Upload 10 files
Browse files100 epoch training model on CDC Covid19 data (final qa pairs - details at https://github.com/jasondeden/capstone)
- .gitattributes +1 -0
- config.json +32 -0
- final_qa_pairs_dev.json +0 -0
- final_qa_pairs_train.json +0 -0
- pytorch_model.bin +3 -0
- special_tokens_map.json +6 -0
- tokenizer.json +3 -0
- tokenizer_config.json +11 -0
- train_results.json +8 -0
- trainer_state.json +1531 -0
- training_args.bin +3 -0
.gitattributes
CHANGED
|
@@ -32,3 +32,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
|
| 32 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 33 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 34 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
|
|
| 32 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 33 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 34 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
| 35 |
+
tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
config.json
ADDED
|
@@ -0,0 +1,32 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "bigscience/bloom-560m",
|
| 3 |
+
"apply_residual_connection_post_layernorm": false,
|
| 4 |
+
"architectures": [
|
| 5 |
+
"BloomForQuestionAnswering"
|
| 6 |
+
],
|
| 7 |
+
"attention_dropout": 0.0,
|
| 8 |
+
"attention_softmax_in_fp32": true,
|
| 9 |
+
"bias_dropout_fusion": true,
|
| 10 |
+
"bos_token_id": 1,
|
| 11 |
+
"eos_token_id": 2,
|
| 12 |
+
"hidden_dropout": 0.0,
|
| 13 |
+
"hidden_size": 1024,
|
| 14 |
+
"initializer_range": 0.02,
|
| 15 |
+
"layer_norm_epsilon": 1e-05,
|
| 16 |
+
"masked_softmax_fusion": true,
|
| 17 |
+
"model_type": "bloom",
|
| 18 |
+
"n_head": 16,
|
| 19 |
+
"n_inner": null,
|
| 20 |
+
"n_layer": 24,
|
| 21 |
+
"offset_alibi": 100,
|
| 22 |
+
"pad_token_id": 3,
|
| 23 |
+
"pretraining_tp": 1,
|
| 24 |
+
"skip_bias_add": true,
|
| 25 |
+
"skip_bias_add_qkv": false,
|
| 26 |
+
"slow_but_exact": false,
|
| 27 |
+
"torch_dtype": "float32",
|
| 28 |
+
"transformers_version": "4.24.0.dev0",
|
| 29 |
+
"unk_token_id": 0,
|
| 30 |
+
"use_cache": true,
|
| 31 |
+
"vocab_size": 250880
|
| 32 |
+
}
|
final_qa_pairs_dev.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
final_qa_pairs_train.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:81bdd8f215a3551748871694928e4c6bdc7b05091e54de4a99c09dec7f0db40b
|
| 3 |
+
size 2236963891
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"eos_token": "</s>",
|
| 4 |
+
"pad_token": "<pad>",
|
| 5 |
+
"unk_token": "<unk>"
|
| 6 |
+
}
|
tokenizer.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8f6efc66e73f1fd69da4f436e48befb519fdff3fe18910850c1d41bd862293a5
|
| 3 |
+
size 14500443
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,11 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"bos_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"name_or_path": "bigscience/bloom-560m",
|
| 6 |
+
"pad_token": "<pad>",
|
| 7 |
+
"padding_side": "left",
|
| 8 |
+
"special_tokens_map_file": null,
|
| 9 |
+
"tokenizer_class": "BloomTokenizer",
|
| 10 |
+
"unk_token": "<unk>"
|
| 11 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 100.0,
|
| 3 |
+
"train_loss": 1.9559951608499904,
|
| 4 |
+
"train_runtime": 50817.2182,
|
| 5 |
+
"train_samples": 7536,
|
| 6 |
+
"train_samples_per_second": 14.83,
|
| 7 |
+
"train_steps_per_second": 2.472
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1531 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 100.0,
|
| 5 |
+
"global_step": 125600,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.4,
|
| 12 |
+
"learning_rate": 2.988296178343949e-05,
|
| 13 |
+
"loss": 7.6975,
|
| 14 |
+
"step": 500
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.8,
|
| 18 |
+
"learning_rate": 2.9763535031847134e-05,
|
| 19 |
+
"loss": 4.413,
|
| 20 |
+
"step": 1000
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 1.19,
|
| 24 |
+
"learning_rate": 2.964410828025478e-05,
|
| 25 |
+
"loss": 3.5707,
|
| 26 |
+
"step": 1500
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 1.59,
|
| 30 |
+
"learning_rate": 2.952468152866242e-05,
|
| 31 |
+
"loss": 3.4966,
|
| 32 |
+
"step": 2000
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 1.99,
|
| 36 |
+
"learning_rate": 2.9405254777070064e-05,
|
| 37 |
+
"loss": 3.7477,
|
| 38 |
+
"step": 2500
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 2.39,
|
| 42 |
+
"learning_rate": 2.9285828025477707e-05,
|
| 43 |
+
"loss": 3.4618,
|
| 44 |
+
"step": 3000
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 2.79,
|
| 48 |
+
"learning_rate": 2.916640127388535e-05,
|
| 49 |
+
"loss": 3.4774,
|
| 50 |
+
"step": 3500
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 3.18,
|
| 54 |
+
"learning_rate": 2.9046974522292994e-05,
|
| 55 |
+
"loss": 3.4214,
|
| 56 |
+
"step": 4000
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 3.58,
|
| 60 |
+
"learning_rate": 2.8927786624203823e-05,
|
| 61 |
+
"loss": 2.5773,
|
| 62 |
+
"step": 4500
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 3.98,
|
| 66 |
+
"learning_rate": 2.8808359872611467e-05,
|
| 67 |
+
"loss": 3.2549,
|
| 68 |
+
"step": 5000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 4.38,
|
| 72 |
+
"learning_rate": 2.868893312101911e-05,
|
| 73 |
+
"loss": 2.4922,
|
| 74 |
+
"step": 5500
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 4.78,
|
| 78 |
+
"learning_rate": 2.856950636942675e-05,
|
| 79 |
+
"loss": 2.6262,
|
| 80 |
+
"step": 6000
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 5.18,
|
| 84 |
+
"learning_rate": 2.845031847133758e-05,
|
| 85 |
+
"loss": 2.5596,
|
| 86 |
+
"step": 6500
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 5.57,
|
| 90 |
+
"learning_rate": 2.8331130573248408e-05,
|
| 91 |
+
"loss": 2.7913,
|
| 92 |
+
"step": 7000
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 5.97,
|
| 96 |
+
"learning_rate": 2.821170382165605e-05,
|
| 97 |
+
"loss": 3.0112,
|
| 98 |
+
"step": 7500
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 6.37,
|
| 102 |
+
"learning_rate": 2.8092277070063698e-05,
|
| 103 |
+
"loss": 2.6159,
|
| 104 |
+
"step": 8000
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 6.77,
|
| 108 |
+
"learning_rate": 2.7972850318471338e-05,
|
| 109 |
+
"loss": 3.9984,
|
| 110 |
+
"step": 8500
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 7.17,
|
| 114 |
+
"learning_rate": 2.785342356687898e-05,
|
| 115 |
+
"loss": 2.6099,
|
| 116 |
+
"step": 9000
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 7.56,
|
| 120 |
+
"learning_rate": 2.7733996815286625e-05,
|
| 121 |
+
"loss": 2.5667,
|
| 122 |
+
"step": 9500
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 7.96,
|
| 126 |
+
"learning_rate": 2.7614570063694268e-05,
|
| 127 |
+
"loss": 2.5934,
|
| 128 |
+
"step": 10000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 8.36,
|
| 132 |
+
"learning_rate": 2.7495382165605094e-05,
|
| 133 |
+
"loss": 2.7481,
|
| 134 |
+
"step": 10500
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 8.76,
|
| 138 |
+
"learning_rate": 2.737595541401274e-05,
|
| 139 |
+
"loss": 3.0303,
|
| 140 |
+
"step": 11000
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 9.16,
|
| 144 |
+
"learning_rate": 2.7256528662420384e-05,
|
| 145 |
+
"loss": 2.8187,
|
| 146 |
+
"step": 11500
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 9.55,
|
| 150 |
+
"learning_rate": 2.7137101910828027e-05,
|
| 151 |
+
"loss": 2.8765,
|
| 152 |
+
"step": 12000
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 9.95,
|
| 156 |
+
"learning_rate": 2.7017914012738853e-05,
|
| 157 |
+
"loss": 2.8891,
|
| 158 |
+
"step": 12500
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 10.35,
|
| 162 |
+
"learning_rate": 2.6898487261146496e-05,
|
| 163 |
+
"loss": 2.6341,
|
| 164 |
+
"step": 13000
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 10.75,
|
| 168 |
+
"learning_rate": 2.677906050955414e-05,
|
| 169 |
+
"loss": 2.9864,
|
| 170 |
+
"step": 13500
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 11.15,
|
| 174 |
+
"learning_rate": 2.6659633757961786e-05,
|
| 175 |
+
"loss": 2.3945,
|
| 176 |
+
"step": 14000
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 11.54,
|
| 180 |
+
"learning_rate": 2.6540207006369426e-05,
|
| 181 |
+
"loss": 2.4969,
|
| 182 |
+
"step": 14500
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 11.94,
|
| 186 |
+
"learning_rate": 2.6421019108280255e-05,
|
| 187 |
+
"loss": 3.4386,
|
| 188 |
+
"step": 15000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 12.34,
|
| 192 |
+
"learning_rate": 2.63015923566879e-05,
|
| 193 |
+
"loss": 2.1529,
|
| 194 |
+
"step": 15500
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 12.74,
|
| 198 |
+
"learning_rate": 2.6182165605095542e-05,
|
| 199 |
+
"loss": 2.4111,
|
| 200 |
+
"step": 16000
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 13.14,
|
| 204 |
+
"learning_rate": 2.6062738853503186e-05,
|
| 205 |
+
"loss": 2.1625,
|
| 206 |
+
"step": 16500
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 13.54,
|
| 210 |
+
"learning_rate": 2.594355095541401e-05,
|
| 211 |
+
"loss": 2.221,
|
| 212 |
+
"step": 17000
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 13.93,
|
| 216 |
+
"learning_rate": 2.5824124203821655e-05,
|
| 217 |
+
"loss": 2.3075,
|
| 218 |
+
"step": 17500
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 14.33,
|
| 222 |
+
"learning_rate": 2.57046974522293e-05,
|
| 223 |
+
"loss": 2.0438,
|
| 224 |
+
"step": 18000
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 14.73,
|
| 228 |
+
"learning_rate": 2.5585270700636945e-05,
|
| 229 |
+
"loss": 2.2311,
|
| 230 |
+
"step": 18500
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 15.13,
|
| 234 |
+
"learning_rate": 2.5465843949044585e-05,
|
| 235 |
+
"loss": 2.3121,
|
| 236 |
+
"step": 19000
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 15.53,
|
| 240 |
+
"learning_rate": 2.5346417197452228e-05,
|
| 241 |
+
"loss": 1.8979,
|
| 242 |
+
"step": 19500
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 15.92,
|
| 246 |
+
"learning_rate": 2.5226990445859875e-05,
|
| 247 |
+
"loss": 2.0793,
|
| 248 |
+
"step": 20000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 16.32,
|
| 252 |
+
"learning_rate": 2.5107563694267518e-05,
|
| 253 |
+
"loss": 2.4946,
|
| 254 |
+
"step": 20500
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 16.72,
|
| 258 |
+
"learning_rate": 2.4988136942675158e-05,
|
| 259 |
+
"loss": 2.6039,
|
| 260 |
+
"step": 21000
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 17.12,
|
| 264 |
+
"learning_rate": 2.4868949044585987e-05,
|
| 265 |
+
"loss": 2.3325,
|
| 266 |
+
"step": 21500
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 17.52,
|
| 270 |
+
"learning_rate": 2.4749761146496816e-05,
|
| 271 |
+
"loss": 2.5405,
|
| 272 |
+
"step": 22000
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 17.91,
|
| 276 |
+
"learning_rate": 2.463033439490446e-05,
|
| 277 |
+
"loss": 2.2666,
|
| 278 |
+
"step": 22500
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 18.31,
|
| 282 |
+
"learning_rate": 2.4510907643312103e-05,
|
| 283 |
+
"loss": 1.8855,
|
| 284 |
+
"step": 23000
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 18.71,
|
| 288 |
+
"learning_rate": 2.439171974522293e-05,
|
| 289 |
+
"loss": 2.5188,
|
| 290 |
+
"step": 23500
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 19.11,
|
| 294 |
+
"learning_rate": 2.4272292993630572e-05,
|
| 295 |
+
"loss": 2.0856,
|
| 296 |
+
"step": 24000
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 19.51,
|
| 300 |
+
"learning_rate": 2.415286624203822e-05,
|
| 301 |
+
"loss": 2.04,
|
| 302 |
+
"step": 24500
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 19.9,
|
| 306 |
+
"learning_rate": 2.4033439490445862e-05,
|
| 307 |
+
"loss": 2.0085,
|
| 308 |
+
"step": 25000
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 20.3,
|
| 312 |
+
"learning_rate": 2.3914012738853502e-05,
|
| 313 |
+
"loss": 2.2144,
|
| 314 |
+
"step": 25500
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 20.7,
|
| 318 |
+
"learning_rate": 2.3794585987261145e-05,
|
| 319 |
+
"loss": 2.1031,
|
| 320 |
+
"step": 26000
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 21.1,
|
| 324 |
+
"learning_rate": 2.3675159235668792e-05,
|
| 325 |
+
"loss": 2.1336,
|
| 326 |
+
"step": 26500
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 21.5,
|
| 330 |
+
"learning_rate": 2.3555732484076436e-05,
|
| 331 |
+
"loss": 2.3206,
|
| 332 |
+
"step": 27000
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 21.89,
|
| 336 |
+
"learning_rate": 2.3436783439490447e-05,
|
| 337 |
+
"loss": 2.0794,
|
| 338 |
+
"step": 27500
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 22.29,
|
| 342 |
+
"learning_rate": 2.3317356687898087e-05,
|
| 343 |
+
"loss": 2.0694,
|
| 344 |
+
"step": 28000
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 22.69,
|
| 348 |
+
"learning_rate": 2.3197929936305734e-05,
|
| 349 |
+
"loss": 2.0866,
|
| 350 |
+
"step": 28500
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 23.09,
|
| 354 |
+
"learning_rate": 2.3078503184713377e-05,
|
| 355 |
+
"loss": 2.2699,
|
| 356 |
+
"step": 29000
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 23.49,
|
| 360 |
+
"learning_rate": 2.295907643312102e-05,
|
| 361 |
+
"loss": 2.0621,
|
| 362 |
+
"step": 29500
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 23.89,
|
| 366 |
+
"learning_rate": 2.283964968152866e-05,
|
| 367 |
+
"loss": 2.1662,
|
| 368 |
+
"step": 30000
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 24.28,
|
| 372 |
+
"learning_rate": 2.2720222929936307e-05,
|
| 373 |
+
"loss": 1.9482,
|
| 374 |
+
"step": 30500
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 24.68,
|
| 378 |
+
"learning_rate": 2.260079617834395e-05,
|
| 379 |
+
"loss": 1.9606,
|
| 380 |
+
"step": 31000
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 25.08,
|
| 384 |
+
"learning_rate": 2.2481369426751594e-05,
|
| 385 |
+
"loss": 2.2075,
|
| 386 |
+
"step": 31500
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 25.48,
|
| 390 |
+
"learning_rate": 2.2361942675159234e-05,
|
| 391 |
+
"loss": 1.9507,
|
| 392 |
+
"step": 32000
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 25.88,
|
| 396 |
+
"learning_rate": 2.224251592356688e-05,
|
| 397 |
+
"loss": 1.99,
|
| 398 |
+
"step": 32500
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 26.27,
|
| 402 |
+
"learning_rate": 2.2123089171974524e-05,
|
| 403 |
+
"loss": 2.0022,
|
| 404 |
+
"step": 33000
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 26.67,
|
| 408 |
+
"learning_rate": 2.2003662420382167e-05,
|
| 409 |
+
"loss": 2.0491,
|
| 410 |
+
"step": 33500
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 27.07,
|
| 414 |
+
"learning_rate": 2.1884235668789807e-05,
|
| 415 |
+
"loss": 2.1034,
|
| 416 |
+
"step": 34000
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 27.47,
|
| 420 |
+
"learning_rate": 2.1765047770700636e-05,
|
| 421 |
+
"loss": 2.0613,
|
| 422 |
+
"step": 34500
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 27.87,
|
| 426 |
+
"learning_rate": 2.164562101910828e-05,
|
| 427 |
+
"loss": 2.0392,
|
| 428 |
+
"step": 35000
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 28.26,
|
| 432 |
+
"learning_rate": 2.1526194267515926e-05,
|
| 433 |
+
"loss": 1.9726,
|
| 434 |
+
"step": 35500
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 28.66,
|
| 438 |
+
"learning_rate": 2.1406767515923566e-05,
|
| 439 |
+
"loss": 1.6027,
|
| 440 |
+
"step": 36000
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 29.06,
|
| 444 |
+
"learning_rate": 2.128734076433121e-05,
|
| 445 |
+
"loss": 2.5043,
|
| 446 |
+
"step": 36500
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 29.46,
|
| 450 |
+
"learning_rate": 2.1167914012738853e-05,
|
| 451 |
+
"loss": 2.111,
|
| 452 |
+
"step": 37000
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 29.86,
|
| 456 |
+
"learning_rate": 2.10484872611465e-05,
|
| 457 |
+
"loss": 2.0965,
|
| 458 |
+
"step": 37500
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 30.25,
|
| 462 |
+
"learning_rate": 2.092906050955414e-05,
|
| 463 |
+
"loss": 1.9389,
|
| 464 |
+
"step": 38000
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 30.65,
|
| 468 |
+
"learning_rate": 2.0809633757961783e-05,
|
| 469 |
+
"loss": 2.1853,
|
| 470 |
+
"step": 38500
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 31.05,
|
| 474 |
+
"learning_rate": 2.0690207006369427e-05,
|
| 475 |
+
"loss": 1.946,
|
| 476 |
+
"step": 39000
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 31.45,
|
| 480 |
+
"learning_rate": 2.0570780254777073e-05,
|
| 481 |
+
"loss": 2.0071,
|
| 482 |
+
"step": 39500
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 31.85,
|
| 486 |
+
"learning_rate": 2.0451831210191085e-05,
|
| 487 |
+
"loss": 2.2209,
|
| 488 |
+
"step": 40000
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 32.25,
|
| 492 |
+
"learning_rate": 2.0332404458598725e-05,
|
| 493 |
+
"loss": 2.0641,
|
| 494 |
+
"step": 40500
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 32.64,
|
| 498 |
+
"learning_rate": 2.021297770700637e-05,
|
| 499 |
+
"loss": 1.9674,
|
| 500 |
+
"step": 41000
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 33.04,
|
| 504 |
+
"learning_rate": 2.0093550955414015e-05,
|
| 505 |
+
"loss": 2.0169,
|
| 506 |
+
"step": 41500
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 33.44,
|
| 510 |
+
"learning_rate": 1.9974124203821658e-05,
|
| 511 |
+
"loss": 2.0076,
|
| 512 |
+
"step": 42000
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 33.84,
|
| 516 |
+
"learning_rate": 1.9854936305732484e-05,
|
| 517 |
+
"loss": 2.2492,
|
| 518 |
+
"step": 42500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 34.24,
|
| 522 |
+
"learning_rate": 1.9735509554140127e-05,
|
| 523 |
+
"loss": 1.7799,
|
| 524 |
+
"step": 43000
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 34.63,
|
| 528 |
+
"learning_rate": 1.9616321656050956e-05,
|
| 529 |
+
"loss": 2.1029,
|
| 530 |
+
"step": 43500
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 35.03,
|
| 534 |
+
"learning_rate": 1.94968949044586e-05,
|
| 535 |
+
"loss": 1.8489,
|
| 536 |
+
"step": 44000
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 35.43,
|
| 540 |
+
"learning_rate": 1.937770700636943e-05,
|
| 541 |
+
"loss": 1.6914,
|
| 542 |
+
"step": 44500
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 35.83,
|
| 546 |
+
"learning_rate": 1.925828025477707e-05,
|
| 547 |
+
"loss": 2.4834,
|
| 548 |
+
"step": 45000
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 36.23,
|
| 552 |
+
"learning_rate": 1.9138853503184712e-05,
|
| 553 |
+
"loss": 1.7828,
|
| 554 |
+
"step": 45500
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 36.62,
|
| 558 |
+
"learning_rate": 1.901942675159236e-05,
|
| 559 |
+
"loss": 1.7454,
|
| 560 |
+
"step": 46000
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 37.02,
|
| 564 |
+
"learning_rate": 1.8900000000000002e-05,
|
| 565 |
+
"loss": 2.2061,
|
| 566 |
+
"step": 46500
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 37.42,
|
| 570 |
+
"learning_rate": 1.8780573248407642e-05,
|
| 571 |
+
"loss": 1.9779,
|
| 572 |
+
"step": 47000
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 37.82,
|
| 576 |
+
"learning_rate": 1.8661146496815285e-05,
|
| 577 |
+
"loss": 1.9194,
|
| 578 |
+
"step": 47500
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 38.22,
|
| 582 |
+
"learning_rate": 1.8541719745222932e-05,
|
| 583 |
+
"loss": 1.9433,
|
| 584 |
+
"step": 48000
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 38.61,
|
| 588 |
+
"learning_rate": 1.8422292993630575e-05,
|
| 589 |
+
"loss": 1.9242,
|
| 590 |
+
"step": 48500
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 39.01,
|
| 594 |
+
"learning_rate": 1.8302866242038215e-05,
|
| 595 |
+
"loss": 1.9324,
|
| 596 |
+
"step": 49000
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 39.41,
|
| 600 |
+
"learning_rate": 1.818343949044586e-05,
|
| 601 |
+
"loss": 1.9326,
|
| 602 |
+
"step": 49500
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 39.81,
|
| 606 |
+
"learning_rate": 1.8064012738853506e-05,
|
| 607 |
+
"loss": 1.8975,
|
| 608 |
+
"step": 50000
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 40.21,
|
| 612 |
+
"learning_rate": 1.794458598726115e-05,
|
| 613 |
+
"loss": 2.0671,
|
| 614 |
+
"step": 50500
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 40.61,
|
| 618 |
+
"learning_rate": 1.782515923566879e-05,
|
| 619 |
+
"loss": 2.209,
|
| 620 |
+
"step": 51000
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 41.0,
|
| 624 |
+
"learning_rate": 1.7705971337579618e-05,
|
| 625 |
+
"loss": 1.9638,
|
| 626 |
+
"step": 51500
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 41.4,
|
| 630 |
+
"learning_rate": 1.758654458598726e-05,
|
| 631 |
+
"loss": 1.6372,
|
| 632 |
+
"step": 52000
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 41.8,
|
| 636 |
+
"learning_rate": 1.7467117834394905e-05,
|
| 637 |
+
"loss": 2.1012,
|
| 638 |
+
"step": 52500
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 42.2,
|
| 642 |
+
"learning_rate": 1.7347691082802548e-05,
|
| 643 |
+
"loss": 2.1207,
|
| 644 |
+
"step": 53000
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 42.6,
|
| 648 |
+
"learning_rate": 1.722826433121019e-05,
|
| 649 |
+
"loss": 1.8135,
|
| 650 |
+
"step": 53500
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 42.99,
|
| 654 |
+
"learning_rate": 1.7108837579617835e-05,
|
| 655 |
+
"loss": 2.1076,
|
| 656 |
+
"step": 54000
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 43.39,
|
| 660 |
+
"learning_rate": 1.6989410828025478e-05,
|
| 661 |
+
"loss": 1.8442,
|
| 662 |
+
"step": 54500
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 43.79,
|
| 666 |
+
"learning_rate": 1.686998407643312e-05,
|
| 667 |
+
"loss": 1.8598,
|
| 668 |
+
"step": 55000
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 44.19,
|
| 672 |
+
"learning_rate": 1.6750557324840765e-05,
|
| 673 |
+
"loss": 1.8612,
|
| 674 |
+
"step": 55500
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 44.59,
|
| 678 |
+
"learning_rate": 1.6631369426751594e-05,
|
| 679 |
+
"loss": 1.7965,
|
| 680 |
+
"step": 56000
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 44.98,
|
| 684 |
+
"learning_rate": 1.6511942675159237e-05,
|
| 685 |
+
"loss": 1.8743,
|
| 686 |
+
"step": 56500
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 45.38,
|
| 690 |
+
"learning_rate": 1.6392515923566877e-05,
|
| 691 |
+
"loss": 1.8581,
|
| 692 |
+
"step": 57000
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 45.78,
|
| 696 |
+
"learning_rate": 1.627308917197452e-05,
|
| 697 |
+
"loss": 1.8212,
|
| 698 |
+
"step": 57500
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 46.18,
|
| 702 |
+
"learning_rate": 1.6153662420382167e-05,
|
| 703 |
+
"loss": 1.7475,
|
| 704 |
+
"step": 58000
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 46.58,
|
| 708 |
+
"learning_rate": 1.603423566878981e-05,
|
| 709 |
+
"loss": 1.5722,
|
| 710 |
+
"step": 58500
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 46.97,
|
| 714 |
+
"learning_rate": 1.591480891719745e-05,
|
| 715 |
+
"loss": 1.7929,
|
| 716 |
+
"step": 59000
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 47.37,
|
| 720 |
+
"learning_rate": 1.5795382165605094e-05,
|
| 721 |
+
"loss": 1.7854,
|
| 722 |
+
"step": 59500
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 47.77,
|
| 726 |
+
"learning_rate": 1.5676194267515923e-05,
|
| 727 |
+
"loss": 1.6686,
|
| 728 |
+
"step": 60000
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 48.17,
|
| 732 |
+
"learning_rate": 1.5556767515923566e-05,
|
| 733 |
+
"loss": 1.8969,
|
| 734 |
+
"step": 60500
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 48.57,
|
| 738 |
+
"learning_rate": 1.543734076433121e-05,
|
| 739 |
+
"loss": 1.6989,
|
| 740 |
+
"step": 61000
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 48.96,
|
| 744 |
+
"learning_rate": 1.5317914012738853e-05,
|
| 745 |
+
"loss": 1.9204,
|
| 746 |
+
"step": 61500
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 49.36,
|
| 750 |
+
"learning_rate": 1.5198726114649682e-05,
|
| 751 |
+
"loss": 1.7297,
|
| 752 |
+
"step": 62000
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 49.76,
|
| 756 |
+
"learning_rate": 1.5079299363057326e-05,
|
| 757 |
+
"loss": 1.6225,
|
| 758 |
+
"step": 62500
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 50.16,
|
| 762 |
+
"learning_rate": 1.4959872611464969e-05,
|
| 763 |
+
"loss": 1.8492,
|
| 764 |
+
"step": 63000
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 50.56,
|
| 768 |
+
"learning_rate": 1.4840445859872612e-05,
|
| 769 |
+
"loss": 1.948,
|
| 770 |
+
"step": 63500
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 50.96,
|
| 774 |
+
"learning_rate": 1.472125796178344e-05,
|
| 775 |
+
"loss": 1.4926,
|
| 776 |
+
"step": 64000
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 51.35,
|
| 780 |
+
"learning_rate": 1.4601831210191083e-05,
|
| 781 |
+
"loss": 1.8296,
|
| 782 |
+
"step": 64500
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 51.75,
|
| 786 |
+
"learning_rate": 1.4482404458598726e-05,
|
| 787 |
+
"loss": 1.672,
|
| 788 |
+
"step": 65000
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 52.15,
|
| 792 |
+
"learning_rate": 1.436297770700637e-05,
|
| 793 |
+
"loss": 1.7793,
|
| 794 |
+
"step": 65500
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 52.55,
|
| 798 |
+
"learning_rate": 1.4243789808917199e-05,
|
| 799 |
+
"loss": 1.6569,
|
| 800 |
+
"step": 66000
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 52.95,
|
| 804 |
+
"learning_rate": 1.412436305732484e-05,
|
| 805 |
+
"loss": 2.2555,
|
| 806 |
+
"step": 66500
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 53.34,
|
| 810 |
+
"learning_rate": 1.4004936305732486e-05,
|
| 811 |
+
"loss": 1.7448,
|
| 812 |
+
"step": 67000
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 53.74,
|
| 816 |
+
"learning_rate": 1.3885509554140127e-05,
|
| 817 |
+
"loss": 1.8515,
|
| 818 |
+
"step": 67500
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 54.14,
|
| 822 |
+
"learning_rate": 1.3766321656050956e-05,
|
| 823 |
+
"loss": 1.5126,
|
| 824 |
+
"step": 68000
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 54.54,
|
| 828 |
+
"learning_rate": 1.3646894904458598e-05,
|
| 829 |
+
"loss": 1.8567,
|
| 830 |
+
"step": 68500
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 54.94,
|
| 834 |
+
"learning_rate": 1.3527468152866243e-05,
|
| 835 |
+
"loss": 1.8848,
|
| 836 |
+
"step": 69000
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 55.33,
|
| 840 |
+
"learning_rate": 1.340828025477707e-05,
|
| 841 |
+
"loss": 1.6216,
|
| 842 |
+
"step": 69500
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 55.73,
|
| 846 |
+
"learning_rate": 1.3288853503184714e-05,
|
| 847 |
+
"loss": 2.0117,
|
| 848 |
+
"step": 70000
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 56.13,
|
| 852 |
+
"learning_rate": 1.3169426751592357e-05,
|
| 853 |
+
"loss": 1.8113,
|
| 854 |
+
"step": 70500
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 56.53,
|
| 858 |
+
"learning_rate": 1.305e-05,
|
| 859 |
+
"loss": 1.7053,
|
| 860 |
+
"step": 71000
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 56.93,
|
| 864 |
+
"learning_rate": 1.2930573248407644e-05,
|
| 865 |
+
"loss": 1.7271,
|
| 866 |
+
"step": 71500
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 57.32,
|
| 870 |
+
"learning_rate": 1.2811385350318471e-05,
|
| 871 |
+
"loss": 1.6382,
|
| 872 |
+
"step": 72000
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 57.72,
|
| 876 |
+
"learning_rate": 1.2691958598726116e-05,
|
| 877 |
+
"loss": 1.5688,
|
| 878 |
+
"step": 72500
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 58.12,
|
| 882 |
+
"learning_rate": 1.2572531847133758e-05,
|
| 883 |
+
"loss": 1.6947,
|
| 884 |
+
"step": 73000
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 58.52,
|
| 888 |
+
"learning_rate": 1.2453105095541403e-05,
|
| 889 |
+
"loss": 1.5709,
|
| 890 |
+
"step": 73500
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 58.92,
|
| 894 |
+
"learning_rate": 1.2333678343949045e-05,
|
| 895 |
+
"loss": 1.8282,
|
| 896 |
+
"step": 74000
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 59.32,
|
| 900 |
+
"learning_rate": 1.221425159235669e-05,
|
| 901 |
+
"loss": 1.6447,
|
| 902 |
+
"step": 74500
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 59.71,
|
| 906 |
+
"learning_rate": 1.2095063694267515e-05,
|
| 907 |
+
"loss": 2.1066,
|
| 908 |
+
"step": 75000
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 60.11,
|
| 912 |
+
"learning_rate": 1.197563694267516e-05,
|
| 913 |
+
"loss": 1.7348,
|
| 914 |
+
"step": 75500
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 60.51,
|
| 918 |
+
"learning_rate": 1.1856210191082802e-05,
|
| 919 |
+
"loss": 1.5,
|
| 920 |
+
"step": 76000
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 60.91,
|
| 924 |
+
"learning_rate": 1.1736783439490447e-05,
|
| 925 |
+
"loss": 1.8146,
|
| 926 |
+
"step": 76500
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 61.31,
|
| 930 |
+
"learning_rate": 1.1617356687898089e-05,
|
| 931 |
+
"loss": 1.7271,
|
| 932 |
+
"step": 77000
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 61.7,
|
| 936 |
+
"learning_rate": 1.1497929936305734e-05,
|
| 937 |
+
"loss": 1.4879,
|
| 938 |
+
"step": 77500
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 62.1,
|
| 942 |
+
"learning_rate": 1.1378503184713375e-05,
|
| 943 |
+
"loss": 1.7706,
|
| 944 |
+
"step": 78000
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 62.5,
|
| 948 |
+
"learning_rate": 1.125907643312102e-05,
|
| 949 |
+
"loss": 1.7463,
|
| 950 |
+
"step": 78500
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 62.9,
|
| 954 |
+
"learning_rate": 1.1139888535031846e-05,
|
| 955 |
+
"loss": 1.4587,
|
| 956 |
+
"step": 79000
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 63.3,
|
| 960 |
+
"learning_rate": 1.1020461783439491e-05,
|
| 961 |
+
"loss": 1.754,
|
| 962 |
+
"step": 79500
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 63.69,
|
| 966 |
+
"learning_rate": 1.0901035031847133e-05,
|
| 967 |
+
"loss": 1.6684,
|
| 968 |
+
"step": 80000
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 64.09,
|
| 972 |
+
"learning_rate": 1.0781847133757962e-05,
|
| 973 |
+
"loss": 1.4427,
|
| 974 |
+
"step": 80500
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 64.49,
|
| 978 |
+
"learning_rate": 1.0662420382165605e-05,
|
| 979 |
+
"loss": 1.5719,
|
| 980 |
+
"step": 81000
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 64.89,
|
| 984 |
+
"learning_rate": 1.0542993630573249e-05,
|
| 985 |
+
"loss": 1.7179,
|
| 986 |
+
"step": 81500
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 65.29,
|
| 990 |
+
"learning_rate": 1.0423566878980892e-05,
|
| 991 |
+
"loss": 1.6181,
|
| 992 |
+
"step": 82000
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 65.68,
|
| 996 |
+
"learning_rate": 1.0304140127388535e-05,
|
| 997 |
+
"loss": 1.4319,
|
| 998 |
+
"step": 82500
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 66.08,
|
| 1002 |
+
"learning_rate": 1.0185191082802548e-05,
|
| 1003 |
+
"loss": 1.78,
|
| 1004 |
+
"step": 83000
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 66.48,
|
| 1008 |
+
"learning_rate": 1.0065764331210192e-05,
|
| 1009 |
+
"loss": 1.5097,
|
| 1010 |
+
"step": 83500
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 66.88,
|
| 1014 |
+
"learning_rate": 9.946337579617835e-06,
|
| 1015 |
+
"loss": 1.6796,
|
| 1016 |
+
"step": 84000
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 67.28,
|
| 1020 |
+
"learning_rate": 9.826910828025479e-06,
|
| 1021 |
+
"loss": 1.7987,
|
| 1022 |
+
"step": 84500
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 67.68,
|
| 1026 |
+
"learning_rate": 9.707484076433122e-06,
|
| 1027 |
+
"loss": 1.5529,
|
| 1028 |
+
"step": 85000
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 68.07,
|
| 1032 |
+
"learning_rate": 9.588057324840764e-06,
|
| 1033 |
+
"loss": 1.7102,
|
| 1034 |
+
"step": 85500
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 68.47,
|
| 1038 |
+
"learning_rate": 9.468630573248409e-06,
|
| 1039 |
+
"loss": 1.7089,
|
| 1040 |
+
"step": 86000
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 68.87,
|
| 1044 |
+
"learning_rate": 9.34920382165605e-06,
|
| 1045 |
+
"loss": 1.5696,
|
| 1046 |
+
"step": 86500
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 69.27,
|
| 1050 |
+
"learning_rate": 9.229777070063695e-06,
|
| 1051 |
+
"loss": 1.4092,
|
| 1052 |
+
"step": 87000
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 69.67,
|
| 1056 |
+
"learning_rate": 9.110589171974523e-06,
|
| 1057 |
+
"loss": 1.7112,
|
| 1058 |
+
"step": 87500
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 70.06,
|
| 1062 |
+
"learning_rate": 8.991162420382166e-06,
|
| 1063 |
+
"loss": 1.6769,
|
| 1064 |
+
"step": 88000
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 70.46,
|
| 1068 |
+
"learning_rate": 8.87173566878981e-06,
|
| 1069 |
+
"loss": 1.7474,
|
| 1070 |
+
"step": 88500
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 70.86,
|
| 1074 |
+
"learning_rate": 8.752547770700637e-06,
|
| 1075 |
+
"loss": 1.3906,
|
| 1076 |
+
"step": 89000
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 71.26,
|
| 1080 |
+
"learning_rate": 8.63312101910828e-06,
|
| 1081 |
+
"loss": 1.6714,
|
| 1082 |
+
"step": 89500
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 71.66,
|
| 1086 |
+
"learning_rate": 8.513694267515923e-06,
|
| 1087 |
+
"loss": 1.6018,
|
| 1088 |
+
"step": 90000
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 72.05,
|
| 1092 |
+
"learning_rate": 8.394267515923567e-06,
|
| 1093 |
+
"loss": 1.4916,
|
| 1094 |
+
"step": 90500
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 72.45,
|
| 1098 |
+
"learning_rate": 8.27484076433121e-06,
|
| 1099 |
+
"loss": 1.8035,
|
| 1100 |
+
"step": 91000
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 72.85,
|
| 1104 |
+
"learning_rate": 8.155414012738854e-06,
|
| 1105 |
+
"loss": 1.3387,
|
| 1106 |
+
"step": 91500
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 73.25,
|
| 1110 |
+
"learning_rate": 8.035987261146497e-06,
|
| 1111 |
+
"loss": 1.5553,
|
| 1112 |
+
"step": 92000
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 73.65,
|
| 1116 |
+
"learning_rate": 7.91656050955414e-06,
|
| 1117 |
+
"loss": 1.5534,
|
| 1118 |
+
"step": 92500
|
| 1119 |
+
},
|
| 1120 |
+
{
|
| 1121 |
+
"epoch": 74.04,
|
| 1122 |
+
"learning_rate": 7.797133757961784e-06,
|
| 1123 |
+
"loss": 1.3926,
|
| 1124 |
+
"step": 93000
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 74.44,
|
| 1128 |
+
"learning_rate": 7.677707006369427e-06,
|
| 1129 |
+
"loss": 1.5332,
|
| 1130 |
+
"step": 93500
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 74.84,
|
| 1134 |
+
"learning_rate": 7.5582802547770704e-06,
|
| 1135 |
+
"loss": 1.5935,
|
| 1136 |
+
"step": 94000
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 75.24,
|
| 1140 |
+
"learning_rate": 7.438853503184713e-06,
|
| 1141 |
+
"loss": 1.6671,
|
| 1142 |
+
"step": 94500
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 75.64,
|
| 1146 |
+
"learning_rate": 7.319665605095542e-06,
|
| 1147 |
+
"loss": 1.662,
|
| 1148 |
+
"step": 95000
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 76.04,
|
| 1152 |
+
"learning_rate": 7.200238853503185e-06,
|
| 1153 |
+
"loss": 1.6435,
|
| 1154 |
+
"step": 95500
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 76.43,
|
| 1158 |
+
"learning_rate": 7.080812101910829e-06,
|
| 1159 |
+
"loss": 1.5282,
|
| 1160 |
+
"step": 96000
|
| 1161 |
+
},
|
| 1162 |
+
{
|
| 1163 |
+
"epoch": 76.83,
|
| 1164 |
+
"learning_rate": 6.961385350318472e-06,
|
| 1165 |
+
"loss": 1.6214,
|
| 1166 |
+
"step": 96500
|
| 1167 |
+
},
|
| 1168 |
+
{
|
| 1169 |
+
"epoch": 77.23,
|
| 1170 |
+
"learning_rate": 6.842197452229299e-06,
|
| 1171 |
+
"loss": 1.2999,
|
| 1172 |
+
"step": 97000
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 77.63,
|
| 1176 |
+
"learning_rate": 6.722770700636943e-06,
|
| 1177 |
+
"loss": 1.4861,
|
| 1178 |
+
"step": 97500
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 78.03,
|
| 1182 |
+
"learning_rate": 6.603343949044586e-06,
|
| 1183 |
+
"loss": 1.493,
|
| 1184 |
+
"step": 98000
|
| 1185 |
+
},
|
| 1186 |
+
{
|
| 1187 |
+
"epoch": 78.42,
|
| 1188 |
+
"learning_rate": 6.484156050955414e-06,
|
| 1189 |
+
"loss": 1.5782,
|
| 1190 |
+
"step": 98500
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 78.82,
|
| 1194 |
+
"learning_rate": 6.364729299363058e-06,
|
| 1195 |
+
"loss": 1.6043,
|
| 1196 |
+
"step": 99000
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 79.22,
|
| 1200 |
+
"learning_rate": 6.245302547770701e-06,
|
| 1201 |
+
"loss": 1.7383,
|
| 1202 |
+
"step": 99500
|
| 1203 |
+
},
|
| 1204 |
+
{
|
| 1205 |
+
"epoch": 79.62,
|
| 1206 |
+
"learning_rate": 6.1258757961783444e-06,
|
| 1207 |
+
"loss": 1.5453,
|
| 1208 |
+
"step": 100000
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 80.02,
|
| 1212 |
+
"learning_rate": 6.006449044585987e-06,
|
| 1213 |
+
"loss": 1.6063,
|
| 1214 |
+
"step": 100500
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 80.41,
|
| 1218 |
+
"learning_rate": 5.88702229299363e-06,
|
| 1219 |
+
"loss": 1.5471,
|
| 1220 |
+
"step": 101000
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 80.81,
|
| 1224 |
+
"learning_rate": 5.767595541401274e-06,
|
| 1225 |
+
"loss": 1.6526,
|
| 1226 |
+
"step": 101500
|
| 1227 |
+
},
|
| 1228 |
+
{
|
| 1229 |
+
"epoch": 81.21,
|
| 1230 |
+
"learning_rate": 5.648168789808917e-06,
|
| 1231 |
+
"loss": 1.4817,
|
| 1232 |
+
"step": 102000
|
| 1233 |
+
},
|
| 1234 |
+
{
|
| 1235 |
+
"epoch": 81.61,
|
| 1236 |
+
"learning_rate": 5.528980891719745e-06,
|
| 1237 |
+
"loss": 1.3771,
|
| 1238 |
+
"step": 102500
|
| 1239 |
+
},
|
| 1240 |
+
{
|
| 1241 |
+
"epoch": 82.01,
|
| 1242 |
+
"learning_rate": 5.409554140127389e-06,
|
| 1243 |
+
"loss": 1.5052,
|
| 1244 |
+
"step": 103000
|
| 1245 |
+
},
|
| 1246 |
+
{
|
| 1247 |
+
"epoch": 82.4,
|
| 1248 |
+
"learning_rate": 5.290127388535032e-06,
|
| 1249 |
+
"loss": 1.7024,
|
| 1250 |
+
"step": 103500
|
| 1251 |
+
},
|
| 1252 |
+
{
|
| 1253 |
+
"epoch": 82.8,
|
| 1254 |
+
"learning_rate": 5.170700636942675e-06,
|
| 1255 |
+
"loss": 1.6985,
|
| 1256 |
+
"step": 104000
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 83.2,
|
| 1260 |
+
"learning_rate": 5.051512738853503e-06,
|
| 1261 |
+
"loss": 1.5487,
|
| 1262 |
+
"step": 104500
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 83.6,
|
| 1266 |
+
"learning_rate": 4.932085987261146e-06,
|
| 1267 |
+
"loss": 1.3429,
|
| 1268 |
+
"step": 105000
|
| 1269 |
+
},
|
| 1270 |
+
{
|
| 1271 |
+
"epoch": 84.0,
|
| 1272 |
+
"learning_rate": 4.812659235668789e-06,
|
| 1273 |
+
"loss": 1.6937,
|
| 1274 |
+
"step": 105500
|
| 1275 |
+
},
|
| 1276 |
+
{
|
| 1277 |
+
"epoch": 84.39,
|
| 1278 |
+
"learning_rate": 4.6934713375796184e-06,
|
| 1279 |
+
"loss": 1.5639,
|
| 1280 |
+
"step": 106000
|
| 1281 |
+
},
|
| 1282 |
+
{
|
| 1283 |
+
"epoch": 84.79,
|
| 1284 |
+
"learning_rate": 4.574044585987262e-06,
|
| 1285 |
+
"loss": 1.5466,
|
| 1286 |
+
"step": 106500
|
| 1287 |
+
},
|
| 1288 |
+
{
|
| 1289 |
+
"epoch": 85.19,
|
| 1290 |
+
"learning_rate": 4.454617834394905e-06,
|
| 1291 |
+
"loss": 1.4726,
|
| 1292 |
+
"step": 107000
|
| 1293 |
+
},
|
| 1294 |
+
{
|
| 1295 |
+
"epoch": 85.59,
|
| 1296 |
+
"learning_rate": 4.3351910828025485e-06,
|
| 1297 |
+
"loss": 1.5633,
|
| 1298 |
+
"step": 107500
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 85.99,
|
| 1302 |
+
"learning_rate": 4.215764331210192e-06,
|
| 1303 |
+
"loss": 1.5687,
|
| 1304 |
+
"step": 108000
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 86.39,
|
| 1308 |
+
"learning_rate": 4.096337579617834e-06,
|
| 1309 |
+
"loss": 1.2633,
|
| 1310 |
+
"step": 108500
|
| 1311 |
+
},
|
| 1312 |
+
{
|
| 1313 |
+
"epoch": 86.78,
|
| 1314 |
+
"learning_rate": 3.976910828025478e-06,
|
| 1315 |
+
"loss": 1.5035,
|
| 1316 |
+
"step": 109000
|
| 1317 |
+
},
|
| 1318 |
+
{
|
| 1319 |
+
"epoch": 87.18,
|
| 1320 |
+
"learning_rate": 3.857484076433121e-06,
|
| 1321 |
+
"loss": 1.6182,
|
| 1322 |
+
"step": 109500
|
| 1323 |
+
},
|
| 1324 |
+
{
|
| 1325 |
+
"epoch": 87.58,
|
| 1326 |
+
"learning_rate": 3.7380573248407645e-06,
|
| 1327 |
+
"loss": 1.6295,
|
| 1328 |
+
"step": 110000
|
| 1329 |
+
},
|
| 1330 |
+
{
|
| 1331 |
+
"epoch": 87.98,
|
| 1332 |
+
"learning_rate": 3.618630573248408e-06,
|
| 1333 |
+
"loss": 1.5837,
|
| 1334 |
+
"step": 110500
|
| 1335 |
+
},
|
| 1336 |
+
{
|
| 1337 |
+
"epoch": 88.38,
|
| 1338 |
+
"learning_rate": 3.4992038216560512e-06,
|
| 1339 |
+
"loss": 1.3664,
|
| 1340 |
+
"step": 111000
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 88.77,
|
| 1344 |
+
"learning_rate": 3.3797770700636946e-06,
|
| 1345 |
+
"loss": 1.6751,
|
| 1346 |
+
"step": 111500
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 89.17,
|
| 1350 |
+
"learning_rate": 3.2605891719745224e-06,
|
| 1351 |
+
"loss": 1.5701,
|
| 1352 |
+
"step": 112000
|
| 1353 |
+
},
|
| 1354 |
+
{
|
| 1355 |
+
"epoch": 89.57,
|
| 1356 |
+
"learning_rate": 3.1411624203821653e-06,
|
| 1357 |
+
"loss": 1.4553,
|
| 1358 |
+
"step": 112500
|
| 1359 |
+
},
|
| 1360 |
+
{
|
| 1361 |
+
"epoch": 89.97,
|
| 1362 |
+
"learning_rate": 3.0217356687898087e-06,
|
| 1363 |
+
"loss": 1.4953,
|
| 1364 |
+
"step": 113000
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 90.37,
|
| 1368 |
+
"learning_rate": 2.902308917197452e-06,
|
| 1369 |
+
"loss": 1.6721,
|
| 1370 |
+
"step": 113500
|
| 1371 |
+
},
|
| 1372 |
+
{
|
| 1373 |
+
"epoch": 90.76,
|
| 1374 |
+
"learning_rate": 2.7831210191082802e-06,
|
| 1375 |
+
"loss": 1.4171,
|
| 1376 |
+
"step": 114000
|
| 1377 |
+
},
|
| 1378 |
+
{
|
| 1379 |
+
"epoch": 91.16,
|
| 1380 |
+
"learning_rate": 2.6636942675159236e-06,
|
| 1381 |
+
"loss": 1.8271,
|
| 1382 |
+
"step": 114500
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 91.56,
|
| 1386 |
+
"learning_rate": 2.544267515923567e-06,
|
| 1387 |
+
"loss": 1.6508,
|
| 1388 |
+
"step": 115000
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 91.96,
|
| 1392 |
+
"learning_rate": 2.4248407643312103e-06,
|
| 1393 |
+
"loss": 1.5092,
|
| 1394 |
+
"step": 115500
|
| 1395 |
+
},
|
| 1396 |
+
{
|
| 1397 |
+
"epoch": 92.36,
|
| 1398 |
+
"learning_rate": 2.305652866242038e-06,
|
| 1399 |
+
"loss": 1.5075,
|
| 1400 |
+
"step": 116000
|
| 1401 |
+
},
|
| 1402 |
+
{
|
| 1403 |
+
"epoch": 92.75,
|
| 1404 |
+
"learning_rate": 2.1864649681528663e-06,
|
| 1405 |
+
"loss": 1.5702,
|
| 1406 |
+
"step": 116500
|
| 1407 |
+
},
|
| 1408 |
+
{
|
| 1409 |
+
"epoch": 93.15,
|
| 1410 |
+
"learning_rate": 2.0670382165605097e-06,
|
| 1411 |
+
"loss": 1.5646,
|
| 1412 |
+
"step": 117000
|
| 1413 |
+
},
|
| 1414 |
+
{
|
| 1415 |
+
"epoch": 93.55,
|
| 1416 |
+
"learning_rate": 1.947611464968153e-06,
|
| 1417 |
+
"loss": 1.4956,
|
| 1418 |
+
"step": 117500
|
| 1419 |
+
},
|
| 1420 |
+
{
|
| 1421 |
+
"epoch": 93.95,
|
| 1422 |
+
"learning_rate": 1.8281847133757964e-06,
|
| 1423 |
+
"loss": 1.5299,
|
| 1424 |
+
"step": 118000
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 94.35,
|
| 1428 |
+
"learning_rate": 1.7087579617834395e-06,
|
| 1429 |
+
"loss": 1.5027,
|
| 1430 |
+
"step": 118500
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 94.75,
|
| 1434 |
+
"learning_rate": 1.5893312101910827e-06,
|
| 1435 |
+
"loss": 1.5207,
|
| 1436 |
+
"step": 119000
|
| 1437 |
+
},
|
| 1438 |
+
{
|
| 1439 |
+
"epoch": 95.14,
|
| 1440 |
+
"learning_rate": 1.469904458598726e-06,
|
| 1441 |
+
"loss": 1.3726,
|
| 1442 |
+
"step": 119500
|
| 1443 |
+
},
|
| 1444 |
+
{
|
| 1445 |
+
"epoch": 95.54,
|
| 1446 |
+
"learning_rate": 1.3504777070063694e-06,
|
| 1447 |
+
"loss": 1.3402,
|
| 1448 |
+
"step": 120000
|
| 1449 |
+
},
|
| 1450 |
+
{
|
| 1451 |
+
"epoch": 95.94,
|
| 1452 |
+
"learning_rate": 1.2312898089171974e-06,
|
| 1453 |
+
"loss": 1.7752,
|
| 1454 |
+
"step": 120500
|
| 1455 |
+
},
|
| 1456 |
+
{
|
| 1457 |
+
"epoch": 96.34,
|
| 1458 |
+
"learning_rate": 1.1121019108280256e-06,
|
| 1459 |
+
"loss": 1.6895,
|
| 1460 |
+
"step": 121000
|
| 1461 |
+
},
|
| 1462 |
+
{
|
| 1463 |
+
"epoch": 96.74,
|
| 1464 |
+
"learning_rate": 9.926751592356687e-07,
|
| 1465 |
+
"loss": 1.633,
|
| 1466 |
+
"step": 121500
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 97.13,
|
| 1470 |
+
"learning_rate": 8.732484076433121e-07,
|
| 1471 |
+
"loss": 1.6033,
|
| 1472 |
+
"step": 122000
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 97.53,
|
| 1476 |
+
"learning_rate": 7.538216560509554e-07,
|
| 1477 |
+
"loss": 1.3726,
|
| 1478 |
+
"step": 122500
|
| 1479 |
+
},
|
| 1480 |
+
{
|
| 1481 |
+
"epoch": 97.93,
|
| 1482 |
+
"learning_rate": 6.343949044585987e-07,
|
| 1483 |
+
"loss": 1.3725,
|
| 1484 |
+
"step": 123000
|
| 1485 |
+
},
|
| 1486 |
+
{
|
| 1487 |
+
"epoch": 98.33,
|
| 1488 |
+
"learning_rate": 5.149681528662421e-07,
|
| 1489 |
+
"loss": 1.5944,
|
| 1490 |
+
"step": 123500
|
| 1491 |
+
},
|
| 1492 |
+
{
|
| 1493 |
+
"epoch": 98.73,
|
| 1494 |
+
"learning_rate": 3.9554140127388536e-07,
|
| 1495 |
+
"loss": 1.466,
|
| 1496 |
+
"step": 124000
|
| 1497 |
+
},
|
| 1498 |
+
{
|
| 1499 |
+
"epoch": 99.12,
|
| 1500 |
+
"learning_rate": 2.7611464968152867e-07,
|
| 1501 |
+
"loss": 1.421,
|
| 1502 |
+
"step": 124500
|
| 1503 |
+
},
|
| 1504 |
+
{
|
| 1505 |
+
"epoch": 99.52,
|
| 1506 |
+
"learning_rate": 1.569267515923567e-07,
|
| 1507 |
+
"loss": 1.6345,
|
| 1508 |
+
"step": 125000
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 99.92,
|
| 1512 |
+
"learning_rate": 3.7500000000000005e-08,
|
| 1513 |
+
"loss": 1.5021,
|
| 1514 |
+
"step": 125500
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 100.0,
|
| 1518 |
+
"step": 125600,
|
| 1519 |
+
"total_flos": 5.249087478816768e+17,
|
| 1520 |
+
"train_loss": 1.9559951608499904,
|
| 1521 |
+
"train_runtime": 50817.2182,
|
| 1522 |
+
"train_samples_per_second": 14.83,
|
| 1523 |
+
"train_steps_per_second": 2.472
|
| 1524 |
+
}
|
| 1525 |
+
],
|
| 1526 |
+
"max_steps": 125600,
|
| 1527 |
+
"num_train_epochs": 100,
|
| 1528 |
+
"total_flos": 5.249087478816768e+17,
|
| 1529 |
+
"trial_name": null,
|
| 1530 |
+
"trial_params": null
|
| 1531 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:722d35f28b4dd1e22b34520f074b1f7b2eff0b4ea5f06fa37c01d1bf17e85112
|
| 3 |
+
size 3439
|