Public Release
Browse files- all_results.json +15 -0
- config.json +25 -0
- eval_results.json +10 -0
- pytorch_model.bin +3 -0
- special_tokens_map.json +7 -0
- tokenizer.json +0 -0
- tokenizer_config.json +14 -0
- train_results.json +8 -0
- trainer_state.json +1147 -0
- training_args.bin +3 -0
- vocab.txt +0 -0
all_results.json
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 127.03,
|
| 3 |
+
"eval_accuracy": 0.7441575486960593,
|
| 4 |
+
"eval_loss": 1.09375,
|
| 5 |
+
"eval_runtime": 330.033,
|
| 6 |
+
"eval_samples": 127448,
|
| 7 |
+
"eval_samples_per_second": 386.167,
|
| 8 |
+
"eval_steps_per_second": 1.509,
|
| 9 |
+
"perplexity": 2.985448539365356,
|
| 10 |
+
"train_loss": 0.015288788411458334,
|
| 11 |
+
"train_runtime": 1005.2623,
|
| 12 |
+
"train_samples": 3021595,
|
| 13 |
+
"train_samples_per_second": 381989.844,
|
| 14 |
+
"train_steps_per_second": 186.518
|
| 15 |
+
}
|
config.json
ADDED
|
@@ -0,0 +1,25 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "/grand/projects/SolarWindowsADSP/pankaj/polaris_outputs/BERT-cased-from-scratch/checkpoint-185000",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"BertForMaskedLM"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"classifier_dropout": null,
|
| 8 |
+
"hidden_act": "gelu",
|
| 9 |
+
"hidden_dropout_prob": 0.1,
|
| 10 |
+
"hidden_size": 768,
|
| 11 |
+
"initializer_range": 0.02,
|
| 12 |
+
"intermediate_size": 3072,
|
| 13 |
+
"layer_norm_eps": 1e-12,
|
| 14 |
+
"max_position_embeddings": 512,
|
| 15 |
+
"model_type": "bert",
|
| 16 |
+
"num_attention_heads": 12,
|
| 17 |
+
"num_hidden_layers": 12,
|
| 18 |
+
"pad_token_id": 0,
|
| 19 |
+
"position_embedding_type": "absolute",
|
| 20 |
+
"torch_dtype": "float16",
|
| 21 |
+
"transformers_version": "4.25.1",
|
| 22 |
+
"type_vocab_size": 2,
|
| 23 |
+
"use_cache": true,
|
| 24 |
+
"vocab_size": 30522
|
| 25 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 127.03,
|
| 3 |
+
"eval_accuracy": 0.7441575486960593,
|
| 4 |
+
"eval_loss": 1.09375,
|
| 5 |
+
"eval_runtime": 330.033,
|
| 6 |
+
"eval_samples": 127448,
|
| 7 |
+
"eval_samples_per_second": 386.167,
|
| 8 |
+
"eval_steps_per_second": 1.509,
|
| 9 |
+
"perplexity": 2.985448539365356
|
| 10 |
+
}
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:65ac0c5e1b4aa32e7715508b24d2e25710d94e6d83880f9ab0ea22629c017b59
|
| 3 |
+
size 219061539
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"cls_token": "[CLS]",
|
| 3 |
+
"mask_token": "[MASK]",
|
| 4 |
+
"pad_token": "[PAD]",
|
| 5 |
+
"sep_token": "[SEP]",
|
| 6 |
+
"unk_token": "[UNK]"
|
| 7 |
+
}
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"cls_token": "[CLS]",
|
| 3 |
+
"do_lower_case": false,
|
| 4 |
+
"mask_token": "[MASK]",
|
| 5 |
+
"model_max_length": 512,
|
| 6 |
+
"name_or_path": "/grand/projects/SolarWindowsADSP/pankaj/polaris_outputs/tokenizer-bert-base-cased-all",
|
| 7 |
+
"pad_token": "[PAD]",
|
| 8 |
+
"sep_token": "[SEP]",
|
| 9 |
+
"special_tokens_map_file": null,
|
| 10 |
+
"strip_accents": null,
|
| 11 |
+
"tokenize_chinese_chars": true,
|
| 12 |
+
"tokenizer_class": "BertTokenizer",
|
| 13 |
+
"unk_token": "[UNK]"
|
| 14 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 127.03,
|
| 3 |
+
"train_loss": 0.015288788411458334,
|
| 4 |
+
"train_runtime": 1005.2623,
|
| 5 |
+
"train_samples": 3021595,
|
| 6 |
+
"train_samples_per_second": 381989.844,
|
| 7 |
+
"train_steps_per_second": 186.518
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1147 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 127.03252032520325,
|
| 5 |
+
"global_step": 187500,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.68,
|
| 12 |
+
"learning_rate": 7.864000000000001e-06,
|
| 13 |
+
"loss": 8.3573,
|
| 14 |
+
"step": 1000
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 1.36,
|
| 18 |
+
"learning_rate": 1.5856e-05,
|
| 19 |
+
"loss": 6.6355,
|
| 20 |
+
"step": 2000
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 2.03,
|
| 24 |
+
"learning_rate": 2.3848e-05,
|
| 25 |
+
"loss": 6.2822,
|
| 26 |
+
"step": 3000
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 2.71,
|
| 30 |
+
"learning_rate": 3.184000000000001e-05,
|
| 31 |
+
"loss": 6.0633,
|
| 32 |
+
"step": 4000
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 3.39,
|
| 36 |
+
"learning_rate": 3.9832e-05,
|
| 37 |
+
"loss": 5.9146,
|
| 38 |
+
"step": 5000
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 4.07,
|
| 42 |
+
"learning_rate": 4.7824e-05,
|
| 43 |
+
"loss": 5.8117,
|
| 44 |
+
"step": 6000
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 4.74,
|
| 48 |
+
"learning_rate": 5.5816e-05,
|
| 49 |
+
"loss": 5.7342,
|
| 50 |
+
"step": 7000
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 5.42,
|
| 54 |
+
"learning_rate": 6.380800000000001e-05,
|
| 55 |
+
"loss": 5.6725,
|
| 56 |
+
"step": 8000
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 6.1,
|
| 60 |
+
"learning_rate": 7.18e-05,
|
| 61 |
+
"loss": 5.6293,
|
| 62 |
+
"step": 9000
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 6.78,
|
| 66 |
+
"learning_rate": 7.9792e-05,
|
| 67 |
+
"loss": 5.594,
|
| 68 |
+
"step": 10000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 7.45,
|
| 72 |
+
"learning_rate": 7.95614647887324e-05,
|
| 73 |
+
"loss": 5.5631,
|
| 74 |
+
"step": 11000
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 8.13,
|
| 78 |
+
"learning_rate": 7.911121126760565e-05,
|
| 79 |
+
"loss": 5.5393,
|
| 80 |
+
"step": 12000
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 8.81,
|
| 84 |
+
"learning_rate": 7.866095774647888e-05,
|
| 85 |
+
"loss": 5.5223,
|
| 86 |
+
"step": 13000
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 9.49,
|
| 90 |
+
"learning_rate": 7.821070422535212e-05,
|
| 91 |
+
"loss": 5.5064,
|
| 92 |
+
"step": 14000
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 10.16,
|
| 96 |
+
"learning_rate": 7.776045070422535e-05,
|
| 97 |
+
"loss": 5.4923,
|
| 98 |
+
"step": 15000
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 10.84,
|
| 102 |
+
"learning_rate": 7.73101971830986e-05,
|
| 103 |
+
"loss": 5.482,
|
| 104 |
+
"step": 16000
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 11.52,
|
| 108 |
+
"learning_rate": 7.685994366197184e-05,
|
| 109 |
+
"loss": 5.4725,
|
| 110 |
+
"step": 17000
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 12.2,
|
| 114 |
+
"learning_rate": 7.640969014084507e-05,
|
| 115 |
+
"loss": 5.4627,
|
| 116 |
+
"step": 18000
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 12.87,
|
| 120 |
+
"learning_rate": 7.595943661971832e-05,
|
| 121 |
+
"loss": 5.4567,
|
| 122 |
+
"step": 19000
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 13.55,
|
| 126 |
+
"learning_rate": 7.550918309859156e-05,
|
| 127 |
+
"loss": 5.4456,
|
| 128 |
+
"step": 20000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 14.23,
|
| 132 |
+
"learning_rate": 7.505892957746479e-05,
|
| 133 |
+
"loss": 5.4407,
|
| 134 |
+
"step": 21000
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 14.91,
|
| 138 |
+
"learning_rate": 7.460867605633804e-05,
|
| 139 |
+
"loss": 5.4347,
|
| 140 |
+
"step": 22000
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 15.58,
|
| 144 |
+
"learning_rate": 7.415842253521126e-05,
|
| 145 |
+
"loss": 5.4284,
|
| 146 |
+
"step": 23000
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 16.26,
|
| 150 |
+
"learning_rate": 7.370816901408451e-05,
|
| 151 |
+
"loss": 5.424,
|
| 152 |
+
"step": 24000
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 16.94,
|
| 156 |
+
"learning_rate": 7.325791549295775e-05,
|
| 157 |
+
"loss": 5.1332,
|
| 158 |
+
"step": 25000
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 17.62,
|
| 162 |
+
"learning_rate": 7.2807661971831e-05,
|
| 163 |
+
"loss": 4.5278,
|
| 164 |
+
"step": 26000
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 18.29,
|
| 168 |
+
"learning_rate": 7.235740845070423e-05,
|
| 169 |
+
"loss": 3.2777,
|
| 170 |
+
"step": 27000
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 18.97,
|
| 174 |
+
"learning_rate": 7.190715492957747e-05,
|
| 175 |
+
"loss": 2.4646,
|
| 176 |
+
"step": 28000
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 19.65,
|
| 180 |
+
"learning_rate": 7.145690140845072e-05,
|
| 181 |
+
"loss": 2.2162,
|
| 182 |
+
"step": 29000
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 20.33,
|
| 186 |
+
"learning_rate": 7.100664788732395e-05,
|
| 187 |
+
"loss": 2.0715,
|
| 188 |
+
"step": 30000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 21.0,
|
| 192 |
+
"learning_rate": 7.055639436619719e-05,
|
| 193 |
+
"loss": 1.9643,
|
| 194 |
+
"step": 31000
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 21.68,
|
| 198 |
+
"learning_rate": 7.010614084507043e-05,
|
| 199 |
+
"loss": 1.8819,
|
| 200 |
+
"step": 32000
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 22.36,
|
| 204 |
+
"learning_rate": 6.965588732394366e-05,
|
| 205 |
+
"loss": 1.8187,
|
| 206 |
+
"step": 33000
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 23.04,
|
| 210 |
+
"learning_rate": 6.920563380281691e-05,
|
| 211 |
+
"loss": 1.7671,
|
| 212 |
+
"step": 34000
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 23.71,
|
| 216 |
+
"learning_rate": 6.875538028169015e-05,
|
| 217 |
+
"loss": 1.7254,
|
| 218 |
+
"step": 35000
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 24.39,
|
| 222 |
+
"learning_rate": 6.830512676056338e-05,
|
| 223 |
+
"loss": 1.6912,
|
| 224 |
+
"step": 36000
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 25.07,
|
| 228 |
+
"learning_rate": 6.785487323943663e-05,
|
| 229 |
+
"loss": 1.6589,
|
| 230 |
+
"step": 37000
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 25.75,
|
| 234 |
+
"learning_rate": 6.740461971830987e-05,
|
| 235 |
+
"loss": 1.631,
|
| 236 |
+
"step": 38000
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 26.42,
|
| 240 |
+
"learning_rate": 6.69543661971831e-05,
|
| 241 |
+
"loss": 1.606,
|
| 242 |
+
"step": 39000
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 27.1,
|
| 246 |
+
"learning_rate": 6.650411267605634e-05,
|
| 247 |
+
"loss": 1.584,
|
| 248 |
+
"step": 40000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 27.78,
|
| 252 |
+
"learning_rate": 6.605385915492959e-05,
|
| 253 |
+
"loss": 1.5641,
|
| 254 |
+
"step": 41000
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 28.46,
|
| 258 |
+
"learning_rate": 6.560360563380282e-05,
|
| 259 |
+
"loss": 1.5455,
|
| 260 |
+
"step": 42000
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 29.13,
|
| 264 |
+
"learning_rate": 6.515335211267606e-05,
|
| 265 |
+
"loss": 1.528,
|
| 266 |
+
"step": 43000
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 29.81,
|
| 270 |
+
"learning_rate": 6.47030985915493e-05,
|
| 271 |
+
"loss": 1.5117,
|
| 272 |
+
"step": 44000
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 30.49,
|
| 276 |
+
"learning_rate": 6.425284507042254e-05,
|
| 277 |
+
"loss": 1.4979,
|
| 278 |
+
"step": 45000
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 31.17,
|
| 282 |
+
"learning_rate": 6.380259154929578e-05,
|
| 283 |
+
"loss": 1.4843,
|
| 284 |
+
"step": 46000
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 31.84,
|
| 288 |
+
"learning_rate": 6.335233802816903e-05,
|
| 289 |
+
"loss": 1.4728,
|
| 290 |
+
"step": 47000
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 32.52,
|
| 294 |
+
"learning_rate": 6.290208450704226e-05,
|
| 295 |
+
"loss": 1.4601,
|
| 296 |
+
"step": 48000
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 33.2,
|
| 300 |
+
"learning_rate": 6.24518309859155e-05,
|
| 301 |
+
"loss": 1.4495,
|
| 302 |
+
"step": 49000
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 33.88,
|
| 306 |
+
"learning_rate": 6.200157746478873e-05,
|
| 307 |
+
"loss": 1.4395,
|
| 308 |
+
"step": 50000
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 34.55,
|
| 312 |
+
"learning_rate": 6.155132394366198e-05,
|
| 313 |
+
"loss": 1.43,
|
| 314 |
+
"step": 51000
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 35.23,
|
| 318 |
+
"learning_rate": 6.110107042253522e-05,
|
| 319 |
+
"loss": 1.4211,
|
| 320 |
+
"step": 52000
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 35.91,
|
| 324 |
+
"learning_rate": 6.0650816901408453e-05,
|
| 325 |
+
"loss": 1.4127,
|
| 326 |
+
"step": 53000
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 36.59,
|
| 330 |
+
"learning_rate": 6.0200563380281696e-05,
|
| 331 |
+
"loss": 1.4031,
|
| 332 |
+
"step": 54000
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 37.26,
|
| 336 |
+
"learning_rate": 5.975030985915493e-05,
|
| 337 |
+
"loss": 1.3951,
|
| 338 |
+
"step": 55000
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 37.94,
|
| 342 |
+
"learning_rate": 5.9300056338028174e-05,
|
| 343 |
+
"loss": 1.3876,
|
| 344 |
+
"step": 56000
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 38.62,
|
| 348 |
+
"learning_rate": 5.8849802816901416e-05,
|
| 349 |
+
"loss": 1.3801,
|
| 350 |
+
"step": 57000
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 39.3,
|
| 354 |
+
"learning_rate": 5.839954929577465e-05,
|
| 355 |
+
"loss": 1.3744,
|
| 356 |
+
"step": 58000
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 39.97,
|
| 360 |
+
"learning_rate": 5.7949295774647894e-05,
|
| 361 |
+
"loss": 1.3673,
|
| 362 |
+
"step": 59000
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 40.65,
|
| 366 |
+
"learning_rate": 5.7499042253521136e-05,
|
| 367 |
+
"loss": 1.3608,
|
| 368 |
+
"step": 60000
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 41.33,
|
| 372 |
+
"learning_rate": 5.7048788732394365e-05,
|
| 373 |
+
"loss": 1.3547,
|
| 374 |
+
"step": 61000
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 42.01,
|
| 378 |
+
"learning_rate": 5.659853521126761e-05,
|
| 379 |
+
"loss": 1.3496,
|
| 380 |
+
"step": 62000
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 42.68,
|
| 384 |
+
"learning_rate": 5.614828169014085e-05,
|
| 385 |
+
"loss": 1.3429,
|
| 386 |
+
"step": 63000
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 43.36,
|
| 390 |
+
"learning_rate": 5.569802816901409e-05,
|
| 391 |
+
"loss": 1.3381,
|
| 392 |
+
"step": 64000
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 44.04,
|
| 396 |
+
"learning_rate": 5.524777464788733e-05,
|
| 397 |
+
"loss": 1.3322,
|
| 398 |
+
"step": 65000
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 44.72,
|
| 402 |
+
"learning_rate": 5.479752112676057e-05,
|
| 403 |
+
"loss": 1.328,
|
| 404 |
+
"step": 66000
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 45.39,
|
| 408 |
+
"learning_rate": 5.434726760563381e-05,
|
| 409 |
+
"loss": 1.3233,
|
| 410 |
+
"step": 67000
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 46.07,
|
| 414 |
+
"learning_rate": 5.389701408450704e-05,
|
| 415 |
+
"loss": 1.319,
|
| 416 |
+
"step": 68000
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 46.75,
|
| 420 |
+
"learning_rate": 5.3446760563380284e-05,
|
| 421 |
+
"loss": 1.3135,
|
| 422 |
+
"step": 69000
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 47.43,
|
| 426 |
+
"learning_rate": 5.2996507042253526e-05,
|
| 427 |
+
"loss": 1.3088,
|
| 428 |
+
"step": 70000
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 48.1,
|
| 432 |
+
"learning_rate": 5.254625352112676e-05,
|
| 433 |
+
"loss": 1.305,
|
| 434 |
+
"step": 71000
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 48.78,
|
| 438 |
+
"learning_rate": 5.2096000000000004e-05,
|
| 439 |
+
"loss": 1.3006,
|
| 440 |
+
"step": 72000
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 49.46,
|
| 444 |
+
"learning_rate": 5.1645746478873246e-05,
|
| 445 |
+
"loss": 1.2968,
|
| 446 |
+
"step": 73000
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 50.14,
|
| 450 |
+
"learning_rate": 5.119549295774648e-05,
|
| 451 |
+
"loss": 1.2933,
|
| 452 |
+
"step": 74000
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 50.81,
|
| 456 |
+
"learning_rate": 5.0745239436619724e-05,
|
| 457 |
+
"loss": 1.2884,
|
| 458 |
+
"step": 75000
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 51.49,
|
| 462 |
+
"learning_rate": 5.029498591549297e-05,
|
| 463 |
+
"loss": 1.2853,
|
| 464 |
+
"step": 76000
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 52.17,
|
| 468 |
+
"learning_rate": 4.9844732394366195e-05,
|
| 469 |
+
"loss": 1.2819,
|
| 470 |
+
"step": 77000
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 52.85,
|
| 474 |
+
"learning_rate": 4.939447887323944e-05,
|
| 475 |
+
"loss": 1.2795,
|
| 476 |
+
"step": 78000
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 53.52,
|
| 480 |
+
"learning_rate": 4.894422535211268e-05,
|
| 481 |
+
"loss": 1.2754,
|
| 482 |
+
"step": 79000
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 54.2,
|
| 486 |
+
"learning_rate": 4.8493971830985916e-05,
|
| 487 |
+
"loss": 1.2715,
|
| 488 |
+
"step": 80000
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 54.88,
|
| 492 |
+
"learning_rate": 4.804371830985916e-05,
|
| 493 |
+
"loss": 1.2691,
|
| 494 |
+
"step": 81000
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 55.56,
|
| 498 |
+
"learning_rate": 4.75934647887324e-05,
|
| 499 |
+
"loss": 1.2652,
|
| 500 |
+
"step": 82000
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 56.23,
|
| 504 |
+
"learning_rate": 4.714321126760564e-05,
|
| 505 |
+
"loss": 1.2631,
|
| 506 |
+
"step": 83000
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 56.91,
|
| 510 |
+
"learning_rate": 4.669295774647888e-05,
|
| 511 |
+
"loss": 1.2598,
|
| 512 |
+
"step": 84000
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 57.59,
|
| 516 |
+
"learning_rate": 4.624270422535212e-05,
|
| 517 |
+
"loss": 1.2575,
|
| 518 |
+
"step": 85000
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 58.27,
|
| 522 |
+
"learning_rate": 4.5792450704225356e-05,
|
| 523 |
+
"loss": 1.2552,
|
| 524 |
+
"step": 86000
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 58.94,
|
| 528 |
+
"learning_rate": 4.534219718309859e-05,
|
| 529 |
+
"loss": 1.2509,
|
| 530 |
+
"step": 87000
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 59.62,
|
| 534 |
+
"learning_rate": 4.4891943661971834e-05,
|
| 535 |
+
"loss": 1.2492,
|
| 536 |
+
"step": 88000
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 60.3,
|
| 540 |
+
"learning_rate": 4.4441690140845077e-05,
|
| 541 |
+
"loss": 1.2468,
|
| 542 |
+
"step": 89000
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 60.98,
|
| 546 |
+
"learning_rate": 4.399143661971831e-05,
|
| 547 |
+
"loss": 1.2441,
|
| 548 |
+
"step": 90000
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 61.65,
|
| 552 |
+
"learning_rate": 4.3541183098591555e-05,
|
| 553 |
+
"loss": 1.2413,
|
| 554 |
+
"step": 91000
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 62.33,
|
| 558 |
+
"learning_rate": 4.30909295774648e-05,
|
| 559 |
+
"loss": 1.2402,
|
| 560 |
+
"step": 92000
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 63.01,
|
| 564 |
+
"learning_rate": 4.2640676056338026e-05,
|
| 565 |
+
"loss": 1.237,
|
| 566 |
+
"step": 93000
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 63.69,
|
| 570 |
+
"learning_rate": 4.219042253521127e-05,
|
| 571 |
+
"loss": 1.2347,
|
| 572 |
+
"step": 94000
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 64.36,
|
| 576 |
+
"learning_rate": 4.174016901408451e-05,
|
| 577 |
+
"loss": 1.2326,
|
| 578 |
+
"step": 95000
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 65.04,
|
| 582 |
+
"learning_rate": 4.1289915492957746e-05,
|
| 583 |
+
"loss": 1.231,
|
| 584 |
+
"step": 96000
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 65.72,
|
| 588 |
+
"learning_rate": 4.083966197183099e-05,
|
| 589 |
+
"loss": 1.2281,
|
| 590 |
+
"step": 97000
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 66.4,
|
| 594 |
+
"learning_rate": 4.038940845070423e-05,
|
| 595 |
+
"loss": 1.2273,
|
| 596 |
+
"step": 98000
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 67.07,
|
| 600 |
+
"learning_rate": 3.9939154929577466e-05,
|
| 601 |
+
"loss": 1.2251,
|
| 602 |
+
"step": 99000
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 67.75,
|
| 606 |
+
"learning_rate": 3.948890140845071e-05,
|
| 607 |
+
"loss": 1.2234,
|
| 608 |
+
"step": 100000
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 68.43,
|
| 612 |
+
"learning_rate": 3.903864788732395e-05,
|
| 613 |
+
"loss": 1.2208,
|
| 614 |
+
"step": 101000
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 69.11,
|
| 618 |
+
"learning_rate": 3.8588394366197187e-05,
|
| 619 |
+
"loss": 1.2195,
|
| 620 |
+
"step": 102000
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 69.78,
|
| 624 |
+
"learning_rate": 3.813814084507042e-05,
|
| 625 |
+
"loss": 1.2184,
|
| 626 |
+
"step": 103000
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 70.46,
|
| 630 |
+
"learning_rate": 3.7687887323943664e-05,
|
| 631 |
+
"loss": 1.2158,
|
| 632 |
+
"step": 104000
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 71.14,
|
| 636 |
+
"learning_rate": 3.723763380281691e-05,
|
| 637 |
+
"loss": 1.2136,
|
| 638 |
+
"step": 105000
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 71.82,
|
| 642 |
+
"learning_rate": 3.678738028169015e-05,
|
| 643 |
+
"loss": 1.2134,
|
| 644 |
+
"step": 106000
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 72.49,
|
| 648 |
+
"learning_rate": 3.6337126760563385e-05,
|
| 649 |
+
"loss": 1.2108,
|
| 650 |
+
"step": 107000
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 73.17,
|
| 654 |
+
"learning_rate": 3.588687323943662e-05,
|
| 655 |
+
"loss": 1.2091,
|
| 656 |
+
"step": 108000
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 73.85,
|
| 660 |
+
"learning_rate": 3.543661971830986e-05,
|
| 661 |
+
"loss": 1.2071,
|
| 662 |
+
"step": 109000
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 74.53,
|
| 666 |
+
"learning_rate": 3.49863661971831e-05,
|
| 667 |
+
"loss": 1.2061,
|
| 668 |
+
"step": 110000
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 75.2,
|
| 672 |
+
"learning_rate": 3.453611267605634e-05,
|
| 673 |
+
"loss": 1.2041,
|
| 674 |
+
"step": 111000
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 75.88,
|
| 678 |
+
"learning_rate": 3.408585915492958e-05,
|
| 679 |
+
"loss": 1.2019,
|
| 680 |
+
"step": 112000
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 76.56,
|
| 684 |
+
"learning_rate": 3.363560563380282e-05,
|
| 685 |
+
"loss": 1.2012,
|
| 686 |
+
"step": 113000
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 77.24,
|
| 690 |
+
"learning_rate": 3.3185352112676054e-05,
|
| 691 |
+
"loss": 1.1996,
|
| 692 |
+
"step": 114000
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 77.91,
|
| 696 |
+
"learning_rate": 3.2735098591549296e-05,
|
| 697 |
+
"loss": 1.1986,
|
| 698 |
+
"step": 115000
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 78.59,
|
| 702 |
+
"learning_rate": 3.228484507042254e-05,
|
| 703 |
+
"loss": 1.1973,
|
| 704 |
+
"step": 116000
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 79.27,
|
| 708 |
+
"learning_rate": 3.183459154929578e-05,
|
| 709 |
+
"loss": 1.1962,
|
| 710 |
+
"step": 117000
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 79.95,
|
| 714 |
+
"learning_rate": 3.138433802816902e-05,
|
| 715 |
+
"loss": 1.194,
|
| 716 |
+
"step": 118000
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 80.62,
|
| 720 |
+
"learning_rate": 3.093408450704225e-05,
|
| 721 |
+
"loss": 1.1932,
|
| 722 |
+
"step": 119000
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 81.3,
|
| 726 |
+
"learning_rate": 3.0483830985915498e-05,
|
| 727 |
+
"loss": 1.1914,
|
| 728 |
+
"step": 120000
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 81.98,
|
| 732 |
+
"learning_rate": 3.0033577464788734e-05,
|
| 733 |
+
"loss": 1.1906,
|
| 734 |
+
"step": 121000
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 82.66,
|
| 738 |
+
"learning_rate": 2.9583323943661973e-05,
|
| 739 |
+
"loss": 1.1879,
|
| 740 |
+
"step": 122000
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 83.33,
|
| 744 |
+
"learning_rate": 2.9133070422535215e-05,
|
| 745 |
+
"loss": 1.189,
|
| 746 |
+
"step": 123000
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 84.01,
|
| 750 |
+
"learning_rate": 2.8682816901408454e-05,
|
| 751 |
+
"loss": 1.1872,
|
| 752 |
+
"step": 124000
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 84.69,
|
| 756 |
+
"learning_rate": 2.8232563380281693e-05,
|
| 757 |
+
"loss": 1.186,
|
| 758 |
+
"step": 125000
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 85.37,
|
| 762 |
+
"learning_rate": 2.7782309859154932e-05,
|
| 763 |
+
"loss": 1.1848,
|
| 764 |
+
"step": 126000
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 86.04,
|
| 768 |
+
"learning_rate": 2.733205633802817e-05,
|
| 769 |
+
"loss": 1.1835,
|
| 770 |
+
"step": 127000
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 86.72,
|
| 774 |
+
"learning_rate": 2.6881802816901413e-05,
|
| 775 |
+
"loss": 1.1831,
|
| 776 |
+
"step": 128000
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 87.4,
|
| 780 |
+
"learning_rate": 2.643154929577465e-05,
|
| 781 |
+
"loss": 1.1811,
|
| 782 |
+
"step": 129000
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 88.08,
|
| 786 |
+
"learning_rate": 2.5981295774647888e-05,
|
| 787 |
+
"loss": 1.1803,
|
| 788 |
+
"step": 130000
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 88.75,
|
| 792 |
+
"learning_rate": 2.553104225352113e-05,
|
| 793 |
+
"loss": 1.1789,
|
| 794 |
+
"step": 131000
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 89.43,
|
| 798 |
+
"learning_rate": 2.508078873239437e-05,
|
| 799 |
+
"loss": 1.1791,
|
| 800 |
+
"step": 132000
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 90.11,
|
| 804 |
+
"learning_rate": 2.4630535211267605e-05,
|
| 805 |
+
"loss": 1.1769,
|
| 806 |
+
"step": 133000
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 90.79,
|
| 810 |
+
"learning_rate": 2.4180281690140847e-05,
|
| 811 |
+
"loss": 1.1766,
|
| 812 |
+
"step": 134000
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 91.46,
|
| 816 |
+
"learning_rate": 2.3730028169014086e-05,
|
| 817 |
+
"loss": 1.1757,
|
| 818 |
+
"step": 135000
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 92.14,
|
| 822 |
+
"learning_rate": 2.327977464788733e-05,
|
| 823 |
+
"loss": 1.1741,
|
| 824 |
+
"step": 136000
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 92.82,
|
| 828 |
+
"learning_rate": 2.2829521126760564e-05,
|
| 829 |
+
"loss": 1.1734,
|
| 830 |
+
"step": 137000
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 93.5,
|
| 834 |
+
"learning_rate": 2.2379267605633803e-05,
|
| 835 |
+
"loss": 1.1733,
|
| 836 |
+
"step": 138000
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 94.17,
|
| 840 |
+
"learning_rate": 2.1929014084507045e-05,
|
| 841 |
+
"loss": 1.1723,
|
| 842 |
+
"step": 139000
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 94.85,
|
| 846 |
+
"learning_rate": 2.1478760563380284e-05,
|
| 847 |
+
"loss": 1.1715,
|
| 848 |
+
"step": 140000
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 95.53,
|
| 852 |
+
"learning_rate": 2.102850704225352e-05,
|
| 853 |
+
"loss": 1.1699,
|
| 854 |
+
"step": 141000
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 96.21,
|
| 858 |
+
"learning_rate": 2.0578253521126762e-05,
|
| 859 |
+
"loss": 1.1689,
|
| 860 |
+
"step": 142000
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 96.88,
|
| 864 |
+
"learning_rate": 2.0128e-05,
|
| 865 |
+
"loss": 1.1691,
|
| 866 |
+
"step": 143000
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 97.56,
|
| 870 |
+
"learning_rate": 1.967774647887324e-05,
|
| 871 |
+
"loss": 1.1675,
|
| 872 |
+
"step": 144000
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 98.24,
|
| 876 |
+
"learning_rate": 1.9227492957746482e-05,
|
| 877 |
+
"loss": 1.1671,
|
| 878 |
+
"step": 145000
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 98.92,
|
| 882 |
+
"learning_rate": 1.877723943661972e-05,
|
| 883 |
+
"loss": 1.167,
|
| 884 |
+
"step": 146000
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 99.59,
|
| 888 |
+
"learning_rate": 1.8326985915492957e-05,
|
| 889 |
+
"loss": 1.1659,
|
| 890 |
+
"step": 147000
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 100.27,
|
| 894 |
+
"learning_rate": 1.78767323943662e-05,
|
| 895 |
+
"loss": 1.1653,
|
| 896 |
+
"step": 148000
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 100.95,
|
| 900 |
+
"learning_rate": 1.7426478873239438e-05,
|
| 901 |
+
"loss": 1.1647,
|
| 902 |
+
"step": 149000
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 101.63,
|
| 906 |
+
"learning_rate": 1.6976225352112677e-05,
|
| 907 |
+
"loss": 1.1639,
|
| 908 |
+
"step": 150000
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 102.3,
|
| 912 |
+
"learning_rate": 1.6525971830985916e-05,
|
| 913 |
+
"loss": 1.163,
|
| 914 |
+
"step": 151000
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 102.98,
|
| 918 |
+
"learning_rate": 1.6075718309859155e-05,
|
| 919 |
+
"loss": 1.1615,
|
| 920 |
+
"step": 152000
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 103.66,
|
| 924 |
+
"learning_rate": 1.5625464788732398e-05,
|
| 925 |
+
"loss": 1.1615,
|
| 926 |
+
"step": 153000
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 104.34,
|
| 930 |
+
"learning_rate": 1.5175211267605635e-05,
|
| 931 |
+
"loss": 1.1595,
|
| 932 |
+
"step": 154000
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 105.01,
|
| 936 |
+
"learning_rate": 1.4724957746478874e-05,
|
| 937 |
+
"loss": 1.1603,
|
| 938 |
+
"step": 155000
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 105.69,
|
| 942 |
+
"learning_rate": 1.4274704225352114e-05,
|
| 943 |
+
"loss": 1.1597,
|
| 944 |
+
"step": 156000
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 106.37,
|
| 948 |
+
"learning_rate": 1.3824450704225353e-05,
|
| 949 |
+
"loss": 1.1591,
|
| 950 |
+
"step": 157000
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 107.05,
|
| 954 |
+
"learning_rate": 1.3374197183098592e-05,
|
| 955 |
+
"loss": 1.1586,
|
| 956 |
+
"step": 158000
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 107.72,
|
| 960 |
+
"learning_rate": 1.2923943661971831e-05,
|
| 961 |
+
"loss": 1.1576,
|
| 962 |
+
"step": 159000
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 108.4,
|
| 966 |
+
"learning_rate": 1.2473690140845072e-05,
|
| 967 |
+
"loss": 1.1563,
|
| 968 |
+
"step": 160000
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 109.08,
|
| 972 |
+
"learning_rate": 1.2023436619718311e-05,
|
| 973 |
+
"loss": 1.1559,
|
| 974 |
+
"step": 161000
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 109.76,
|
| 978 |
+
"learning_rate": 1.157318309859155e-05,
|
| 979 |
+
"loss": 1.1559,
|
| 980 |
+
"step": 162000
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 110.43,
|
| 984 |
+
"learning_rate": 1.1122929577464789e-05,
|
| 985 |
+
"loss": 1.1543,
|
| 986 |
+
"step": 163000
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 111.11,
|
| 990 |
+
"learning_rate": 1.067267605633803e-05,
|
| 991 |
+
"loss": 1.1553,
|
| 992 |
+
"step": 164000
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 111.79,
|
| 996 |
+
"learning_rate": 1.022242253521127e-05,
|
| 997 |
+
"loss": 1.154,
|
| 998 |
+
"step": 165000
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 112.47,
|
| 1002 |
+
"learning_rate": 9.772169014084507e-06,
|
| 1003 |
+
"loss": 1.1535,
|
| 1004 |
+
"step": 166000
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 113.14,
|
| 1008 |
+
"learning_rate": 9.321915492957746e-06,
|
| 1009 |
+
"loss": 1.1536,
|
| 1010 |
+
"step": 167000
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 113.82,
|
| 1014 |
+
"learning_rate": 8.871661971830987e-06,
|
| 1015 |
+
"loss": 1.1525,
|
| 1016 |
+
"step": 168000
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 114.5,
|
| 1020 |
+
"learning_rate": 8.421408450704226e-06,
|
| 1021 |
+
"loss": 1.1524,
|
| 1022 |
+
"step": 169000
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 115.18,
|
| 1026 |
+
"learning_rate": 7.971154929577467e-06,
|
| 1027 |
+
"loss": 1.152,
|
| 1028 |
+
"step": 170000
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 115.85,
|
| 1032 |
+
"learning_rate": 7.520901408450705e-06,
|
| 1033 |
+
"loss": 1.1514,
|
| 1034 |
+
"step": 171000
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 116.53,
|
| 1038 |
+
"learning_rate": 7.070647887323944e-06,
|
| 1039 |
+
"loss": 1.151,
|
| 1040 |
+
"step": 172000
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 117.21,
|
| 1044 |
+
"learning_rate": 6.620394366197184e-06,
|
| 1045 |
+
"loss": 1.1504,
|
| 1046 |
+
"step": 173000
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 117.89,
|
| 1050 |
+
"learning_rate": 6.170140845070423e-06,
|
| 1051 |
+
"loss": 1.1495,
|
| 1052 |
+
"step": 174000
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 118.56,
|
| 1056 |
+
"learning_rate": 5.719887323943662e-06,
|
| 1057 |
+
"loss": 1.1494,
|
| 1058 |
+
"step": 175000
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 119.24,
|
| 1062 |
+
"learning_rate": 5.269633802816901e-06,
|
| 1063 |
+
"loss": 1.1493,
|
| 1064 |
+
"step": 176000
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 119.92,
|
| 1068 |
+
"learning_rate": 4.819380281690141e-06,
|
| 1069 |
+
"loss": 1.149,
|
| 1070 |
+
"step": 177000
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 120.6,
|
| 1074 |
+
"learning_rate": 4.369126760563381e-06,
|
| 1075 |
+
"loss": 1.1485,
|
| 1076 |
+
"step": 178000
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 121.27,
|
| 1080 |
+
"learning_rate": 3.91887323943662e-06,
|
| 1081 |
+
"loss": 1.1482,
|
| 1082 |
+
"step": 179000
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 121.95,
|
| 1086 |
+
"learning_rate": 3.4686197183098598e-06,
|
| 1087 |
+
"loss": 1.148,
|
| 1088 |
+
"step": 180000
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 122.63,
|
| 1092 |
+
"learning_rate": 3.018366197183099e-06,
|
| 1093 |
+
"loss": 1.1477,
|
| 1094 |
+
"step": 181000
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 123.31,
|
| 1098 |
+
"learning_rate": 2.568112676056338e-06,
|
| 1099 |
+
"loss": 1.147,
|
| 1100 |
+
"step": 182000
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 123.98,
|
| 1104 |
+
"learning_rate": 2.1178591549295775e-06,
|
| 1105 |
+
"loss": 1.1466,
|
| 1106 |
+
"step": 183000
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 124.66,
|
| 1110 |
+
"learning_rate": 1.6676056338028171e-06,
|
| 1111 |
+
"loss": 1.1475,
|
| 1112 |
+
"step": 184000
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 125.34,
|
| 1116 |
+
"learning_rate": 1.2173521126760563e-06,
|
| 1117 |
+
"loss": 1.1459,
|
| 1118 |
+
"step": 185000
|
| 1119 |
+
},
|
| 1120 |
+
{
|
| 1121 |
+
"epoch": 126.02,
|
| 1122 |
+
"learning_rate": 7.670985915492958e-07,
|
| 1123 |
+
"loss": 1.1466,
|
| 1124 |
+
"step": 186000
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 126.69,
|
| 1128 |
+
"learning_rate": 3.1684507042253523e-07,
|
| 1129 |
+
"loss": 1.1462,
|
| 1130 |
+
"step": 187000
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 127.03,
|
| 1134 |
+
"step": 187500,
|
| 1135 |
+
"total_flos": 1.0102999800563001e+20,
|
| 1136 |
+
"train_loss": 0.015288788411458334,
|
| 1137 |
+
"train_runtime": 1005.2623,
|
| 1138 |
+
"train_samples_per_second": 381989.844,
|
| 1139 |
+
"train_steps_per_second": 186.518
|
| 1140 |
+
}
|
| 1141 |
+
],
|
| 1142 |
+
"max_steps": 187500,
|
| 1143 |
+
"num_train_epochs": 128,
|
| 1144 |
+
"total_flos": 1.0102999800563001e+20,
|
| 1145 |
+
"trial_name": null,
|
| 1146 |
+
"trial_params": null
|
| 1147 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ee8c43514f1fb076e72597a11051e1c6004c6743e285a911ac2542dec211a913
|
| 3 |
+
size 4271
|
vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|