.1875,
|
| 637 |
+
"learning_rate": 1.6156614753256583e-05,
|
| 638 |
+
"loss": 0.7126,
|
| 639 |
+
"step": 900
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.29992172372595066,
|
| 643 |
+
"grad_norm": 14.8125,
|
| 644 |
+
"learning_rate": 1.6045991148623752e-05,
|
| 645 |
+
"loss": 0.778,
|
| 646 |
+
"step": 910
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.30321756684381823,
|
| 650 |
+
"grad_norm": 10.375,
|
| 651 |
+
"learning_rate": 1.5934188866037017e-05,
|
| 652 |
+
"loss": 0.7234,
|
| 653 |
+
"step": 920
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.3065134099616858,
|
| 657 |
+
"grad_norm": 11.1875,
|
| 658 |
+
"learning_rate": 1.5821229701572897e-05,
|
| 659 |
+
"loss": 0.8043,
|
| 660 |
+
"step": 930
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 0.3098092530795534,
|
| 664 |
+
"grad_norm": 12.9375,
|
| 665 |
+
"learning_rate": 1.570713567684432e-05,
|
| 666 |
+
"loss": 0.7271,
|
| 667 |
+
"step": 940
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.313105096197421,
|
| 671 |
+
"grad_norm": 4.21875,
|
| 672 |
+
"learning_rate": 1.5591929034707468e-05,
|
| 673 |
+
"loss": 0.6921,
|
| 674 |
+
"step": 950
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.31640093931528857,
|
| 678 |
+
"grad_norm": 8.625,
|
| 679 |
+
"learning_rate": 1.5475632234925505e-05,
|
| 680 |
+
"loss": 0.78,
|
| 681 |
+
"step": 960
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.3196967824331562,
|
| 685 |
+
"grad_norm": 7.34375,
|
| 686 |
+
"learning_rate": 1.5358267949789968e-05,
|
| 687 |
+
"loss": 0.7616,
|
| 688 |
+
"step": 970
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.32299262555102376,
|
| 692 |
+
"grad_norm": 7.8125,
|
| 693 |
+
"learning_rate": 1.5239859059700794e-05,
|
| 694 |
+
"loss": 0.7442,
|
| 695 |
+
"step": 980
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.3262884686688914,
|
| 699 |
+
"grad_norm": 12.5625,
|
| 700 |
+
"learning_rate": 1.5120428648705716e-05,
|
| 701 |
+
"loss": 0.7209,
|
| 702 |
+
"step": 990
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 0.32958431178675895,
|
| 706 |
+
"grad_norm": 11.3125,
|
| 707 |
+
"learning_rate": 1.5000000000000002e-05,
|
| 708 |
+
"loss": 0.7091,
|
| 709 |
+
"step": 1000
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.3328801549046265,
|
| 713 |
+
"grad_norm": 5.03125,
|
| 714 |
+
"learning_rate": 1.4878596591387329e-05,
|
| 715 |
+
"loss": 0.7082,
|
| 716 |
+
"step": 1010
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.33617599802249415,
|
| 720 |
+
"grad_norm": 4.03125,
|
| 721 |
+
"learning_rate": 1.4756242090702756e-05,
|
| 722 |
+
"loss": 0.6973,
|
| 723 |
+
"step": 1020
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.3394718411403617,
|
| 727 |
+
"grad_norm": 13.625,
|
| 728 |
+
"learning_rate": 1.463296035119862e-05,
|
| 729 |
+
"loss": 0.7182,
|
| 730 |
+
"step": 1030
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.3427676842582293,
|
| 734 |
+
"grad_norm": 15.6875,
|
| 735 |
+
"learning_rate": 1.4508775406894308e-05,
|
| 736 |
+
"loss": 0.6958,
|
| 737 |
+
"step": 1040
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.3460635273760969,
|
| 741 |
+
"grad_norm": 16.75,
|
| 742 |
+
"learning_rate": 1.4383711467890776e-05,
|
| 743 |
+
"loss": 0.6861,
|
| 744 |
+
"step": 1050
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 0.3493593704939645,
|
| 748 |
+
"grad_norm": 5.0,
|
| 749 |
+
"learning_rate": 1.4257792915650728e-05,
|
| 750 |
+
"loss": 0.675,
|
| 751 |
+
"step": 1060
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.35265521361183205,
|
| 755 |
+
"grad_norm": 10.125,
|
| 756 |
+
"learning_rate": 1.413104429824542e-05,
|
| 757 |
+
"loss": 0.7937,
|
| 758 |
+
"step": 1070
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.3559510567296997,
|
| 762 |
+
"grad_norm": 4.28125,
|
| 763 |
+
"learning_rate": 1.4003490325568953e-05,
|
| 764 |
+
"loss": 0.7161,
|
| 765 |
+
"step": 1080
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.35924689984756725,
|
| 769 |
+
"grad_norm": 7.03125,
|
| 770 |
+
"learning_rate": 1.3875155864521031e-05,
|
| 771 |
+
"loss": 0.713,
|
| 772 |
+
"step": 1090
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 0.36254274296543487,
|
| 776 |
+
"grad_norm": 19.0,
|
| 777 |
+
"learning_rate": 1.3746065934159123e-05,
|
| 778 |
+
"loss": 0.6866,
|
| 779 |
+
"step": 1100
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.36583858608330244,
|
| 783 |
+
"grad_norm": 11.5625,
|
| 784 |
+
"learning_rate": 1.3616245700820922e-05,
|
| 785 |
+
"loss": 0.7129,
|
| 786 |
+
"step": 1110
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 0.36913442920117,
|
| 790 |
+
"grad_norm": 4.03125,
|
| 791 |
+
"learning_rate": 1.3485720473218153e-05,
|
| 792 |
+
"loss": 0.7217,
|
| 793 |
+
"step": 1120
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.37243027231903764,
|
| 797 |
+
"grad_norm": 7.78125,
|
| 798 |
+
"learning_rate": 1.3354515697502552e-05,
|
| 799 |
+
"loss": 0.7008,
|
| 800 |
+
"step": 1130
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.3757261154369052,
|
| 804 |
+
"grad_norm": 9.25,
|
| 805 |
+
"learning_rate": 1.3222656952305113e-05,
|
| 806 |
+
"loss": 0.6741,
|
| 807 |
+
"step": 1140
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.3790219585547728,
|
| 811 |
+
"grad_norm": 7.3125,
|
| 812 |
+
"learning_rate": 1.3090169943749475e-05,
|
| 813 |
+
"loss": 0.7249,
|
| 814 |
+
"step": 1150
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 0.3823178016726404,
|
| 818 |
+
"grad_norm": 4.53125,
|
| 819 |
+
"learning_rate": 1.2957080500440469e-05,
|
| 820 |
+
"loss": 0.6616,
|
| 821 |
+
"step": 1160
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.38561364479050797,
|
| 825 |
+
"grad_norm": 3.5625,
|
| 826 |
+
"learning_rate": 1.2823414568428767e-05,
|
| 827 |
+
"loss": 0.6848,
|
| 828 |
+
"step": 1170
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 0.38890948790837554,
|
| 832 |
+
"grad_norm": 9.875,
|
| 833 |
+
"learning_rate": 1.2689198206152657e-05,
|
| 834 |
+
"loss": 0.6417,
|
| 835 |
+
"step": 1180
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.39220533102624316,
|
| 839 |
+
"grad_norm": 8.75,
|
| 840 |
+
"learning_rate": 1.2554457579357906e-05,
|
| 841 |
+
"loss": 0.6812,
|
| 842 |
+
"step": 1190
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.39550117414411073,
|
| 846 |
+
"grad_norm": 6.3125,
|
| 847 |
+
"learning_rate": 1.2419218955996677e-05,
|
| 848 |
+
"loss": 0.6304,
|
| 849 |
+
"step": 1200
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.3987970172619783,
|
| 853 |
+
"grad_norm": 5.28125,
|
| 854 |
+
"learning_rate": 1.2283508701106559e-05,
|
| 855 |
+
"loss": 0.6765,
|
| 856 |
+
"step": 1210
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 0.40209286037984593,
|
| 860 |
+
"grad_norm": 6.4375,
|
| 861 |
+
"learning_rate": 1.2147353271670634e-05,
|
| 862 |
+
"loss": 0.6552,
|
| 863 |
+
"step": 1220
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.4053887034977135,
|
| 867 |
+
"grad_norm": 9.375,
|
| 868 |
+
"learning_rate": 1.2010779211459649e-05,
|
| 869 |
+
"loss": 0.6989,
|
| 870 |
+
"step": 1230
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 0.4086845466155811,
|
| 874 |
+
"grad_norm": 4.875,
|
| 875 |
+
"learning_rate": 1.187381314585725e-05,
|
| 876 |
+
"loss": 0.6349,
|
| 877 |
+
"step": 1240
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.4119803897334487,
|
| 881 |
+
"grad_norm": 13.625,
|
| 882 |
+
"learning_rate": 1.1736481776669307e-05,
|
| 883 |
+
"loss": 0.6537,
|
| 884 |
+
"step": 1250
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.41527623285131626,
|
| 888 |
+
"grad_norm": 3.609375,
|
| 889 |
+
"learning_rate": 1.159881187691835e-05,
|
| 890 |
+
"loss": 0.6216,
|
| 891 |
+
"step": 1260
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.4185720759691839,
|
| 895 |
+
"grad_norm": 9.9375,
|
| 896 |
+
"learning_rate": 1.1460830285624119e-05,
|
| 897 |
+
"loss": 0.6572,
|
| 898 |
+
"step": 1270
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 0.42186791908705146,
|
| 902 |
+
"grad_norm": 6.84375,
|
| 903 |
+
"learning_rate": 1.1322563902571227e-05,
|
| 904 |
+
"loss": 0.6511,
|
| 905 |
+
"step": 1280
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 0.425163762204919,
|
| 909 |
+
"grad_norm": 7.09375,
|
| 910 |
+
"learning_rate": 1.1184039683065014e-05,
|
| 911 |
+
"loss": 0.6671,
|
| 912 |
+
"step": 1290
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"epoch": 0.42845960532278665,
|
| 916 |
+
"grad_norm": 4.875,
|
| 917 |
+
"learning_rate": 1.1045284632676535e-05,
|
| 918 |
+
"loss": 0.6279,
|
| 919 |
+
"step": 1300
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.4317554484406542,
|
| 923 |
+
"grad_norm": 8.6875,
|
| 924 |
+
"learning_rate": 1.0906325801977804e-05,
|
| 925 |
+
"loss": 0.6442,
|
| 926 |
+
"step": 1310
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 0.4350512915585218,
|
| 930 |
+
"grad_norm": 22.75,
|
| 931 |
+
"learning_rate": 1.0767190281268187e-05,
|
| 932 |
+
"loss": 0.6786,
|
| 933 |
+
"step": 1320
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.4383471346763894,
|
| 937 |
+
"grad_norm": 5.625,
|
| 938 |
+
"learning_rate": 1.0627905195293135e-05,
|
| 939 |
+
"loss": 0.6482,
|
| 940 |
+
"step": 1330
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 0.441642977794257,
|
| 944 |
+
"grad_norm": 10.5,
|
| 945 |
+
"learning_rate": 1.0488497697956134e-05,
|
| 946 |
+
"loss": 0.6611,
|
| 947 |
+
"step": 1340
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.44493882091212456,
|
| 951 |
+
"grad_norm": 3.109375,
|
| 952 |
+
"learning_rate": 1.0348994967025012e-05,
|
| 953 |
+
"loss": 0.6093,
|
| 954 |
+
"step": 1350
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 0.4482346640299922,
|
| 958 |
+
"grad_norm": 4.21875,
|
| 959 |
+
"learning_rate": 1.0209424198833571e-05,
|
| 960 |
+
"loss": 0.7027,
|
| 961 |
+
"step": 1360
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.45153050714785975,
|
| 965 |
+
"grad_norm": 3.546875,
|
| 966 |
+
"learning_rate": 1.0069812602979617e-05,
|
| 967 |
+
"loss": 0.665,
|
| 968 |
+
"step": 1370
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.4548263502657274,
|
| 972 |
+
"grad_norm": 8.0,
|
| 973 |
+
"learning_rate": 9.930187397020385e-06,
|
| 974 |
+
"loss": 0.6277,
|
| 975 |
+
"step": 1380
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.45812219338359494,
|
| 979 |
+
"grad_norm": 12.9375,
|
| 980 |
+
"learning_rate": 9.790575801166432e-06,
|
| 981 |
+
"loss": 0.6452,
|
| 982 |
+
"step": 1390
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 0.4614180365014625,
|
| 986 |
+
"grad_norm": 8.75,
|
| 987 |
+
"learning_rate": 9.651005032974994e-06,
|
| 988 |
+
"loss": 0.6493,
|
| 989 |
+
"step": 1400
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 0.46471387961933014,
|
| 993 |
+
"grad_norm": 5.53125,
|
| 994 |
+
"learning_rate": 9.511502302043867e-06,
|
| 995 |
+
"loss": 0.6633,
|
| 996 |
+
"step": 1410
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 0.4680097227371977,
|
| 1000 |
+
"grad_norm": 5.4375,
|
| 1001 |
+
"learning_rate": 9.372094804706867e-06,
|
| 1002 |
+
"loss": 0.6568,
|
| 1003 |
+
"step": 1420
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 0.4713055658550653,
|
| 1007 |
+
"grad_norm": 8.1875,
|
| 1008 |
+
"learning_rate": 9.232809718731815e-06,
|
| 1009 |
+
"loss": 0.6294,
|
| 1010 |
+
"step": 1430
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.4746014089729329,
|
| 1014 |
+
"grad_norm": 6.3125,
|
| 1015 |
+
"learning_rate": 9.093674198022201e-06,
|
| 1016 |
+
"loss": 0.6133,
|
| 1017 |
+
"step": 1440
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.4778972520908005,
|
| 1021 |
+
"grad_norm": 4.8125,
|
| 1022 |
+
"learning_rate": 8.954715367323468e-06,
|
| 1023 |
+
"loss": 0.6392,
|
| 1024 |
+
"step": 1450
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 0.48119309520866804,
|
| 1028 |
+
"grad_norm": 3.34375,
|
| 1029 |
+
"learning_rate": 8.815960316934991e-06,
|
| 1030 |
+
"loss": 0.6103,
|
| 1031 |
+
"step": 1460
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 0.48448893832653567,
|
| 1035 |
+
"grad_norm": 3.375,
|
| 1036 |
+
"learning_rate": 8.677436097428775e-06,
|
| 1037 |
+
"loss": 0.6069,
|
| 1038 |
+
"step": 1470
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 0.48778478144440324,
|
| 1042 |
+
"grad_norm": 11.4375,
|
| 1043 |
+
"learning_rate": 8.539169714375885e-06,
|
| 1044 |
+
"loss": 0.7006,
|
| 1045 |
+
"step": 1480
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 0.49108062456227086,
|
| 1049 |
+
"grad_norm": 6.34375,
|
| 1050 |
+
"learning_rate": 8.401188123081653e-06,
|
| 1051 |
+
"loss": 0.6163,
|
| 1052 |
+
"step": 1490
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 0.49437646768013843,
|
| 1056 |
+
"grad_norm": 7.78125,
|
| 1057 |
+
"learning_rate": 8.263518223330698e-06,
|
| 1058 |
+
"loss": 0.6063,
|
| 1059 |
+
"step": 1500
|
| 1060 |
+
}
|
| 1061 |
+
],
|
| 1062 |
+
"logging_steps": 10,
|
| 1063 |
+
"max_steps": 2500,
|
| 1064 |
+
"num_input_tokens_seen": 0,
|
| 1065 |
+
"num_train_epochs": 1,
|
| 1066 |
+
"save_steps": 500,
|
| 1067 |
+
"stateful_callbacks": {
|
| 1068 |
+
"TrainerControl": {
|
| 1069 |
+
"args": {
|
| 1070 |
+
"should_epoch_stop": false,
|
| 1071 |
+
"should_evaluate": false,
|
| 1072 |
+
"should_log": false,
|
| 1073 |
+
"should_save": true,
|
| 1074 |
+
"should_training_stop": false
|
| 1075 |
+
},
|
| 1076 |
+
"attributes": {}
|
| 1077 |
+
}
|
| 1078 |
+
},
|
| 1079 |
+
"total_flos": 5.73984932364288e+17,
|
| 1080 |
+
"train_batch_size": 4,
|
| 1081 |
+
"trial_name": null,
|
| 1082 |
+
"trial_params": null
|
| 1083 |
+
}
|
checkpoint-1500/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7067b18e12e53f801fe59dd35886f8daf20b398c5f95d75002b19445049d5042
|
| 3 |
+
size 5368
|