Commit
•
d4bc4a9
1
Parent(s):
9676cd2
Training in progress, epoch 1
Browse files- config.json +34 -0
- model.safetensors +3 -0
- runs/Apr25_17-49-45_NCATS-2265430-P/events.out.tfevents.1714081800.NCATS-2265430-P.25784.0 +3 -0
- special_tokens_map.json +7 -0
- tmp-checkpoint-397/config.json +34 -0
- tmp-checkpoint-397/model.safetensors +3 -0
- tmp-checkpoint-397/optimizer.pt +3 -0
- tmp-checkpoint-397/rng_state.pth +3 -0
- tmp-checkpoint-397/scheduler.pt +3 -0
- tmp-checkpoint-397/special_tokens_map.json +7 -0
- tmp-checkpoint-397/tokenizer.json +0 -0
- tmp-checkpoint-397/tokenizer_config.json +57 -0
- tmp-checkpoint-397/trainer_state.json +2812 -0
- tmp-checkpoint-397/training_args.bin +3 -0
- tmp-checkpoint-397/vocab.txt +0 -0
- tokenizer.json +0 -0
- tokenizer_config.json +57 -0
- training_args.bin +3 -0
- vocab.txt +0 -0
config.json
ADDED
|
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "microsoft/BiomedNLP-BiomedBERT-base-uncased-abstract",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"BertForSequenceClassification"
|
| 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 |
+
"id2label": {
|
| 12 |
+
"0": "the paper is not a primary experimental study in rare disease or the study is not directly investigating the natural history of a disease",
|
| 13 |
+
"1": "its primary contribution centers on observing the time course of a rare disease"
|
| 14 |
+
},
|
| 15 |
+
"initializer_range": 0.02,
|
| 16 |
+
"intermediate_size": 3072,
|
| 17 |
+
"label2id": {
|
| 18 |
+
"its primary contribution centers on observing the time course of a rare disease": 1,
|
| 19 |
+
"the paper is not a primary experimental study in rare disease or the study is not directly investigating the natural history of a disease": 0
|
| 20 |
+
},
|
| 21 |
+
"layer_norm_eps": 1e-12,
|
| 22 |
+
"max_position_embeddings": 512,
|
| 23 |
+
"model_type": "bert",
|
| 24 |
+
"num_attention_heads": 12,
|
| 25 |
+
"num_hidden_layers": 12,
|
| 26 |
+
"pad_token_id": 0,
|
| 27 |
+
"position_embedding_type": "absolute",
|
| 28 |
+
"problem_type": "single_label_classification",
|
| 29 |
+
"torch_dtype": "float32",
|
| 30 |
+
"transformers_version": "4.38.2",
|
| 31 |
+
"type_vocab_size": 2,
|
| 32 |
+
"use_cache": true,
|
| 33 |
+
"vocab_size": 30522
|
| 34 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:950320f993d0d5c8cbecae5c4562a0e1188c4d6705e12e78eefc6279ad50b928
|
| 3 |
+
size 437958648
|
runs/Apr25_17-49-45_NCATS-2265430-P/events.out.tfevents.1714081800.NCATS-2265430-P.25784.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d6e51242ddad93e7ff2b907d65fa39748ea1a4e02db431236c3f0cf5412b3b1b
|
| 3 |
+
size 88904
|
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 |
+
}
|
tmp-checkpoint-397/config.json
ADDED
|
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "microsoft/BiomedNLP-BiomedBERT-base-uncased-abstract",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"BertForSequenceClassification"
|
| 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 |
+
"id2label": {
|
| 12 |
+
"0": "the paper is not a primary experimental study in rare disease or the study is not directly investigating the natural history of a disease",
|
| 13 |
+
"1": "its primary contribution centers on observing the time course of a rare disease"
|
| 14 |
+
},
|
| 15 |
+
"initializer_range": 0.02,
|
| 16 |
+
"intermediate_size": 3072,
|
| 17 |
+
"label2id": {
|
| 18 |
+
"its primary contribution centers on observing the time course of a rare disease": 1,
|
| 19 |
+
"the paper is not a primary experimental study in rare disease or the study is not directly investigating the natural history of a disease": 0
|
| 20 |
+
},
|
| 21 |
+
"layer_norm_eps": 1e-12,
|
| 22 |
+
"max_position_embeddings": 512,
|
| 23 |
+
"model_type": "bert",
|
| 24 |
+
"num_attention_heads": 12,
|
| 25 |
+
"num_hidden_layers": 12,
|
| 26 |
+
"pad_token_id": 0,
|
| 27 |
+
"position_embedding_type": "absolute",
|
| 28 |
+
"problem_type": "single_label_classification",
|
| 29 |
+
"torch_dtype": "float32",
|
| 30 |
+
"transformers_version": "4.38.2",
|
| 31 |
+
"type_vocab_size": 2,
|
| 32 |
+
"use_cache": true,
|
| 33 |
+
"vocab_size": 30522
|
| 34 |
+
}
|
tmp-checkpoint-397/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:950320f993d0d5c8cbecae5c4562a0e1188c4d6705e12e78eefc6279ad50b928
|
| 3 |
+
size 437958648
|
tmp-checkpoint-397/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:72f9c6bc2ada3603093a027f5e30dfa276677b091fb95c15d614640a5d4fed40
|
| 3 |
+
size 876032698
|
tmp-checkpoint-397/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1bea22e306bb9123207acf5d1b4932eb169b7b856f6aa40e9a3d8dcbfc07f6cb
|
| 3 |
+
size 13990
|
tmp-checkpoint-397/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bd75744655bb8dc28bb36c7457f2082d688170aba52437ee81f2d1f5516b79ad
|
| 3 |
+
size 1064
|
tmp-checkpoint-397/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 |
+
}
|
tmp-checkpoint-397/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tmp-checkpoint-397/tokenizer_config.json
ADDED
|
@@ -0,0 +1,57 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "[PAD]",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "[UNK]",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "[CLS]",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "[SEP]",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"4": {
|
| 36 |
+
"content": "[MASK]",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"clean_up_tokenization_spaces": true,
|
| 45 |
+
"cls_token": "[CLS]",
|
| 46 |
+
"do_basic_tokenize": true,
|
| 47 |
+
"do_lower_case": true,
|
| 48 |
+
"mask_token": "[MASK]",
|
| 49 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 50 |
+
"never_split": null,
|
| 51 |
+
"pad_token": "[PAD]",
|
| 52 |
+
"sep_token": "[SEP]",
|
| 53 |
+
"strip_accents": null,
|
| 54 |
+
"tokenize_chinese_chars": true,
|
| 55 |
+
"tokenizer_class": "BertTokenizer",
|
| 56 |
+
"unk_token": "[UNK]"
|
| 57 |
+
}
|
tmp-checkpoint-397/trainer_state.json
ADDED
|
@@ -0,0 +1,2812 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.4741531312465668,
|
| 3 |
+
"best_model_checkpoint": "NHS-BiomedNLP-BiomedBERT-hypop\\checkpoint-397",
|
| 4 |
+
"epoch": 1.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 397,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.0,
|
| 13 |
+
"grad_norm": 9.947269439697266,
|
| 14 |
+
"learning_rate": 2.9987405541561713e-05,
|
| 15 |
+
"loss": 0.7027,
|
| 16 |
+
"step": 1
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.01,
|
| 20 |
+
"grad_norm": 8.814902305603027,
|
| 21 |
+
"learning_rate": 2.997481108312343e-05,
|
| 22 |
+
"loss": 0.6097,
|
| 23 |
+
"step": 2
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.01,
|
| 27 |
+
"grad_norm": 14.668534278869629,
|
| 28 |
+
"learning_rate": 2.9962216624685138e-05,
|
| 29 |
+
"loss": 0.7259,
|
| 30 |
+
"step": 3
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.01,
|
| 34 |
+
"grad_norm": 6.714921951293945,
|
| 35 |
+
"learning_rate": 2.994962216624685e-05,
|
| 36 |
+
"loss": 0.6384,
|
| 37 |
+
"step": 4
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.01,
|
| 41 |
+
"grad_norm": 7.754816055297852,
|
| 42 |
+
"learning_rate": 2.9937027707808566e-05,
|
| 43 |
+
"loss": 0.5809,
|
| 44 |
+
"step": 5
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.02,
|
| 48 |
+
"grad_norm": 8.383195877075195,
|
| 49 |
+
"learning_rate": 2.992443324937028e-05,
|
| 50 |
+
"loss": 0.5908,
|
| 51 |
+
"step": 6
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.02,
|
| 55 |
+
"grad_norm": 6.874183654785156,
|
| 56 |
+
"learning_rate": 2.9911838790931988e-05,
|
| 57 |
+
"loss": 0.5422,
|
| 58 |
+
"step": 7
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.02,
|
| 62 |
+
"grad_norm": 7.232607841491699,
|
| 63 |
+
"learning_rate": 2.9899244332493704e-05,
|
| 64 |
+
"loss": 0.6751,
|
| 65 |
+
"step": 8
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.02,
|
| 69 |
+
"grad_norm": 8.095083236694336,
|
| 70 |
+
"learning_rate": 2.9886649874055416e-05,
|
| 71 |
+
"loss": 0.5845,
|
| 72 |
+
"step": 9
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.03,
|
| 76 |
+
"grad_norm": 10.181581497192383,
|
| 77 |
+
"learning_rate": 2.987405541561713e-05,
|
| 78 |
+
"loss": 0.4039,
|
| 79 |
+
"step": 10
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.03,
|
| 83 |
+
"grad_norm": 7.251787185668945,
|
| 84 |
+
"learning_rate": 2.9861460957178844e-05,
|
| 85 |
+
"loss": 0.5025,
|
| 86 |
+
"step": 11
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.03,
|
| 90 |
+
"grad_norm": 11.612983703613281,
|
| 91 |
+
"learning_rate": 2.9848866498740553e-05,
|
| 92 |
+
"loss": 0.658,
|
| 93 |
+
"step": 12
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.03,
|
| 97 |
+
"grad_norm": 13.866792678833008,
|
| 98 |
+
"learning_rate": 2.9836272040302266e-05,
|
| 99 |
+
"loss": 0.5777,
|
| 100 |
+
"step": 13
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.04,
|
| 104 |
+
"grad_norm": 9.340723037719727,
|
| 105 |
+
"learning_rate": 2.982367758186398e-05,
|
| 106 |
+
"loss": 0.6047,
|
| 107 |
+
"step": 14
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.04,
|
| 111 |
+
"grad_norm": 10.717991828918457,
|
| 112 |
+
"learning_rate": 2.9811083123425694e-05,
|
| 113 |
+
"loss": 0.5881,
|
| 114 |
+
"step": 15
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.04,
|
| 118 |
+
"grad_norm": 10.31982421875,
|
| 119 |
+
"learning_rate": 2.9798488664987406e-05,
|
| 120 |
+
"loss": 0.4501,
|
| 121 |
+
"step": 16
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.04,
|
| 125 |
+
"grad_norm": 9.57770824432373,
|
| 126 |
+
"learning_rate": 2.978589420654912e-05,
|
| 127 |
+
"loss": 0.4792,
|
| 128 |
+
"step": 17
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.05,
|
| 132 |
+
"grad_norm": 9.465398788452148,
|
| 133 |
+
"learning_rate": 2.977329974811083e-05,
|
| 134 |
+
"loss": 0.5718,
|
| 135 |
+
"step": 18
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.05,
|
| 139 |
+
"grad_norm": 16.856624603271484,
|
| 140 |
+
"learning_rate": 2.9760705289672544e-05,
|
| 141 |
+
"loss": 0.7295,
|
| 142 |
+
"step": 19
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.05,
|
| 146 |
+
"grad_norm": 11.629205703735352,
|
| 147 |
+
"learning_rate": 2.974811083123426e-05,
|
| 148 |
+
"loss": 0.5327,
|
| 149 |
+
"step": 20
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.05,
|
| 153 |
+
"grad_norm": 8.08765697479248,
|
| 154 |
+
"learning_rate": 2.973551637279597e-05,
|
| 155 |
+
"loss": 0.6436,
|
| 156 |
+
"step": 21
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.06,
|
| 160 |
+
"grad_norm": 11.187536239624023,
|
| 161 |
+
"learning_rate": 2.972292191435768e-05,
|
| 162 |
+
"loss": 0.5741,
|
| 163 |
+
"step": 22
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.06,
|
| 167 |
+
"grad_norm": 12.667988777160645,
|
| 168 |
+
"learning_rate": 2.9710327455919397e-05,
|
| 169 |
+
"loss": 0.5328,
|
| 170 |
+
"step": 23
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.06,
|
| 174 |
+
"grad_norm": 8.135946273803711,
|
| 175 |
+
"learning_rate": 2.969773299748111e-05,
|
| 176 |
+
"loss": 0.436,
|
| 177 |
+
"step": 24
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.06,
|
| 181 |
+
"grad_norm": 8.500922203063965,
|
| 182 |
+
"learning_rate": 2.968513853904282e-05,
|
| 183 |
+
"loss": 0.6446,
|
| 184 |
+
"step": 25
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.07,
|
| 188 |
+
"grad_norm": 8.38931941986084,
|
| 189 |
+
"learning_rate": 2.9672544080604534e-05,
|
| 190 |
+
"loss": 0.4106,
|
| 191 |
+
"step": 26
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.07,
|
| 195 |
+
"grad_norm": 7.952789783477783,
|
| 196 |
+
"learning_rate": 2.9659949622166247e-05,
|
| 197 |
+
"loss": 0.5269,
|
| 198 |
+
"step": 27
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.07,
|
| 202 |
+
"grad_norm": 9.8812255859375,
|
| 203 |
+
"learning_rate": 2.964735516372796e-05,
|
| 204 |
+
"loss": 0.5114,
|
| 205 |
+
"step": 28
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.07,
|
| 209 |
+
"grad_norm": 8.14013671875,
|
| 210 |
+
"learning_rate": 2.9634760705289675e-05,
|
| 211 |
+
"loss": 0.3894,
|
| 212 |
+
"step": 29
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.08,
|
| 216 |
+
"grad_norm": 10.428199768066406,
|
| 217 |
+
"learning_rate": 2.9622166246851387e-05,
|
| 218 |
+
"loss": 0.5967,
|
| 219 |
+
"step": 30
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.08,
|
| 223 |
+
"grad_norm": 9.175215721130371,
|
| 224 |
+
"learning_rate": 2.9609571788413096e-05,
|
| 225 |
+
"loss": 0.4978,
|
| 226 |
+
"step": 31
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.08,
|
| 230 |
+
"grad_norm": 8.572174072265625,
|
| 231 |
+
"learning_rate": 2.9596977329974812e-05,
|
| 232 |
+
"loss": 0.6002,
|
| 233 |
+
"step": 32
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.08,
|
| 237 |
+
"grad_norm": 7.942105293273926,
|
| 238 |
+
"learning_rate": 2.9584382871536524e-05,
|
| 239 |
+
"loss": 0.4905,
|
| 240 |
+
"step": 33
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.09,
|
| 244 |
+
"grad_norm": 8.92253303527832,
|
| 245 |
+
"learning_rate": 2.9571788413098237e-05,
|
| 246 |
+
"loss": 0.378,
|
| 247 |
+
"step": 34
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.09,
|
| 251 |
+
"grad_norm": 10.829577445983887,
|
| 252 |
+
"learning_rate": 2.955919395465995e-05,
|
| 253 |
+
"loss": 0.5745,
|
| 254 |
+
"step": 35
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.09,
|
| 258 |
+
"grad_norm": 7.60324764251709,
|
| 259 |
+
"learning_rate": 2.9546599496221662e-05,
|
| 260 |
+
"loss": 0.3207,
|
| 261 |
+
"step": 36
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.09,
|
| 265 |
+
"grad_norm": 6.186948776245117,
|
| 266 |
+
"learning_rate": 2.9534005037783378e-05,
|
| 267 |
+
"loss": 0.292,
|
| 268 |
+
"step": 37
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.1,
|
| 272 |
+
"grad_norm": 7.360220909118652,
|
| 273 |
+
"learning_rate": 2.952141057934509e-05,
|
| 274 |
+
"loss": 0.7034,
|
| 275 |
+
"step": 38
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.1,
|
| 279 |
+
"grad_norm": 13.63427734375,
|
| 280 |
+
"learning_rate": 2.9508816120906802e-05,
|
| 281 |
+
"loss": 0.4843,
|
| 282 |
+
"step": 39
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.1,
|
| 286 |
+
"grad_norm": 11.388775825500488,
|
| 287 |
+
"learning_rate": 2.9496221662468515e-05,
|
| 288 |
+
"loss": 0.4537,
|
| 289 |
+
"step": 40
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.1,
|
| 293 |
+
"grad_norm": 13.440176963806152,
|
| 294 |
+
"learning_rate": 2.9483627204030227e-05,
|
| 295 |
+
"loss": 0.624,
|
| 296 |
+
"step": 41
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.11,
|
| 300 |
+
"grad_norm": 13.364521026611328,
|
| 301 |
+
"learning_rate": 2.947103274559194e-05,
|
| 302 |
+
"loss": 0.3167,
|
| 303 |
+
"step": 42
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.11,
|
| 307 |
+
"grad_norm": 9.809293746948242,
|
| 308 |
+
"learning_rate": 2.9458438287153656e-05,
|
| 309 |
+
"loss": 0.5945,
|
| 310 |
+
"step": 43
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.11,
|
| 314 |
+
"grad_norm": 14.221176147460938,
|
| 315 |
+
"learning_rate": 2.9445843828715368e-05,
|
| 316 |
+
"loss": 0.3922,
|
| 317 |
+
"step": 44
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.11,
|
| 321 |
+
"grad_norm": 9.054633140563965,
|
| 322 |
+
"learning_rate": 2.9433249370277077e-05,
|
| 323 |
+
"loss": 0.4583,
|
| 324 |
+
"step": 45
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.12,
|
| 328 |
+
"grad_norm": 15.587240219116211,
|
| 329 |
+
"learning_rate": 2.9420654911838793e-05,
|
| 330 |
+
"loss": 0.3764,
|
| 331 |
+
"step": 46
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.12,
|
| 335 |
+
"grad_norm": 19.746070861816406,
|
| 336 |
+
"learning_rate": 2.9408060453400505e-05,
|
| 337 |
+
"loss": 0.3708,
|
| 338 |
+
"step": 47
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.12,
|
| 342 |
+
"grad_norm": 7.772696018218994,
|
| 343 |
+
"learning_rate": 2.9395465994962218e-05,
|
| 344 |
+
"loss": 0.3178,
|
| 345 |
+
"step": 48
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.12,
|
| 349 |
+
"grad_norm": 13.086981773376465,
|
| 350 |
+
"learning_rate": 2.938287153652393e-05,
|
| 351 |
+
"loss": 0.4907,
|
| 352 |
+
"step": 49
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.13,
|
| 356 |
+
"grad_norm": 8.691929817199707,
|
| 357 |
+
"learning_rate": 2.9370277078085643e-05,
|
| 358 |
+
"loss": 0.2839,
|
| 359 |
+
"step": 50
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.13,
|
| 363 |
+
"grad_norm": 9.099030494689941,
|
| 364 |
+
"learning_rate": 2.9357682619647355e-05,
|
| 365 |
+
"loss": 0.4787,
|
| 366 |
+
"step": 51
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.13,
|
| 370 |
+
"grad_norm": 17.202730178833008,
|
| 371 |
+
"learning_rate": 2.934508816120907e-05,
|
| 372 |
+
"loss": 0.5272,
|
| 373 |
+
"step": 52
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.13,
|
| 377 |
+
"grad_norm": 14.722758293151855,
|
| 378 |
+
"learning_rate": 2.9332493702770783e-05,
|
| 379 |
+
"loss": 0.3865,
|
| 380 |
+
"step": 53
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.14,
|
| 384 |
+
"grad_norm": 8.29536247253418,
|
| 385 |
+
"learning_rate": 2.9319899244332492e-05,
|
| 386 |
+
"loss": 0.6003,
|
| 387 |
+
"step": 54
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.14,
|
| 391 |
+
"grad_norm": 10.03609848022461,
|
| 392 |
+
"learning_rate": 2.9307304785894208e-05,
|
| 393 |
+
"loss": 0.5586,
|
| 394 |
+
"step": 55
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.14,
|
| 398 |
+
"grad_norm": 15.111539840698242,
|
| 399 |
+
"learning_rate": 2.929471032745592e-05,
|
| 400 |
+
"loss": 0.3587,
|
| 401 |
+
"step": 56
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.14,
|
| 405 |
+
"grad_norm": 10.219131469726562,
|
| 406 |
+
"learning_rate": 2.9282115869017633e-05,
|
| 407 |
+
"loss": 0.4541,
|
| 408 |
+
"step": 57
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.15,
|
| 412 |
+
"grad_norm": 9.36807918548584,
|
| 413 |
+
"learning_rate": 2.926952141057935e-05,
|
| 414 |
+
"loss": 0.448,
|
| 415 |
+
"step": 58
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.15,
|
| 419 |
+
"grad_norm": 12.47349739074707,
|
| 420 |
+
"learning_rate": 2.9256926952141058e-05,
|
| 421 |
+
"loss": 0.585,
|
| 422 |
+
"step": 59
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.15,
|
| 426 |
+
"grad_norm": 7.583310604095459,
|
| 427 |
+
"learning_rate": 2.924433249370277e-05,
|
| 428 |
+
"loss": 0.2274,
|
| 429 |
+
"step": 60
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.15,
|
| 433 |
+
"grad_norm": 13.765729904174805,
|
| 434 |
+
"learning_rate": 2.9231738035264486e-05,
|
| 435 |
+
"loss": 0.4415,
|
| 436 |
+
"step": 61
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.16,
|
| 440 |
+
"grad_norm": 22.002029418945312,
|
| 441 |
+
"learning_rate": 2.92191435768262e-05,
|
| 442 |
+
"loss": 0.4643,
|
| 443 |
+
"step": 62
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.16,
|
| 447 |
+
"grad_norm": 15.840831756591797,
|
| 448 |
+
"learning_rate": 2.9206549118387908e-05,
|
| 449 |
+
"loss": 0.9231,
|
| 450 |
+
"step": 63
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.16,
|
| 454 |
+
"grad_norm": 22.16878318786621,
|
| 455 |
+
"learning_rate": 2.9193954659949623e-05,
|
| 456 |
+
"loss": 0.6477,
|
| 457 |
+
"step": 64
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.16,
|
| 461 |
+
"grad_norm": 10.195225715637207,
|
| 462 |
+
"learning_rate": 2.9181360201511336e-05,
|
| 463 |
+
"loss": 0.4786,
|
| 464 |
+
"step": 65
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.17,
|
| 468 |
+
"grad_norm": 11.7352294921875,
|
| 469 |
+
"learning_rate": 2.9168765743073048e-05,
|
| 470 |
+
"loss": 0.4849,
|
| 471 |
+
"step": 66
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.17,
|
| 475 |
+
"grad_norm": 12.2020902633667,
|
| 476 |
+
"learning_rate": 2.9156171284634764e-05,
|
| 477 |
+
"loss": 0.3908,
|
| 478 |
+
"step": 67
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.17,
|
| 482 |
+
"grad_norm": 13.499194145202637,
|
| 483 |
+
"learning_rate": 2.9143576826196473e-05,
|
| 484 |
+
"loss": 0.6328,
|
| 485 |
+
"step": 68
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.17,
|
| 489 |
+
"grad_norm": 20.535781860351562,
|
| 490 |
+
"learning_rate": 2.9130982367758185e-05,
|
| 491 |
+
"loss": 0.6886,
|
| 492 |
+
"step": 69
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.18,
|
| 496 |
+
"grad_norm": 11.518393516540527,
|
| 497 |
+
"learning_rate": 2.91183879093199e-05,
|
| 498 |
+
"loss": 0.6269,
|
| 499 |
+
"step": 70
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.18,
|
| 503 |
+
"grad_norm": 10.17136001586914,
|
| 504 |
+
"learning_rate": 2.9105793450881614e-05,
|
| 505 |
+
"loss": 0.5077,
|
| 506 |
+
"step": 71
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.18,
|
| 510 |
+
"grad_norm": 6.502359390258789,
|
| 511 |
+
"learning_rate": 2.9093198992443326e-05,
|
| 512 |
+
"loss": 0.5269,
|
| 513 |
+
"step": 72
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.18,
|
| 517 |
+
"grad_norm": 16.268077850341797,
|
| 518 |
+
"learning_rate": 2.908060453400504e-05,
|
| 519 |
+
"loss": 0.6138,
|
| 520 |
+
"step": 73
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.19,
|
| 524 |
+
"grad_norm": 11.319491386413574,
|
| 525 |
+
"learning_rate": 2.906801007556675e-05,
|
| 526 |
+
"loss": 0.4758,
|
| 527 |
+
"step": 74
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.19,
|
| 531 |
+
"grad_norm": 7.1629319190979,
|
| 532 |
+
"learning_rate": 2.9055415617128463e-05,
|
| 533 |
+
"loss": 0.4717,
|
| 534 |
+
"step": 75
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.19,
|
| 538 |
+
"grad_norm": 17.689434051513672,
|
| 539 |
+
"learning_rate": 2.904282115869018e-05,
|
| 540 |
+
"loss": 0.6062,
|
| 541 |
+
"step": 76
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.19,
|
| 545 |
+
"grad_norm": 16.246288299560547,
|
| 546 |
+
"learning_rate": 2.903022670025189e-05,
|
| 547 |
+
"loss": 0.4756,
|
| 548 |
+
"step": 77
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.2,
|
| 552 |
+
"grad_norm": 12.266759872436523,
|
| 553 |
+
"learning_rate": 2.90176322418136e-05,
|
| 554 |
+
"loss": 0.4271,
|
| 555 |
+
"step": 78
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.2,
|
| 559 |
+
"grad_norm": 7.64720344543457,
|
| 560 |
+
"learning_rate": 2.9005037783375317e-05,
|
| 561 |
+
"loss": 0.5505,
|
| 562 |
+
"step": 79
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.2,
|
| 566 |
+
"grad_norm": 6.0152153968811035,
|
| 567 |
+
"learning_rate": 2.899244332493703e-05,
|
| 568 |
+
"loss": 0.4172,
|
| 569 |
+
"step": 80
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.2,
|
| 573 |
+
"grad_norm": 10.39532470703125,
|
| 574 |
+
"learning_rate": 2.897984886649874e-05,
|
| 575 |
+
"loss": 0.659,
|
| 576 |
+
"step": 81
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 0.21,
|
| 580 |
+
"grad_norm": 6.066320896148682,
|
| 581 |
+
"learning_rate": 2.8967254408060454e-05,
|
| 582 |
+
"loss": 0.5022,
|
| 583 |
+
"step": 82
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.21,
|
| 587 |
+
"grad_norm": 9.511946678161621,
|
| 588 |
+
"learning_rate": 2.8954659949622166e-05,
|
| 589 |
+
"loss": 0.3605,
|
| 590 |
+
"step": 83
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.21,
|
| 594 |
+
"grad_norm": 5.521829605102539,
|
| 595 |
+
"learning_rate": 2.894206549118388e-05,
|
| 596 |
+
"loss": 0.2804,
|
| 597 |
+
"step": 84
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.21,
|
| 601 |
+
"grad_norm": 6.357683181762695,
|
| 602 |
+
"learning_rate": 2.8929471032745595e-05,
|
| 603 |
+
"loss": 0.343,
|
| 604 |
+
"step": 85
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 0.22,
|
| 608 |
+
"grad_norm": 4.500522136688232,
|
| 609 |
+
"learning_rate": 2.8916876574307307e-05,
|
| 610 |
+
"loss": 0.2611,
|
| 611 |
+
"step": 86
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.22,
|
| 615 |
+
"grad_norm": 7.982411861419678,
|
| 616 |
+
"learning_rate": 2.8904282115869016e-05,
|
| 617 |
+
"loss": 0.291,
|
| 618 |
+
"step": 87
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.22,
|
| 622 |
+
"grad_norm": 5.47086763381958,
|
| 623 |
+
"learning_rate": 2.8891687657430732e-05,
|
| 624 |
+
"loss": 0.2171,
|
| 625 |
+
"step": 88
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.22,
|
| 629 |
+
"grad_norm": 6.41118860244751,
|
| 630 |
+
"learning_rate": 2.8879093198992444e-05,
|
| 631 |
+
"loss": 0.2698,
|
| 632 |
+
"step": 89
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.23,
|
| 636 |
+
"grad_norm": 7.404595375061035,
|
| 637 |
+
"learning_rate": 2.8866498740554157e-05,
|
| 638 |
+
"loss": 0.3499,
|
| 639 |
+
"step": 90
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.23,
|
| 643 |
+
"grad_norm": 9.283653259277344,
|
| 644 |
+
"learning_rate": 2.885390428211587e-05,
|
| 645 |
+
"loss": 0.3149,
|
| 646 |
+
"step": 91
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.23,
|
| 650 |
+
"grad_norm": 4.450091361999512,
|
| 651 |
+
"learning_rate": 2.884130982367758e-05,
|
| 652 |
+
"loss": 0.2221,
|
| 653 |
+
"step": 92
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.23,
|
| 657 |
+
"grad_norm": 13.71849536895752,
|
| 658 |
+
"learning_rate": 2.8828715365239294e-05,
|
| 659 |
+
"loss": 0.3377,
|
| 660 |
+
"step": 93
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 0.24,
|
| 664 |
+
"grad_norm": 5.049820423126221,
|
| 665 |
+
"learning_rate": 2.881612090680101e-05,
|
| 666 |
+
"loss": 0.1135,
|
| 667 |
+
"step": 94
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.24,
|
| 671 |
+
"grad_norm": 10.600882530212402,
|
| 672 |
+
"learning_rate": 2.8803526448362722e-05,
|
| 673 |
+
"loss": 0.313,
|
| 674 |
+
"step": 95
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.24,
|
| 678 |
+
"grad_norm": 9.384568214416504,
|
| 679 |
+
"learning_rate": 2.879093198992443e-05,
|
| 680 |
+
"loss": 0.4214,
|
| 681 |
+
"step": 96
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.24,
|
| 685 |
+
"grad_norm": 14.095718383789062,
|
| 686 |
+
"learning_rate": 2.8778337531486147e-05,
|
| 687 |
+
"loss": 0.3608,
|
| 688 |
+
"step": 97
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.25,
|
| 692 |
+
"grad_norm": 12.552332878112793,
|
| 693 |
+
"learning_rate": 2.876574307304786e-05,
|
| 694 |
+
"loss": 0.6839,
|
| 695 |
+
"step": 98
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.25,
|
| 699 |
+
"grad_norm": 12.589719772338867,
|
| 700 |
+
"learning_rate": 2.8753148614609572e-05,
|
| 701 |
+
"loss": 0.3692,
|
| 702 |
+
"step": 99
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 0.25,
|
| 706 |
+
"grad_norm": 10.545365333557129,
|
| 707 |
+
"learning_rate": 2.8740554156171288e-05,
|
| 708 |
+
"loss": 0.6336,
|
| 709 |
+
"step": 100
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.25,
|
| 713 |
+
"grad_norm": 4.529080390930176,
|
| 714 |
+
"learning_rate": 2.8727959697732997e-05,
|
| 715 |
+
"loss": 0.0726,
|
| 716 |
+
"step": 101
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.26,
|
| 720 |
+
"grad_norm": 10.031196594238281,
|
| 721 |
+
"learning_rate": 2.871536523929471e-05,
|
| 722 |
+
"loss": 0.5852,
|
| 723 |
+
"step": 102
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.26,
|
| 727 |
+
"grad_norm": 7.78066873550415,
|
| 728 |
+
"learning_rate": 2.8702770780856425e-05,
|
| 729 |
+
"loss": 0.3664,
|
| 730 |
+
"step": 103
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.26,
|
| 734 |
+
"grad_norm": 9.381865501403809,
|
| 735 |
+
"learning_rate": 2.8690176322418137e-05,
|
| 736 |
+
"loss": 0.6564,
|
| 737 |
+
"step": 104
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.26,
|
| 741 |
+
"grad_norm": 11.74254035949707,
|
| 742 |
+
"learning_rate": 2.8677581863979846e-05,
|
| 743 |
+
"loss": 0.4658,
|
| 744 |
+
"step": 105
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 0.27,
|
| 748 |
+
"grad_norm": 11.351922988891602,
|
| 749 |
+
"learning_rate": 2.8664987405541562e-05,
|
| 750 |
+
"loss": 0.5164,
|
| 751 |
+
"step": 106
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.27,
|
| 755 |
+
"grad_norm": 13.590174674987793,
|
| 756 |
+
"learning_rate": 2.8652392947103275e-05,
|
| 757 |
+
"loss": 0.5728,
|
| 758 |
+
"step": 107
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.27,
|
| 762 |
+
"grad_norm": 10.585956573486328,
|
| 763 |
+
"learning_rate": 2.8639798488664987e-05,
|
| 764 |
+
"loss": 0.2159,
|
| 765 |
+
"step": 108
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.27,
|
| 769 |
+
"grad_norm": 5.4417219161987305,
|
| 770 |
+
"learning_rate": 2.8627204030226703e-05,
|
| 771 |
+
"loss": 0.3464,
|
| 772 |
+
"step": 109
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 0.28,
|
| 776 |
+
"grad_norm": 8.342070579528809,
|
| 777 |
+
"learning_rate": 2.8614609571788412e-05,
|
| 778 |
+
"loss": 0.5985,
|
| 779 |
+
"step": 110
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.28,
|
| 783 |
+
"grad_norm": 6.407668590545654,
|
| 784 |
+
"learning_rate": 2.8602015113350128e-05,
|
| 785 |
+
"loss": 0.347,
|
| 786 |
+
"step": 111
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 0.28,
|
| 790 |
+
"grad_norm": 9.523656845092773,
|
| 791 |
+
"learning_rate": 2.858942065491184e-05,
|
| 792 |
+
"loss": 0.4596,
|
| 793 |
+
"step": 112
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.28,
|
| 797 |
+
"grad_norm": 7.752206325531006,
|
| 798 |
+
"learning_rate": 2.8576826196473553e-05,
|
| 799 |
+
"loss": 0.5771,
|
| 800 |
+
"step": 113
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.29,
|
| 804 |
+
"grad_norm": 10.980051040649414,
|
| 805 |
+
"learning_rate": 2.8564231738035265e-05,
|
| 806 |
+
"loss": 0.7089,
|
| 807 |
+
"step": 114
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.29,
|
| 811 |
+
"grad_norm": 7.721608638763428,
|
| 812 |
+
"learning_rate": 2.8551637279596978e-05,
|
| 813 |
+
"loss": 0.3957,
|
| 814 |
+
"step": 115
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 0.29,
|
| 818 |
+
"grad_norm": 5.920357704162598,
|
| 819 |
+
"learning_rate": 2.853904282115869e-05,
|
| 820 |
+
"loss": 0.3956,
|
| 821 |
+
"step": 116
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.29,
|
| 825 |
+
"grad_norm": 6.167881965637207,
|
| 826 |
+
"learning_rate": 2.8526448362720406e-05,
|
| 827 |
+
"loss": 0.2437,
|
| 828 |
+
"step": 117
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 0.3,
|
| 832 |
+
"grad_norm": 18.24216651916504,
|
| 833 |
+
"learning_rate": 2.8513853904282118e-05,
|
| 834 |
+
"loss": 0.6398,
|
| 835 |
+
"step": 118
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.3,
|
| 839 |
+
"grad_norm": 9.355070114135742,
|
| 840 |
+
"learning_rate": 2.8501259445843827e-05,
|
| 841 |
+
"loss": 0.7392,
|
| 842 |
+
"step": 119
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.3,
|
| 846 |
+
"grad_norm": 7.304947853088379,
|
| 847 |
+
"learning_rate": 2.8488664987405543e-05,
|
| 848 |
+
"loss": 0.5611,
|
| 849 |
+
"step": 120
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.3,
|
| 853 |
+
"grad_norm": 6.211945533752441,
|
| 854 |
+
"learning_rate": 2.8476070528967256e-05,
|
| 855 |
+
"loss": 0.3727,
|
| 856 |
+
"step": 121
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 0.31,
|
| 860 |
+
"grad_norm": 11.621819496154785,
|
| 861 |
+
"learning_rate": 2.8463476070528968e-05,
|
| 862 |
+
"loss": 0.5283,
|
| 863 |
+
"step": 122
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.31,
|
| 867 |
+
"grad_norm": 7.048833847045898,
|
| 868 |
+
"learning_rate": 2.8450881612090684e-05,
|
| 869 |
+
"loss": 0.6998,
|
| 870 |
+
"step": 123
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 0.31,
|
| 874 |
+
"grad_norm": 4.535017967224121,
|
| 875 |
+
"learning_rate": 2.8438287153652393e-05,
|
| 876 |
+
"loss": 0.431,
|
| 877 |
+
"step": 124
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.31,
|
| 881 |
+
"grad_norm": 4.307721138000488,
|
| 882 |
+
"learning_rate": 2.8425692695214105e-05,
|
| 883 |
+
"loss": 0.2508,
|
| 884 |
+
"step": 125
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.32,
|
| 888 |
+
"grad_norm": 7.520541191101074,
|
| 889 |
+
"learning_rate": 2.841309823677582e-05,
|
| 890 |
+
"loss": 0.2839,
|
| 891 |
+
"step": 126
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.32,
|
| 895 |
+
"grad_norm": 6.359663486480713,
|
| 896 |
+
"learning_rate": 2.8400503778337533e-05,
|
| 897 |
+
"loss": 0.2315,
|
| 898 |
+
"step": 127
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 0.32,
|
| 902 |
+
"grad_norm": 5.880284786224365,
|
| 903 |
+
"learning_rate": 2.8387909319899243e-05,
|
| 904 |
+
"loss": 0.2354,
|
| 905 |
+
"step": 128
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 0.32,
|
| 909 |
+
"grad_norm": 14.193618774414062,
|
| 910 |
+
"learning_rate": 2.837531486146096e-05,
|
| 911 |
+
"loss": 0.7256,
|
| 912 |
+
"step": 129
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"epoch": 0.33,
|
| 916 |
+
"grad_norm": 4.47697639465332,
|
| 917 |
+
"learning_rate": 2.836272040302267e-05,
|
| 918 |
+
"loss": 0.1521,
|
| 919 |
+
"step": 130
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.33,
|
| 923 |
+
"grad_norm": 8.092729568481445,
|
| 924 |
+
"learning_rate": 2.8350125944584383e-05,
|
| 925 |
+
"loss": 0.6047,
|
| 926 |
+
"step": 131
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 0.33,
|
| 930 |
+
"grad_norm": 7.471700668334961,
|
| 931 |
+
"learning_rate": 2.83375314861461e-05,
|
| 932 |
+
"loss": 0.3359,
|
| 933 |
+
"step": 132
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.34,
|
| 937 |
+
"grad_norm": 21.315391540527344,
|
| 938 |
+
"learning_rate": 2.8324937027707808e-05,
|
| 939 |
+
"loss": 0.6584,
|
| 940 |
+
"step": 133
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 0.34,
|
| 944 |
+
"grad_norm": 18.247236251831055,
|
| 945 |
+
"learning_rate": 2.831234256926952e-05,
|
| 946 |
+
"loss": 0.7327,
|
| 947 |
+
"step": 134
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.34,
|
| 951 |
+
"grad_norm": 13.097492218017578,
|
| 952 |
+
"learning_rate": 2.8299748110831236e-05,
|
| 953 |
+
"loss": 0.6226,
|
| 954 |
+
"step": 135
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 0.34,
|
| 958 |
+
"grad_norm": 5.948911190032959,
|
| 959 |
+
"learning_rate": 2.828715365239295e-05,
|
| 960 |
+
"loss": 0.3375,
|
| 961 |
+
"step": 136
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.35,
|
| 965 |
+
"grad_norm": 13.252030372619629,
|
| 966 |
+
"learning_rate": 2.827455919395466e-05,
|
| 967 |
+
"loss": 0.2128,
|
| 968 |
+
"step": 137
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.35,
|
| 972 |
+
"grad_norm": 10.255802154541016,
|
| 973 |
+
"learning_rate": 2.8261964735516374e-05,
|
| 974 |
+
"loss": 0.1734,
|
| 975 |
+
"step": 138
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.35,
|
| 979 |
+
"grad_norm": 7.478372097015381,
|
| 980 |
+
"learning_rate": 2.8249370277078086e-05,
|
| 981 |
+
"loss": 0.6492,
|
| 982 |
+
"step": 139
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 0.35,
|
| 986 |
+
"grad_norm": 10.903780937194824,
|
| 987 |
+
"learning_rate": 2.82367758186398e-05,
|
| 988 |
+
"loss": 0.9533,
|
| 989 |
+
"step": 140
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 0.36,
|
| 993 |
+
"grad_norm": 8.005816459655762,
|
| 994 |
+
"learning_rate": 2.8224181360201514e-05,
|
| 995 |
+
"loss": 0.3773,
|
| 996 |
+
"step": 141
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 0.36,
|
| 1000 |
+
"grad_norm": 8.27034854888916,
|
| 1001 |
+
"learning_rate": 2.8211586901763223e-05,
|
| 1002 |
+
"loss": 0.6595,
|
| 1003 |
+
"step": 142
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 0.36,
|
| 1007 |
+
"grad_norm": 7.85991907119751,
|
| 1008 |
+
"learning_rate": 2.8198992443324936e-05,
|
| 1009 |
+
"loss": 0.3601,
|
| 1010 |
+
"step": 143
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.36,
|
| 1014 |
+
"grad_norm": 7.255641460418701,
|
| 1015 |
+
"learning_rate": 2.818639798488665e-05,
|
| 1016 |
+
"loss": 0.4024,
|
| 1017 |
+
"step": 144
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.37,
|
| 1021 |
+
"grad_norm": 14.83194351196289,
|
| 1022 |
+
"learning_rate": 2.8173803526448364e-05,
|
| 1023 |
+
"loss": 0.6188,
|
| 1024 |
+
"step": 145
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 0.37,
|
| 1028 |
+
"grad_norm": 5.861283779144287,
|
| 1029 |
+
"learning_rate": 2.8161209068010076e-05,
|
| 1030 |
+
"loss": 0.493,
|
| 1031 |
+
"step": 146
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 0.37,
|
| 1035 |
+
"grad_norm": 5.700588226318359,
|
| 1036 |
+
"learning_rate": 2.814861460957179e-05,
|
| 1037 |
+
"loss": 0.3955,
|
| 1038 |
+
"step": 147
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 0.37,
|
| 1042 |
+
"grad_norm": 13.155322074890137,
|
| 1043 |
+
"learning_rate": 2.81360201511335e-05,
|
| 1044 |
+
"loss": 0.4269,
|
| 1045 |
+
"step": 148
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 0.38,
|
| 1049 |
+
"grad_norm": 8.21208381652832,
|
| 1050 |
+
"learning_rate": 2.8123425692695214e-05,
|
| 1051 |
+
"loss": 0.4352,
|
| 1052 |
+
"step": 149
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 0.38,
|
| 1056 |
+
"grad_norm": 6.507381439208984,
|
| 1057 |
+
"learning_rate": 2.811083123425693e-05,
|
| 1058 |
+
"loss": 0.2584,
|
| 1059 |
+
"step": 150
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.38,
|
| 1063 |
+
"grad_norm": 9.123990058898926,
|
| 1064 |
+
"learning_rate": 2.8098236775818642e-05,
|
| 1065 |
+
"loss": 0.5207,
|
| 1066 |
+
"step": 151
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 0.38,
|
| 1070 |
+
"grad_norm": 10.706199645996094,
|
| 1071 |
+
"learning_rate": 2.808564231738035e-05,
|
| 1072 |
+
"loss": 0.4213,
|
| 1073 |
+
"step": 152
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 0.39,
|
| 1077 |
+
"grad_norm": 5.8628621101379395,
|
| 1078 |
+
"learning_rate": 2.8073047858942067e-05,
|
| 1079 |
+
"loss": 0.3451,
|
| 1080 |
+
"step": 153
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 0.39,
|
| 1084 |
+
"grad_norm": 7.158113956451416,
|
| 1085 |
+
"learning_rate": 2.806045340050378e-05,
|
| 1086 |
+
"loss": 0.3274,
|
| 1087 |
+
"step": 154
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 0.39,
|
| 1091 |
+
"grad_norm": 5.258652210235596,
|
| 1092 |
+
"learning_rate": 2.804785894206549e-05,
|
| 1093 |
+
"loss": 0.3959,
|
| 1094 |
+
"step": 155
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 0.39,
|
| 1098 |
+
"grad_norm": 3.823733329772949,
|
| 1099 |
+
"learning_rate": 2.8035264483627204e-05,
|
| 1100 |
+
"loss": 0.2523,
|
| 1101 |
+
"step": 156
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 0.4,
|
| 1105 |
+
"grad_norm": 8.84554386138916,
|
| 1106 |
+
"learning_rate": 2.8022670025188917e-05,
|
| 1107 |
+
"loss": 0.553,
|
| 1108 |
+
"step": 157
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 0.4,
|
| 1112 |
+
"grad_norm": 6.735053539276123,
|
| 1113 |
+
"learning_rate": 2.801007556675063e-05,
|
| 1114 |
+
"loss": 0.393,
|
| 1115 |
+
"step": 158
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 0.4,
|
| 1119 |
+
"grad_norm": 4.995819091796875,
|
| 1120 |
+
"learning_rate": 2.7997481108312345e-05,
|
| 1121 |
+
"loss": 0.4338,
|
| 1122 |
+
"step": 159
|
| 1123 |
+
},
|
| 1124 |
+
{
|
| 1125 |
+
"epoch": 0.4,
|
| 1126 |
+
"grad_norm": 7.378593444824219,
|
| 1127 |
+
"learning_rate": 2.7984886649874057e-05,
|
| 1128 |
+
"loss": 0.613,
|
| 1129 |
+
"step": 160
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 0.41,
|
| 1133 |
+
"grad_norm": 6.643065452575684,
|
| 1134 |
+
"learning_rate": 2.7972292191435766e-05,
|
| 1135 |
+
"loss": 0.2731,
|
| 1136 |
+
"step": 161
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 0.41,
|
| 1140 |
+
"grad_norm": 9.81127643585205,
|
| 1141 |
+
"learning_rate": 2.7959697732997482e-05,
|
| 1142 |
+
"loss": 0.3793,
|
| 1143 |
+
"step": 162
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 0.41,
|
| 1147 |
+
"grad_norm": 10.082533836364746,
|
| 1148 |
+
"learning_rate": 2.7947103274559194e-05,
|
| 1149 |
+
"loss": 0.6018,
|
| 1150 |
+
"step": 163
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 0.41,
|
| 1154 |
+
"grad_norm": 5.660975933074951,
|
| 1155 |
+
"learning_rate": 2.7934508816120907e-05,
|
| 1156 |
+
"loss": 0.3733,
|
| 1157 |
+
"step": 164
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 0.42,
|
| 1161 |
+
"grad_norm": 12.785099983215332,
|
| 1162 |
+
"learning_rate": 2.7921914357682623e-05,
|
| 1163 |
+
"loss": 0.7492,
|
| 1164 |
+
"step": 165
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 0.42,
|
| 1168 |
+
"grad_norm": 8.646162986755371,
|
| 1169 |
+
"learning_rate": 2.7909319899244332e-05,
|
| 1170 |
+
"loss": 0.5155,
|
| 1171 |
+
"step": 166
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 0.42,
|
| 1175 |
+
"grad_norm": 7.465343952178955,
|
| 1176 |
+
"learning_rate": 2.7896725440806044e-05,
|
| 1177 |
+
"loss": 0.4195,
|
| 1178 |
+
"step": 167
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 0.42,
|
| 1182 |
+
"grad_norm": 6.121210098266602,
|
| 1183 |
+
"learning_rate": 2.788413098236776e-05,
|
| 1184 |
+
"loss": 0.3969,
|
| 1185 |
+
"step": 168
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 0.43,
|
| 1189 |
+
"grad_norm": 8.49669361114502,
|
| 1190 |
+
"learning_rate": 2.7871536523929472e-05,
|
| 1191 |
+
"loss": 0.4192,
|
| 1192 |
+
"step": 169
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 0.43,
|
| 1196 |
+
"grad_norm": 4.441570281982422,
|
| 1197 |
+
"learning_rate": 2.785894206549118e-05,
|
| 1198 |
+
"loss": 0.2789,
|
| 1199 |
+
"step": 170
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 0.43,
|
| 1203 |
+
"grad_norm": 6.6631245613098145,
|
| 1204 |
+
"learning_rate": 2.7846347607052897e-05,
|
| 1205 |
+
"loss": 0.4081,
|
| 1206 |
+
"step": 171
|
| 1207 |
+
},
|
| 1208 |
+
{
|
| 1209 |
+
"epoch": 0.43,
|
| 1210 |
+
"grad_norm": 5.974854469299316,
|
| 1211 |
+
"learning_rate": 2.783375314861461e-05,
|
| 1212 |
+
"loss": 0.3799,
|
| 1213 |
+
"step": 172
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 0.44,
|
| 1217 |
+
"grad_norm": 4.523766040802002,
|
| 1218 |
+
"learning_rate": 2.7821158690176322e-05,
|
| 1219 |
+
"loss": 0.1571,
|
| 1220 |
+
"step": 173
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 0.44,
|
| 1224 |
+
"grad_norm": 6.361337184906006,
|
| 1225 |
+
"learning_rate": 2.7808564231738038e-05,
|
| 1226 |
+
"loss": 0.3327,
|
| 1227 |
+
"step": 174
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 0.44,
|
| 1231 |
+
"grad_norm": 8.463950157165527,
|
| 1232 |
+
"learning_rate": 2.7795969773299747e-05,
|
| 1233 |
+
"loss": 0.5044,
|
| 1234 |
+
"step": 175
|
| 1235 |
+
},
|
| 1236 |
+
{
|
| 1237 |
+
"epoch": 0.44,
|
| 1238 |
+
"grad_norm": 13.052382469177246,
|
| 1239 |
+
"learning_rate": 2.778337531486146e-05,
|
| 1240 |
+
"loss": 0.4517,
|
| 1241 |
+
"step": 176
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 0.45,
|
| 1245 |
+
"grad_norm": 22.617353439331055,
|
| 1246 |
+
"learning_rate": 2.7770780856423175e-05,
|
| 1247 |
+
"loss": 1.0641,
|
| 1248 |
+
"step": 177
|
| 1249 |
+
},
|
| 1250 |
+
{
|
| 1251 |
+
"epoch": 0.45,
|
| 1252 |
+
"grad_norm": 9.117269515991211,
|
| 1253 |
+
"learning_rate": 2.7758186397984888e-05,
|
| 1254 |
+
"loss": 0.3886,
|
| 1255 |
+
"step": 178
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 0.45,
|
| 1259 |
+
"grad_norm": 12.824740409851074,
|
| 1260 |
+
"learning_rate": 2.77455919395466e-05,
|
| 1261 |
+
"loss": 0.3126,
|
| 1262 |
+
"step": 179
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 0.45,
|
| 1266 |
+
"grad_norm": 7.019243240356445,
|
| 1267 |
+
"learning_rate": 2.7732997481108313e-05,
|
| 1268 |
+
"loss": 0.2975,
|
| 1269 |
+
"step": 180
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 0.46,
|
| 1273 |
+
"grad_norm": 7.184720039367676,
|
| 1274 |
+
"learning_rate": 2.7720403022670025e-05,
|
| 1275 |
+
"loss": 0.5372,
|
| 1276 |
+
"step": 181
|
| 1277 |
+
},
|
| 1278 |
+
{
|
| 1279 |
+
"epoch": 0.46,
|
| 1280 |
+
"grad_norm": 9.207572937011719,
|
| 1281 |
+
"learning_rate": 2.770780856423174e-05,
|
| 1282 |
+
"loss": 0.4707,
|
| 1283 |
+
"step": 182
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 0.46,
|
| 1287 |
+
"grad_norm": 10.323878288269043,
|
| 1288 |
+
"learning_rate": 2.7695214105793453e-05,
|
| 1289 |
+
"loss": 0.4206,
|
| 1290 |
+
"step": 183
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 0.46,
|
| 1294 |
+
"grad_norm": 9.704578399658203,
|
| 1295 |
+
"learning_rate": 2.7682619647355162e-05,
|
| 1296 |
+
"loss": 0.4494,
|
| 1297 |
+
"step": 184
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 0.47,
|
| 1301 |
+
"grad_norm": 7.675583362579346,
|
| 1302 |
+
"learning_rate": 2.7670025188916878e-05,
|
| 1303 |
+
"loss": 0.2075,
|
| 1304 |
+
"step": 185
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 0.47,
|
| 1308 |
+
"grad_norm": 5.299776554107666,
|
| 1309 |
+
"learning_rate": 2.765743073047859e-05,
|
| 1310 |
+
"loss": 0.2851,
|
| 1311 |
+
"step": 186
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 0.47,
|
| 1315 |
+
"grad_norm": 8.457651138305664,
|
| 1316 |
+
"learning_rate": 2.7644836272040303e-05,
|
| 1317 |
+
"loss": 0.4121,
|
| 1318 |
+
"step": 187
|
| 1319 |
+
},
|
| 1320 |
+
{
|
| 1321 |
+
"epoch": 0.47,
|
| 1322 |
+
"grad_norm": 5.211854934692383,
|
| 1323 |
+
"learning_rate": 2.763224181360202e-05,
|
| 1324 |
+
"loss": 0.1637,
|
| 1325 |
+
"step": 188
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 0.48,
|
| 1329 |
+
"grad_norm": 18.794628143310547,
|
| 1330 |
+
"learning_rate": 2.7619647355163728e-05,
|
| 1331 |
+
"loss": 0.4972,
|
| 1332 |
+
"step": 189
|
| 1333 |
+
},
|
| 1334 |
+
{
|
| 1335 |
+
"epoch": 0.48,
|
| 1336 |
+
"grad_norm": 8.111665725708008,
|
| 1337 |
+
"learning_rate": 2.760705289672544e-05,
|
| 1338 |
+
"loss": 0.3703,
|
| 1339 |
+
"step": 190
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 0.48,
|
| 1343 |
+
"grad_norm": 14.346935272216797,
|
| 1344 |
+
"learning_rate": 2.7594458438287156e-05,
|
| 1345 |
+
"loss": 0.4615,
|
| 1346 |
+
"step": 191
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 0.48,
|
| 1350 |
+
"grad_norm": 4.972769260406494,
|
| 1351 |
+
"learning_rate": 2.758186397984887e-05,
|
| 1352 |
+
"loss": 0.1846,
|
| 1353 |
+
"step": 192
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 0.49,
|
| 1357 |
+
"grad_norm": 10.939284324645996,
|
| 1358 |
+
"learning_rate": 2.756926952141058e-05,
|
| 1359 |
+
"loss": 0.9461,
|
| 1360 |
+
"step": 193
|
| 1361 |
+
},
|
| 1362 |
+
{
|
| 1363 |
+
"epoch": 0.49,
|
| 1364 |
+
"grad_norm": 18.650239944458008,
|
| 1365 |
+
"learning_rate": 2.7556675062972293e-05,
|
| 1366 |
+
"loss": 0.6751,
|
| 1367 |
+
"step": 194
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 0.49,
|
| 1371 |
+
"grad_norm": 21.654159545898438,
|
| 1372 |
+
"learning_rate": 2.7544080604534006e-05,
|
| 1373 |
+
"loss": 0.6841,
|
| 1374 |
+
"step": 195
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 0.49,
|
| 1378 |
+
"grad_norm": 20.820980072021484,
|
| 1379 |
+
"learning_rate": 2.7531486146095718e-05,
|
| 1380 |
+
"loss": 0.4773,
|
| 1381 |
+
"step": 196
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 0.5,
|
| 1385 |
+
"grad_norm": 22.232995986938477,
|
| 1386 |
+
"learning_rate": 2.7518891687657434e-05,
|
| 1387 |
+
"loss": 0.8045,
|
| 1388 |
+
"step": 197
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 0.5,
|
| 1392 |
+
"grad_norm": 24.284286499023438,
|
| 1393 |
+
"learning_rate": 2.7506297229219143e-05,
|
| 1394 |
+
"loss": 0.574,
|
| 1395 |
+
"step": 198
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 0.5,
|
| 1399 |
+
"grad_norm": 7.939319610595703,
|
| 1400 |
+
"learning_rate": 2.7493702770780855e-05,
|
| 1401 |
+
"loss": 0.2071,
|
| 1402 |
+
"step": 199
|
| 1403 |
+
},
|
| 1404 |
+
{
|
| 1405 |
+
"epoch": 0.5,
|
| 1406 |
+
"grad_norm": 13.728672981262207,
|
| 1407 |
+
"learning_rate": 2.748110831234257e-05,
|
| 1408 |
+
"loss": 0.7237,
|
| 1409 |
+
"step": 200
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 0.51,
|
| 1413 |
+
"grad_norm": 8.86567497253418,
|
| 1414 |
+
"learning_rate": 2.7468513853904284e-05,
|
| 1415 |
+
"loss": 0.2757,
|
| 1416 |
+
"step": 201
|
| 1417 |
+
},
|
| 1418 |
+
{
|
| 1419 |
+
"epoch": 0.51,
|
| 1420 |
+
"grad_norm": 8.1141357421875,
|
| 1421 |
+
"learning_rate": 2.7455919395465996e-05,
|
| 1422 |
+
"loss": 0.2384,
|
| 1423 |
+
"step": 202
|
| 1424 |
+
},
|
| 1425 |
+
{
|
| 1426 |
+
"epoch": 0.51,
|
| 1427 |
+
"grad_norm": 11.418375968933105,
|
| 1428 |
+
"learning_rate": 2.744332493702771e-05,
|
| 1429 |
+
"loss": 0.7057,
|
| 1430 |
+
"step": 203
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 0.51,
|
| 1434 |
+
"grad_norm": 14.610424995422363,
|
| 1435 |
+
"learning_rate": 2.743073047858942e-05,
|
| 1436 |
+
"loss": 0.3707,
|
| 1437 |
+
"step": 204
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 0.52,
|
| 1441 |
+
"grad_norm": 8.905698776245117,
|
| 1442 |
+
"learning_rate": 2.7418136020151133e-05,
|
| 1443 |
+
"loss": 0.6057,
|
| 1444 |
+
"step": 205
|
| 1445 |
+
},
|
| 1446 |
+
{
|
| 1447 |
+
"epoch": 0.52,
|
| 1448 |
+
"grad_norm": 20.418991088867188,
|
| 1449 |
+
"learning_rate": 2.740554156171285e-05,
|
| 1450 |
+
"loss": 0.8607,
|
| 1451 |
+
"step": 206
|
| 1452 |
+
},
|
| 1453 |
+
{
|
| 1454 |
+
"epoch": 0.52,
|
| 1455 |
+
"grad_norm": 13.551456451416016,
|
| 1456 |
+
"learning_rate": 2.7392947103274562e-05,
|
| 1457 |
+
"loss": 0.532,
|
| 1458 |
+
"step": 207
|
| 1459 |
+
},
|
| 1460 |
+
{
|
| 1461 |
+
"epoch": 0.52,
|
| 1462 |
+
"grad_norm": 13.535616874694824,
|
| 1463 |
+
"learning_rate": 2.738035264483627e-05,
|
| 1464 |
+
"loss": 0.6038,
|
| 1465 |
+
"step": 208
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 0.53,
|
| 1469 |
+
"grad_norm": 11.406067848205566,
|
| 1470 |
+
"learning_rate": 2.7367758186397987e-05,
|
| 1471 |
+
"loss": 0.6093,
|
| 1472 |
+
"step": 209
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 0.53,
|
| 1476 |
+
"grad_norm": 10.88662338256836,
|
| 1477 |
+
"learning_rate": 2.73551637279597e-05,
|
| 1478 |
+
"loss": 0.7387,
|
| 1479 |
+
"step": 210
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 0.53,
|
| 1483 |
+
"grad_norm": 17.733755111694336,
|
| 1484 |
+
"learning_rate": 2.734256926952141e-05,
|
| 1485 |
+
"loss": 0.7371,
|
| 1486 |
+
"step": 211
|
| 1487 |
+
},
|
| 1488 |
+
{
|
| 1489 |
+
"epoch": 0.53,
|
| 1490 |
+
"grad_norm": 6.4716901779174805,
|
| 1491 |
+
"learning_rate": 2.7329974811083124e-05,
|
| 1492 |
+
"loss": 0.3157,
|
| 1493 |
+
"step": 212
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"epoch": 0.54,
|
| 1497 |
+
"grad_norm": 3.8103296756744385,
|
| 1498 |
+
"learning_rate": 2.7317380352644836e-05,
|
| 1499 |
+
"loss": 0.2807,
|
| 1500 |
+
"step": 213
|
| 1501 |
+
},
|
| 1502 |
+
{
|
| 1503 |
+
"epoch": 0.54,
|
| 1504 |
+
"grad_norm": 6.197389602661133,
|
| 1505 |
+
"learning_rate": 2.730478589420655e-05,
|
| 1506 |
+
"loss": 0.3373,
|
| 1507 |
+
"step": 214
|
| 1508 |
+
},
|
| 1509 |
+
{
|
| 1510 |
+
"epoch": 0.54,
|
| 1511 |
+
"grad_norm": 6.760047435760498,
|
| 1512 |
+
"learning_rate": 2.7292191435768265e-05,
|
| 1513 |
+
"loss": 0.3521,
|
| 1514 |
+
"step": 215
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 0.54,
|
| 1518 |
+
"grad_norm": 5.889042854309082,
|
| 1519 |
+
"learning_rate": 2.7279596977329977e-05,
|
| 1520 |
+
"loss": 0.276,
|
| 1521 |
+
"step": 216
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 0.55,
|
| 1525 |
+
"grad_norm": 7.752958297729492,
|
| 1526 |
+
"learning_rate": 2.7267002518891686e-05,
|
| 1527 |
+
"loss": 0.5561,
|
| 1528 |
+
"step": 217
|
| 1529 |
+
},
|
| 1530 |
+
{
|
| 1531 |
+
"epoch": 0.55,
|
| 1532 |
+
"grad_norm": 11.166560173034668,
|
| 1533 |
+
"learning_rate": 2.7254408060453402e-05,
|
| 1534 |
+
"loss": 0.3495,
|
| 1535 |
+
"step": 218
|
| 1536 |
+
},
|
| 1537 |
+
{
|
| 1538 |
+
"epoch": 0.55,
|
| 1539 |
+
"grad_norm": 7.210260391235352,
|
| 1540 |
+
"learning_rate": 2.7241813602015114e-05,
|
| 1541 |
+
"loss": 0.3042,
|
| 1542 |
+
"step": 219
|
| 1543 |
+
},
|
| 1544 |
+
{
|
| 1545 |
+
"epoch": 0.55,
|
| 1546 |
+
"grad_norm": 4.527043342590332,
|
| 1547 |
+
"learning_rate": 2.7229219143576827e-05,
|
| 1548 |
+
"loss": 0.2351,
|
| 1549 |
+
"step": 220
|
| 1550 |
+
},
|
| 1551 |
+
{
|
| 1552 |
+
"epoch": 0.56,
|
| 1553 |
+
"grad_norm": 18.023122787475586,
|
| 1554 |
+
"learning_rate": 2.721662468513854e-05,
|
| 1555 |
+
"loss": 0.6469,
|
| 1556 |
+
"step": 221
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 0.56,
|
| 1560 |
+
"grad_norm": 10.936578750610352,
|
| 1561 |
+
"learning_rate": 2.720403022670025e-05,
|
| 1562 |
+
"loss": 0.5859,
|
| 1563 |
+
"step": 222
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 0.56,
|
| 1567 |
+
"grad_norm": 11.246559143066406,
|
| 1568 |
+
"learning_rate": 2.7191435768261964e-05,
|
| 1569 |
+
"loss": 0.4637,
|
| 1570 |
+
"step": 223
|
| 1571 |
+
},
|
| 1572 |
+
{
|
| 1573 |
+
"epoch": 0.56,
|
| 1574 |
+
"grad_norm": 10.577720642089844,
|
| 1575 |
+
"learning_rate": 2.717884130982368e-05,
|
| 1576 |
+
"loss": 0.5009,
|
| 1577 |
+
"step": 224
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"epoch": 0.57,
|
| 1581 |
+
"grad_norm": 4.079160690307617,
|
| 1582 |
+
"learning_rate": 2.7166246851385392e-05,
|
| 1583 |
+
"loss": 0.1622,
|
| 1584 |
+
"step": 225
|
| 1585 |
+
},
|
| 1586 |
+
{
|
| 1587 |
+
"epoch": 0.57,
|
| 1588 |
+
"grad_norm": 7.402092933654785,
|
| 1589 |
+
"learning_rate": 2.71536523929471e-05,
|
| 1590 |
+
"loss": 0.2807,
|
| 1591 |
+
"step": 226
|
| 1592 |
+
},
|
| 1593 |
+
{
|
| 1594 |
+
"epoch": 0.57,
|
| 1595 |
+
"grad_norm": 11.498400688171387,
|
| 1596 |
+
"learning_rate": 2.7141057934508817e-05,
|
| 1597 |
+
"loss": 0.7531,
|
| 1598 |
+
"step": 227
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 0.57,
|
| 1602 |
+
"grad_norm": 4.898861885070801,
|
| 1603 |
+
"learning_rate": 2.712846347607053e-05,
|
| 1604 |
+
"loss": 0.2006,
|
| 1605 |
+
"step": 228
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 0.58,
|
| 1609 |
+
"grad_norm": 10.955698013305664,
|
| 1610 |
+
"learning_rate": 2.7115869017632242e-05,
|
| 1611 |
+
"loss": 0.4176,
|
| 1612 |
+
"step": 229
|
| 1613 |
+
},
|
| 1614 |
+
{
|
| 1615 |
+
"epoch": 0.58,
|
| 1616 |
+
"grad_norm": 9.794878005981445,
|
| 1617 |
+
"learning_rate": 2.7103274559193958e-05,
|
| 1618 |
+
"loss": 0.3908,
|
| 1619 |
+
"step": 230
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"epoch": 0.58,
|
| 1623 |
+
"grad_norm": 15.19322395324707,
|
| 1624 |
+
"learning_rate": 2.7090680100755667e-05,
|
| 1625 |
+
"loss": 0.6566,
|
| 1626 |
+
"step": 231
|
| 1627 |
+
},
|
| 1628 |
+
{
|
| 1629 |
+
"epoch": 0.58,
|
| 1630 |
+
"grad_norm": 12.016098022460938,
|
| 1631 |
+
"learning_rate": 2.707808564231738e-05,
|
| 1632 |
+
"loss": 0.4483,
|
| 1633 |
+
"step": 232
|
| 1634 |
+
},
|
| 1635 |
+
{
|
| 1636 |
+
"epoch": 0.59,
|
| 1637 |
+
"grad_norm": 4.229908466339111,
|
| 1638 |
+
"learning_rate": 2.7065491183879095e-05,
|
| 1639 |
+
"loss": 0.4335,
|
| 1640 |
+
"step": 233
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 0.59,
|
| 1644 |
+
"grad_norm": 10.778432846069336,
|
| 1645 |
+
"learning_rate": 2.7052896725440807e-05,
|
| 1646 |
+
"loss": 0.2455,
|
| 1647 |
+
"step": 234
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 0.59,
|
| 1651 |
+
"grad_norm": 7.074155807495117,
|
| 1652 |
+
"learning_rate": 2.7040302267002517e-05,
|
| 1653 |
+
"loss": 0.211,
|
| 1654 |
+
"step": 235
|
| 1655 |
+
},
|
| 1656 |
+
{
|
| 1657 |
+
"epoch": 0.59,
|
| 1658 |
+
"grad_norm": 6.830694675445557,
|
| 1659 |
+
"learning_rate": 2.7027707808564232e-05,
|
| 1660 |
+
"loss": 0.1891,
|
| 1661 |
+
"step": 236
|
| 1662 |
+
},
|
| 1663 |
+
{
|
| 1664 |
+
"epoch": 0.6,
|
| 1665 |
+
"grad_norm": 8.83045768737793,
|
| 1666 |
+
"learning_rate": 2.7015113350125945e-05,
|
| 1667 |
+
"loss": 0.4168,
|
| 1668 |
+
"step": 237
|
| 1669 |
+
},
|
| 1670 |
+
{
|
| 1671 |
+
"epoch": 0.6,
|
| 1672 |
+
"grad_norm": 8.931233406066895,
|
| 1673 |
+
"learning_rate": 2.7002518891687657e-05,
|
| 1674 |
+
"loss": 0.3596,
|
| 1675 |
+
"step": 238
|
| 1676 |
+
},
|
| 1677 |
+
{
|
| 1678 |
+
"epoch": 0.6,
|
| 1679 |
+
"grad_norm": 11.842453002929688,
|
| 1680 |
+
"learning_rate": 2.6989924433249373e-05,
|
| 1681 |
+
"loss": 0.5041,
|
| 1682 |
+
"step": 239
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 0.6,
|
| 1686 |
+
"grad_norm": 6.133688926696777,
|
| 1687 |
+
"learning_rate": 2.6977329974811082e-05,
|
| 1688 |
+
"loss": 0.5432,
|
| 1689 |
+
"step": 240
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 0.61,
|
| 1693 |
+
"grad_norm": 8.509514808654785,
|
| 1694 |
+
"learning_rate": 2.6964735516372794e-05,
|
| 1695 |
+
"loss": 0.3315,
|
| 1696 |
+
"step": 241
|
| 1697 |
+
},
|
| 1698 |
+
{
|
| 1699 |
+
"epoch": 0.61,
|
| 1700 |
+
"grad_norm": 8.522913932800293,
|
| 1701 |
+
"learning_rate": 2.695214105793451e-05,
|
| 1702 |
+
"loss": 0.3872,
|
| 1703 |
+
"step": 242
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 0.61,
|
| 1707 |
+
"grad_norm": 4.950671672821045,
|
| 1708 |
+
"learning_rate": 2.6939546599496223e-05,
|
| 1709 |
+
"loss": 0.2198,
|
| 1710 |
+
"step": 243
|
| 1711 |
+
},
|
| 1712 |
+
{
|
| 1713 |
+
"epoch": 0.61,
|
| 1714 |
+
"grad_norm": 12.004826545715332,
|
| 1715 |
+
"learning_rate": 2.6926952141057935e-05,
|
| 1716 |
+
"loss": 0.4331,
|
| 1717 |
+
"step": 244
|
| 1718 |
+
},
|
| 1719 |
+
{
|
| 1720 |
+
"epoch": 0.62,
|
| 1721 |
+
"grad_norm": 5.216823577880859,
|
| 1722 |
+
"learning_rate": 2.6914357682619648e-05,
|
| 1723 |
+
"loss": 0.12,
|
| 1724 |
+
"step": 245
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 0.62,
|
| 1728 |
+
"grad_norm": 4.667407512664795,
|
| 1729 |
+
"learning_rate": 2.690176322418136e-05,
|
| 1730 |
+
"loss": 0.3283,
|
| 1731 |
+
"step": 246
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 0.62,
|
| 1735 |
+
"grad_norm": 10.124615669250488,
|
| 1736 |
+
"learning_rate": 2.6889168765743072e-05,
|
| 1737 |
+
"loss": 0.188,
|
| 1738 |
+
"step": 247
|
| 1739 |
+
},
|
| 1740 |
+
{
|
| 1741 |
+
"epoch": 0.62,
|
| 1742 |
+
"grad_norm": 16.799171447753906,
|
| 1743 |
+
"learning_rate": 2.6876574307304788e-05,
|
| 1744 |
+
"loss": 0.5253,
|
| 1745 |
+
"step": 248
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 0.63,
|
| 1749 |
+
"grad_norm": 11.693323135375977,
|
| 1750 |
+
"learning_rate": 2.6863979848866497e-05,
|
| 1751 |
+
"loss": 0.3584,
|
| 1752 |
+
"step": 249
|
| 1753 |
+
},
|
| 1754 |
+
{
|
| 1755 |
+
"epoch": 0.63,
|
| 1756 |
+
"grad_norm": 9.2306489944458,
|
| 1757 |
+
"learning_rate": 2.685138539042821e-05,
|
| 1758 |
+
"loss": 0.2927,
|
| 1759 |
+
"step": 250
|
| 1760 |
+
},
|
| 1761 |
+
{
|
| 1762 |
+
"epoch": 0.63,
|
| 1763 |
+
"grad_norm": 6.715793132781982,
|
| 1764 |
+
"learning_rate": 2.6838790931989926e-05,
|
| 1765 |
+
"loss": 0.3034,
|
| 1766 |
+
"step": 251
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 0.63,
|
| 1770 |
+
"grad_norm": 10.9847993850708,
|
| 1771 |
+
"learning_rate": 2.6826196473551638e-05,
|
| 1772 |
+
"loss": 0.3885,
|
| 1773 |
+
"step": 252
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 0.64,
|
| 1777 |
+
"grad_norm": 9.826873779296875,
|
| 1778 |
+
"learning_rate": 2.681360201511335e-05,
|
| 1779 |
+
"loss": 0.4347,
|
| 1780 |
+
"step": 253
|
| 1781 |
+
},
|
| 1782 |
+
{
|
| 1783 |
+
"epoch": 0.64,
|
| 1784 |
+
"grad_norm": 16.903553009033203,
|
| 1785 |
+
"learning_rate": 2.6801007556675063e-05,
|
| 1786 |
+
"loss": 0.4324,
|
| 1787 |
+
"step": 254
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"epoch": 0.64,
|
| 1791 |
+
"grad_norm": 10.097508430480957,
|
| 1792 |
+
"learning_rate": 2.6788413098236775e-05,
|
| 1793 |
+
"loss": 0.4163,
|
| 1794 |
+
"step": 255
|
| 1795 |
+
},
|
| 1796 |
+
{
|
| 1797 |
+
"epoch": 0.64,
|
| 1798 |
+
"grad_norm": 7.872523307800293,
|
| 1799 |
+
"learning_rate": 2.677581863979849e-05,
|
| 1800 |
+
"loss": 0.5585,
|
| 1801 |
+
"step": 256
|
| 1802 |
+
},
|
| 1803 |
+
{
|
| 1804 |
+
"epoch": 0.65,
|
| 1805 |
+
"grad_norm": 5.794745922088623,
|
| 1806 |
+
"learning_rate": 2.6763224181360204e-05,
|
| 1807 |
+
"loss": 0.3675,
|
| 1808 |
+
"step": 257
|
| 1809 |
+
},
|
| 1810 |
+
{
|
| 1811 |
+
"epoch": 0.65,
|
| 1812 |
+
"grad_norm": 8.695060729980469,
|
| 1813 |
+
"learning_rate": 2.6750629722921916e-05,
|
| 1814 |
+
"loss": 0.2789,
|
| 1815 |
+
"step": 258
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 0.65,
|
| 1819 |
+
"grad_norm": 10.348055839538574,
|
| 1820 |
+
"learning_rate": 2.673803526448363e-05,
|
| 1821 |
+
"loss": 0.4214,
|
| 1822 |
+
"step": 259
|
| 1823 |
+
},
|
| 1824 |
+
{
|
| 1825 |
+
"epoch": 0.65,
|
| 1826 |
+
"grad_norm": 11.409734725952148,
|
| 1827 |
+
"learning_rate": 2.672544080604534e-05,
|
| 1828 |
+
"loss": 0.7848,
|
| 1829 |
+
"step": 260
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"epoch": 0.66,
|
| 1833 |
+
"grad_norm": 10.424241065979004,
|
| 1834 |
+
"learning_rate": 2.6712846347607053e-05,
|
| 1835 |
+
"loss": 0.5209,
|
| 1836 |
+
"step": 261
|
| 1837 |
+
},
|
| 1838 |
+
{
|
| 1839 |
+
"epoch": 0.66,
|
| 1840 |
+
"grad_norm": 5.8873114585876465,
|
| 1841 |
+
"learning_rate": 2.670025188916877e-05,
|
| 1842 |
+
"loss": 0.2265,
|
| 1843 |
+
"step": 262
|
| 1844 |
+
},
|
| 1845 |
+
{
|
| 1846 |
+
"epoch": 0.66,
|
| 1847 |
+
"grad_norm": 11.455771446228027,
|
| 1848 |
+
"learning_rate": 2.6687657430730478e-05,
|
| 1849 |
+
"loss": 0.7453,
|
| 1850 |
+
"step": 263
|
| 1851 |
+
},
|
| 1852 |
+
{
|
| 1853 |
+
"epoch": 0.66,
|
| 1854 |
+
"grad_norm": 12.967717170715332,
|
| 1855 |
+
"learning_rate": 2.667506297229219e-05,
|
| 1856 |
+
"loss": 0.3653,
|
| 1857 |
+
"step": 264
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 0.67,
|
| 1861 |
+
"grad_norm": 8.101251602172852,
|
| 1862 |
+
"learning_rate": 2.6662468513853906e-05,
|
| 1863 |
+
"loss": 0.3649,
|
| 1864 |
+
"step": 265
|
| 1865 |
+
},
|
| 1866 |
+
{
|
| 1867 |
+
"epoch": 0.67,
|
| 1868 |
+
"grad_norm": 10.239398956298828,
|
| 1869 |
+
"learning_rate": 2.664987405541562e-05,
|
| 1870 |
+
"loss": 0.4807,
|
| 1871 |
+
"step": 266
|
| 1872 |
+
},
|
| 1873 |
+
{
|
| 1874 |
+
"epoch": 0.67,
|
| 1875 |
+
"grad_norm": 7.652868270874023,
|
| 1876 |
+
"learning_rate": 2.663727959697733e-05,
|
| 1877 |
+
"loss": 0.4622,
|
| 1878 |
+
"step": 267
|
| 1879 |
+
},
|
| 1880 |
+
{
|
| 1881 |
+
"epoch": 0.68,
|
| 1882 |
+
"grad_norm": 7.320528030395508,
|
| 1883 |
+
"learning_rate": 2.6624685138539044e-05,
|
| 1884 |
+
"loss": 0.3544,
|
| 1885 |
+
"step": 268
|
| 1886 |
+
},
|
| 1887 |
+
{
|
| 1888 |
+
"epoch": 0.68,
|
| 1889 |
+
"grad_norm": 5.603414535522461,
|
| 1890 |
+
"learning_rate": 2.6612090680100756e-05,
|
| 1891 |
+
"loss": 0.225,
|
| 1892 |
+
"step": 269
|
| 1893 |
+
},
|
| 1894 |
+
{
|
| 1895 |
+
"epoch": 0.68,
|
| 1896 |
+
"grad_norm": 6.3062310218811035,
|
| 1897 |
+
"learning_rate": 2.659949622166247e-05,
|
| 1898 |
+
"loss": 0.29,
|
| 1899 |
+
"step": 270
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"epoch": 0.68,
|
| 1903 |
+
"grad_norm": 10.358262062072754,
|
| 1904 |
+
"learning_rate": 2.6586901763224184e-05,
|
| 1905 |
+
"loss": 0.5102,
|
| 1906 |
+
"step": 271
|
| 1907 |
+
},
|
| 1908 |
+
{
|
| 1909 |
+
"epoch": 0.69,
|
| 1910 |
+
"grad_norm": 5.093788146972656,
|
| 1911 |
+
"learning_rate": 2.6574307304785897e-05,
|
| 1912 |
+
"loss": 0.3955,
|
| 1913 |
+
"step": 272
|
| 1914 |
+
},
|
| 1915 |
+
{
|
| 1916 |
+
"epoch": 0.69,
|
| 1917 |
+
"grad_norm": 12.661267280578613,
|
| 1918 |
+
"learning_rate": 2.6561712846347606e-05,
|
| 1919 |
+
"loss": 0.4287,
|
| 1920 |
+
"step": 273
|
| 1921 |
+
},
|
| 1922 |
+
{
|
| 1923 |
+
"epoch": 0.69,
|
| 1924 |
+
"grad_norm": 6.933968544006348,
|
| 1925 |
+
"learning_rate": 2.654911838790932e-05,
|
| 1926 |
+
"loss": 0.3987,
|
| 1927 |
+
"step": 274
|
| 1928 |
+
},
|
| 1929 |
+
{
|
| 1930 |
+
"epoch": 0.69,
|
| 1931 |
+
"grad_norm": 10.704058647155762,
|
| 1932 |
+
"learning_rate": 2.6536523929471034e-05,
|
| 1933 |
+
"loss": 0.3831,
|
| 1934 |
+
"step": 275
|
| 1935 |
+
},
|
| 1936 |
+
{
|
| 1937 |
+
"epoch": 0.7,
|
| 1938 |
+
"grad_norm": 6.526610374450684,
|
| 1939 |
+
"learning_rate": 2.6523929471032746e-05,
|
| 1940 |
+
"loss": 0.3872,
|
| 1941 |
+
"step": 276
|
| 1942 |
+
},
|
| 1943 |
+
{
|
| 1944 |
+
"epoch": 0.7,
|
| 1945 |
+
"grad_norm": 5.706236839294434,
|
| 1946 |
+
"learning_rate": 2.651133501259446e-05,
|
| 1947 |
+
"loss": 0.4442,
|
| 1948 |
+
"step": 277
|
| 1949 |
+
},
|
| 1950 |
+
{
|
| 1951 |
+
"epoch": 0.7,
|
| 1952 |
+
"grad_norm": 7.010911464691162,
|
| 1953 |
+
"learning_rate": 2.649874055415617e-05,
|
| 1954 |
+
"loss": 0.403,
|
| 1955 |
+
"step": 278
|
| 1956 |
+
},
|
| 1957 |
+
{
|
| 1958 |
+
"epoch": 0.7,
|
| 1959 |
+
"grad_norm": 5.128185749053955,
|
| 1960 |
+
"learning_rate": 2.6486146095717884e-05,
|
| 1961 |
+
"loss": 0.2681,
|
| 1962 |
+
"step": 279
|
| 1963 |
+
},
|
| 1964 |
+
{
|
| 1965 |
+
"epoch": 0.71,
|
| 1966 |
+
"grad_norm": 9.200146675109863,
|
| 1967 |
+
"learning_rate": 2.64735516372796e-05,
|
| 1968 |
+
"loss": 0.4966,
|
| 1969 |
+
"step": 280
|
| 1970 |
+
},
|
| 1971 |
+
{
|
| 1972 |
+
"epoch": 0.71,
|
| 1973 |
+
"grad_norm": 10.710809707641602,
|
| 1974 |
+
"learning_rate": 2.6460957178841312e-05,
|
| 1975 |
+
"loss": 0.3411,
|
| 1976 |
+
"step": 281
|
| 1977 |
+
},
|
| 1978 |
+
{
|
| 1979 |
+
"epoch": 0.71,
|
| 1980 |
+
"grad_norm": 6.845553874969482,
|
| 1981 |
+
"learning_rate": 2.644836272040302e-05,
|
| 1982 |
+
"loss": 0.4416,
|
| 1983 |
+
"step": 282
|
| 1984 |
+
},
|
| 1985 |
+
{
|
| 1986 |
+
"epoch": 0.71,
|
| 1987 |
+
"grad_norm": 6.162066459655762,
|
| 1988 |
+
"learning_rate": 2.6435768261964737e-05,
|
| 1989 |
+
"loss": 0.5407,
|
| 1990 |
+
"step": 283
|
| 1991 |
+
},
|
| 1992 |
+
{
|
| 1993 |
+
"epoch": 0.72,
|
| 1994 |
+
"grad_norm": 11.713993072509766,
|
| 1995 |
+
"learning_rate": 2.642317380352645e-05,
|
| 1996 |
+
"loss": 0.9063,
|
| 1997 |
+
"step": 284
|
| 1998 |
+
},
|
| 1999 |
+
{
|
| 2000 |
+
"epoch": 0.72,
|
| 2001 |
+
"grad_norm": 7.366119384765625,
|
| 2002 |
+
"learning_rate": 2.641057934508816e-05,
|
| 2003 |
+
"loss": 0.4344,
|
| 2004 |
+
"step": 285
|
| 2005 |
+
},
|
| 2006 |
+
{
|
| 2007 |
+
"epoch": 0.72,
|
| 2008 |
+
"grad_norm": 7.175751209259033,
|
| 2009 |
+
"learning_rate": 2.6397984886649878e-05,
|
| 2010 |
+
"loss": 0.4297,
|
| 2011 |
+
"step": 286
|
| 2012 |
+
},
|
| 2013 |
+
{
|
| 2014 |
+
"epoch": 0.72,
|
| 2015 |
+
"grad_norm": 14.455536842346191,
|
| 2016 |
+
"learning_rate": 2.6385390428211587e-05,
|
| 2017 |
+
"loss": 0.5623,
|
| 2018 |
+
"step": 287
|
| 2019 |
+
},
|
| 2020 |
+
{
|
| 2021 |
+
"epoch": 0.73,
|
| 2022 |
+
"grad_norm": 7.098339080810547,
|
| 2023 |
+
"learning_rate": 2.63727959697733e-05,
|
| 2024 |
+
"loss": 0.2626,
|
| 2025 |
+
"step": 288
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 0.73,
|
| 2029 |
+
"grad_norm": 9.281996726989746,
|
| 2030 |
+
"learning_rate": 2.6360201511335015e-05,
|
| 2031 |
+
"loss": 0.4656,
|
| 2032 |
+
"step": 289
|
| 2033 |
+
},
|
| 2034 |
+
{
|
| 2035 |
+
"epoch": 0.73,
|
| 2036 |
+
"grad_norm": 8.613694190979004,
|
| 2037 |
+
"learning_rate": 2.6347607052896727e-05,
|
| 2038 |
+
"loss": 0.5498,
|
| 2039 |
+
"step": 290
|
| 2040 |
+
},
|
| 2041 |
+
{
|
| 2042 |
+
"epoch": 0.73,
|
| 2043 |
+
"grad_norm": 4.918963432312012,
|
| 2044 |
+
"learning_rate": 2.6335012594458436e-05,
|
| 2045 |
+
"loss": 0.4148,
|
| 2046 |
+
"step": 291
|
| 2047 |
+
},
|
| 2048 |
+
{
|
| 2049 |
+
"epoch": 0.74,
|
| 2050 |
+
"grad_norm": 12.958881378173828,
|
| 2051 |
+
"learning_rate": 2.6322418136020152e-05,
|
| 2052 |
+
"loss": 0.8399,
|
| 2053 |
+
"step": 292
|
| 2054 |
+
},
|
| 2055 |
+
{
|
| 2056 |
+
"epoch": 0.74,
|
| 2057 |
+
"grad_norm": 6.894360065460205,
|
| 2058 |
+
"learning_rate": 2.6309823677581865e-05,
|
| 2059 |
+
"loss": 0.3625,
|
| 2060 |
+
"step": 293
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"epoch": 0.74,
|
| 2064 |
+
"grad_norm": 11.97496223449707,
|
| 2065 |
+
"learning_rate": 2.6297229219143577e-05,
|
| 2066 |
+
"loss": 0.6333,
|
| 2067 |
+
"step": 294
|
| 2068 |
+
},
|
| 2069 |
+
{
|
| 2070 |
+
"epoch": 0.74,
|
| 2071 |
+
"grad_norm": 6.575906276702881,
|
| 2072 |
+
"learning_rate": 2.6284634760705293e-05,
|
| 2073 |
+
"loss": 0.3271,
|
| 2074 |
+
"step": 295
|
| 2075 |
+
},
|
| 2076 |
+
{
|
| 2077 |
+
"epoch": 0.75,
|
| 2078 |
+
"grad_norm": 8.077478408813477,
|
| 2079 |
+
"learning_rate": 2.6272040302267002e-05,
|
| 2080 |
+
"loss": 0.4545,
|
| 2081 |
+
"step": 296
|
| 2082 |
+
},
|
| 2083 |
+
{
|
| 2084 |
+
"epoch": 0.75,
|
| 2085 |
+
"grad_norm": 7.899360179901123,
|
| 2086 |
+
"learning_rate": 2.6259445843828714e-05,
|
| 2087 |
+
"loss": 0.4815,
|
| 2088 |
+
"step": 297
|
| 2089 |
+
},
|
| 2090 |
+
{
|
| 2091 |
+
"epoch": 0.75,
|
| 2092 |
+
"grad_norm": 9.990239143371582,
|
| 2093 |
+
"learning_rate": 2.624685138539043e-05,
|
| 2094 |
+
"loss": 0.631,
|
| 2095 |
+
"step": 298
|
| 2096 |
+
},
|
| 2097 |
+
{
|
| 2098 |
+
"epoch": 0.75,
|
| 2099 |
+
"grad_norm": 7.54863166809082,
|
| 2100 |
+
"learning_rate": 2.6234256926952142e-05,
|
| 2101 |
+
"loss": 0.5138,
|
| 2102 |
+
"step": 299
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 0.76,
|
| 2106 |
+
"grad_norm": 7.893070220947266,
|
| 2107 |
+
"learning_rate": 2.6221662468513855e-05,
|
| 2108 |
+
"loss": 0.2928,
|
| 2109 |
+
"step": 300
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"epoch": 0.76,
|
| 2113 |
+
"grad_norm": 6.223887920379639,
|
| 2114 |
+
"learning_rate": 2.6209068010075567e-05,
|
| 2115 |
+
"loss": 0.3628,
|
| 2116 |
+
"step": 301
|
| 2117 |
+
},
|
| 2118 |
+
{
|
| 2119 |
+
"epoch": 0.76,
|
| 2120 |
+
"grad_norm": 9.773299217224121,
|
| 2121 |
+
"learning_rate": 2.619647355163728e-05,
|
| 2122 |
+
"loss": 0.5439,
|
| 2123 |
+
"step": 302
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"epoch": 0.76,
|
| 2127 |
+
"grad_norm": 4.369682788848877,
|
| 2128 |
+
"learning_rate": 2.6183879093198992e-05,
|
| 2129 |
+
"loss": 0.2298,
|
| 2130 |
+
"step": 303
|
| 2131 |
+
},
|
| 2132 |
+
{
|
| 2133 |
+
"epoch": 0.77,
|
| 2134 |
+
"grad_norm": 4.92218542098999,
|
| 2135 |
+
"learning_rate": 2.6171284634760708e-05,
|
| 2136 |
+
"loss": 0.2903,
|
| 2137 |
+
"step": 304
|
| 2138 |
+
},
|
| 2139 |
+
{
|
| 2140 |
+
"epoch": 0.77,
|
| 2141 |
+
"grad_norm": 5.855688571929932,
|
| 2142 |
+
"learning_rate": 2.6158690176322417e-05,
|
| 2143 |
+
"loss": 0.3738,
|
| 2144 |
+
"step": 305
|
| 2145 |
+
},
|
| 2146 |
+
{
|
| 2147 |
+
"epoch": 0.77,
|
| 2148 |
+
"grad_norm": 4.964177131652832,
|
| 2149 |
+
"learning_rate": 2.614609571788413e-05,
|
| 2150 |
+
"loss": 0.373,
|
| 2151 |
+
"step": 306
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 0.77,
|
| 2155 |
+
"grad_norm": 8.245409965515137,
|
| 2156 |
+
"learning_rate": 2.6133501259445845e-05,
|
| 2157 |
+
"loss": 0.4829,
|
| 2158 |
+
"step": 307
|
| 2159 |
+
},
|
| 2160 |
+
{
|
| 2161 |
+
"epoch": 0.78,
|
| 2162 |
+
"grad_norm": 8.923148155212402,
|
| 2163 |
+
"learning_rate": 2.6120906801007558e-05,
|
| 2164 |
+
"loss": 0.3148,
|
| 2165 |
+
"step": 308
|
| 2166 |
+
},
|
| 2167 |
+
{
|
| 2168 |
+
"epoch": 0.78,
|
| 2169 |
+
"grad_norm": 9.844786643981934,
|
| 2170 |
+
"learning_rate": 2.610831234256927e-05,
|
| 2171 |
+
"loss": 0.4606,
|
| 2172 |
+
"step": 309
|
| 2173 |
+
},
|
| 2174 |
+
{
|
| 2175 |
+
"epoch": 0.78,
|
| 2176 |
+
"grad_norm": 5.273078441619873,
|
| 2177 |
+
"learning_rate": 2.6095717884130983e-05,
|
| 2178 |
+
"loss": 0.2387,
|
| 2179 |
+
"step": 310
|
| 2180 |
+
},
|
| 2181 |
+
{
|
| 2182 |
+
"epoch": 0.78,
|
| 2183 |
+
"grad_norm": 8.784069061279297,
|
| 2184 |
+
"learning_rate": 2.6083123425692695e-05,
|
| 2185 |
+
"loss": 0.3051,
|
| 2186 |
+
"step": 311
|
| 2187 |
+
},
|
| 2188 |
+
{
|
| 2189 |
+
"epoch": 0.79,
|
| 2190 |
+
"grad_norm": 7.740212917327881,
|
| 2191 |
+
"learning_rate": 2.6070528967254407e-05,
|
| 2192 |
+
"loss": 0.2269,
|
| 2193 |
+
"step": 312
|
| 2194 |
+
},
|
| 2195 |
+
{
|
| 2196 |
+
"epoch": 0.79,
|
| 2197 |
+
"grad_norm": 12.335386276245117,
|
| 2198 |
+
"learning_rate": 2.6057934508816123e-05,
|
| 2199 |
+
"loss": 0.4262,
|
| 2200 |
+
"step": 313
|
| 2201 |
+
},
|
| 2202 |
+
{
|
| 2203 |
+
"epoch": 0.79,
|
| 2204 |
+
"grad_norm": 4.678969383239746,
|
| 2205 |
+
"learning_rate": 2.6045340050377836e-05,
|
| 2206 |
+
"loss": 0.1791,
|
| 2207 |
+
"step": 314
|
| 2208 |
+
},
|
| 2209 |
+
{
|
| 2210 |
+
"epoch": 0.79,
|
| 2211 |
+
"grad_norm": 9.690211296081543,
|
| 2212 |
+
"learning_rate": 2.6032745591939545e-05,
|
| 2213 |
+
"loss": 0.2351,
|
| 2214 |
+
"step": 315
|
| 2215 |
+
},
|
| 2216 |
+
{
|
| 2217 |
+
"epoch": 0.8,
|
| 2218 |
+
"grad_norm": 12.03903865814209,
|
| 2219 |
+
"learning_rate": 2.602015113350126e-05,
|
| 2220 |
+
"loss": 0.8051,
|
| 2221 |
+
"step": 316
|
| 2222 |
+
},
|
| 2223 |
+
{
|
| 2224 |
+
"epoch": 0.8,
|
| 2225 |
+
"grad_norm": 11.595259666442871,
|
| 2226 |
+
"learning_rate": 2.6007556675062973e-05,
|
| 2227 |
+
"loss": 0.3681,
|
| 2228 |
+
"step": 317
|
| 2229 |
+
},
|
| 2230 |
+
{
|
| 2231 |
+
"epoch": 0.8,
|
| 2232 |
+
"grad_norm": 3.8068230152130127,
|
| 2233 |
+
"learning_rate": 2.5994962216624685e-05,
|
| 2234 |
+
"loss": 0.3157,
|
| 2235 |
+
"step": 318
|
| 2236 |
+
},
|
| 2237 |
+
{
|
| 2238 |
+
"epoch": 0.8,
|
| 2239 |
+
"grad_norm": 4.258709907531738,
|
| 2240 |
+
"learning_rate": 2.5982367758186398e-05,
|
| 2241 |
+
"loss": 0.1873,
|
| 2242 |
+
"step": 319
|
| 2243 |
+
},
|
| 2244 |
+
{
|
| 2245 |
+
"epoch": 0.81,
|
| 2246 |
+
"grad_norm": 7.911970615386963,
|
| 2247 |
+
"learning_rate": 2.596977329974811e-05,
|
| 2248 |
+
"loss": 0.2601,
|
| 2249 |
+
"step": 320
|
| 2250 |
+
},
|
| 2251 |
+
{
|
| 2252 |
+
"epoch": 0.81,
|
| 2253 |
+
"grad_norm": 13.647623062133789,
|
| 2254 |
+
"learning_rate": 2.5957178841309823e-05,
|
| 2255 |
+
"loss": 0.6636,
|
| 2256 |
+
"step": 321
|
| 2257 |
+
},
|
| 2258 |
+
{
|
| 2259 |
+
"epoch": 0.81,
|
| 2260 |
+
"grad_norm": 14.344672203063965,
|
| 2261 |
+
"learning_rate": 2.594458438287154e-05,
|
| 2262 |
+
"loss": 0.5215,
|
| 2263 |
+
"step": 322
|
| 2264 |
+
},
|
| 2265 |
+
{
|
| 2266 |
+
"epoch": 0.81,
|
| 2267 |
+
"grad_norm": 6.633630752563477,
|
| 2268 |
+
"learning_rate": 2.593198992443325e-05,
|
| 2269 |
+
"loss": 0.4419,
|
| 2270 |
+
"step": 323
|
| 2271 |
+
},
|
| 2272 |
+
{
|
| 2273 |
+
"epoch": 0.82,
|
| 2274 |
+
"grad_norm": 11.353543281555176,
|
| 2275 |
+
"learning_rate": 2.5919395465994963e-05,
|
| 2276 |
+
"loss": 0.4572,
|
| 2277 |
+
"step": 324
|
| 2278 |
+
},
|
| 2279 |
+
{
|
| 2280 |
+
"epoch": 0.82,
|
| 2281 |
+
"grad_norm": 8.904594421386719,
|
| 2282 |
+
"learning_rate": 2.5906801007556676e-05,
|
| 2283 |
+
"loss": 0.3291,
|
| 2284 |
+
"step": 325
|
| 2285 |
+
},
|
| 2286 |
+
{
|
| 2287 |
+
"epoch": 0.82,
|
| 2288 |
+
"grad_norm": 10.205378532409668,
|
| 2289 |
+
"learning_rate": 2.5894206549118388e-05,
|
| 2290 |
+
"loss": 0.3371,
|
| 2291 |
+
"step": 326
|
| 2292 |
+
},
|
| 2293 |
+
{
|
| 2294 |
+
"epoch": 0.82,
|
| 2295 |
+
"grad_norm": 8.3673095703125,
|
| 2296 |
+
"learning_rate": 2.5881612090680104e-05,
|
| 2297 |
+
"loss": 0.322,
|
| 2298 |
+
"step": 327
|
| 2299 |
+
},
|
| 2300 |
+
{
|
| 2301 |
+
"epoch": 0.83,
|
| 2302 |
+
"grad_norm": 7.420979022979736,
|
| 2303 |
+
"learning_rate": 2.5869017632241813e-05,
|
| 2304 |
+
"loss": 0.3825,
|
| 2305 |
+
"step": 328
|
| 2306 |
+
},
|
| 2307 |
+
{
|
| 2308 |
+
"epoch": 0.83,
|
| 2309 |
+
"grad_norm": 12.808219909667969,
|
| 2310 |
+
"learning_rate": 2.5856423173803526e-05,
|
| 2311 |
+
"loss": 0.5258,
|
| 2312 |
+
"step": 329
|
| 2313 |
+
},
|
| 2314 |
+
{
|
| 2315 |
+
"epoch": 0.83,
|
| 2316 |
+
"grad_norm": 9.836726188659668,
|
| 2317 |
+
"learning_rate": 2.584382871536524e-05,
|
| 2318 |
+
"loss": 0.3549,
|
| 2319 |
+
"step": 330
|
| 2320 |
+
},
|
| 2321 |
+
{
|
| 2322 |
+
"epoch": 0.83,
|
| 2323 |
+
"grad_norm": 6.072993755340576,
|
| 2324 |
+
"learning_rate": 2.5831234256926954e-05,
|
| 2325 |
+
"loss": 0.3907,
|
| 2326 |
+
"step": 331
|
| 2327 |
+
},
|
| 2328 |
+
{
|
| 2329 |
+
"epoch": 0.84,
|
| 2330 |
+
"grad_norm": 20.400632858276367,
|
| 2331 |
+
"learning_rate": 2.5818639798488666e-05,
|
| 2332 |
+
"loss": 0.8538,
|
| 2333 |
+
"step": 332
|
| 2334 |
+
},
|
| 2335 |
+
{
|
| 2336 |
+
"epoch": 0.84,
|
| 2337 |
+
"grad_norm": 9.215600967407227,
|
| 2338 |
+
"learning_rate": 2.580604534005038e-05,
|
| 2339 |
+
"loss": 0.3813,
|
| 2340 |
+
"step": 333
|
| 2341 |
+
},
|
| 2342 |
+
{
|
| 2343 |
+
"epoch": 0.84,
|
| 2344 |
+
"grad_norm": 8.703184127807617,
|
| 2345 |
+
"learning_rate": 2.579345088161209e-05,
|
| 2346 |
+
"loss": 0.6626,
|
| 2347 |
+
"step": 334
|
| 2348 |
+
},
|
| 2349 |
+
{
|
| 2350 |
+
"epoch": 0.84,
|
| 2351 |
+
"grad_norm": 10.990998268127441,
|
| 2352 |
+
"learning_rate": 2.5780856423173803e-05,
|
| 2353 |
+
"loss": 0.4793,
|
| 2354 |
+
"step": 335
|
| 2355 |
+
},
|
| 2356 |
+
{
|
| 2357 |
+
"epoch": 0.85,
|
| 2358 |
+
"grad_norm": 11.248719215393066,
|
| 2359 |
+
"learning_rate": 2.576826196473552e-05,
|
| 2360 |
+
"loss": 0.3488,
|
| 2361 |
+
"step": 336
|
| 2362 |
+
},
|
| 2363 |
+
{
|
| 2364 |
+
"epoch": 0.85,
|
| 2365 |
+
"grad_norm": 6.082599639892578,
|
| 2366 |
+
"learning_rate": 2.5755667506297232e-05,
|
| 2367 |
+
"loss": 0.4904,
|
| 2368 |
+
"step": 337
|
| 2369 |
+
},
|
| 2370 |
+
{
|
| 2371 |
+
"epoch": 0.85,
|
| 2372 |
+
"grad_norm": 6.6478495597839355,
|
| 2373 |
+
"learning_rate": 2.574307304785894e-05,
|
| 2374 |
+
"loss": 0.535,
|
| 2375 |
+
"step": 338
|
| 2376 |
+
},
|
| 2377 |
+
{
|
| 2378 |
+
"epoch": 0.85,
|
| 2379 |
+
"grad_norm": 8.529887199401855,
|
| 2380 |
+
"learning_rate": 2.5730478589420657e-05,
|
| 2381 |
+
"loss": 0.6424,
|
| 2382 |
+
"step": 339
|
| 2383 |
+
},
|
| 2384 |
+
{
|
| 2385 |
+
"epoch": 0.86,
|
| 2386 |
+
"grad_norm": 7.008981704711914,
|
| 2387 |
+
"learning_rate": 2.571788413098237e-05,
|
| 2388 |
+
"loss": 0.2907,
|
| 2389 |
+
"step": 340
|
| 2390 |
+
},
|
| 2391 |
+
{
|
| 2392 |
+
"epoch": 0.86,
|
| 2393 |
+
"grad_norm": 4.586373805999756,
|
| 2394 |
+
"learning_rate": 2.570528967254408e-05,
|
| 2395 |
+
"loss": 0.2298,
|
| 2396 |
+
"step": 341
|
| 2397 |
+
},
|
| 2398 |
+
{
|
| 2399 |
+
"epoch": 0.86,
|
| 2400 |
+
"grad_norm": 6.62999963760376,
|
| 2401 |
+
"learning_rate": 2.5692695214105794e-05,
|
| 2402 |
+
"loss": 0.3508,
|
| 2403 |
+
"step": 342
|
| 2404 |
+
},
|
| 2405 |
+
{
|
| 2406 |
+
"epoch": 0.86,
|
| 2407 |
+
"grad_norm": 9.748456954956055,
|
| 2408 |
+
"learning_rate": 2.5680100755667506e-05,
|
| 2409 |
+
"loss": 0.4157,
|
| 2410 |
+
"step": 343
|
| 2411 |
+
},
|
| 2412 |
+
{
|
| 2413 |
+
"epoch": 0.87,
|
| 2414 |
+
"grad_norm": 6.978835582733154,
|
| 2415 |
+
"learning_rate": 2.566750629722922e-05,
|
| 2416 |
+
"loss": 0.3145,
|
| 2417 |
+
"step": 344
|
| 2418 |
+
},
|
| 2419 |
+
{
|
| 2420 |
+
"epoch": 0.87,
|
| 2421 |
+
"grad_norm": 5.523429870605469,
|
| 2422 |
+
"learning_rate": 2.5654911838790935e-05,
|
| 2423 |
+
"loss": 0.3471,
|
| 2424 |
+
"step": 345
|
| 2425 |
+
},
|
| 2426 |
+
{
|
| 2427 |
+
"epoch": 0.87,
|
| 2428 |
+
"grad_norm": 3.892453193664551,
|
| 2429 |
+
"learning_rate": 2.5642317380352647e-05,
|
| 2430 |
+
"loss": 0.2324,
|
| 2431 |
+
"step": 346
|
| 2432 |
+
},
|
| 2433 |
+
{
|
| 2434 |
+
"epoch": 0.87,
|
| 2435 |
+
"grad_norm": 4.895789623260498,
|
| 2436 |
+
"learning_rate": 2.5629722921914356e-05,
|
| 2437 |
+
"loss": 0.3082,
|
| 2438 |
+
"step": 347
|
| 2439 |
+
},
|
| 2440 |
+
{
|
| 2441 |
+
"epoch": 0.88,
|
| 2442 |
+
"grad_norm": 5.085398197174072,
|
| 2443 |
+
"learning_rate": 2.5617128463476072e-05,
|
| 2444 |
+
"loss": 0.4067,
|
| 2445 |
+
"step": 348
|
| 2446 |
+
},
|
| 2447 |
+
{
|
| 2448 |
+
"epoch": 0.88,
|
| 2449 |
+
"grad_norm": 8.951998710632324,
|
| 2450 |
+
"learning_rate": 2.5604534005037784e-05,
|
| 2451 |
+
"loss": 0.4093,
|
| 2452 |
+
"step": 349
|
| 2453 |
+
},
|
| 2454 |
+
{
|
| 2455 |
+
"epoch": 0.88,
|
| 2456 |
+
"grad_norm": 6.7257537841796875,
|
| 2457 |
+
"learning_rate": 2.5591939546599497e-05,
|
| 2458 |
+
"loss": 0.4623,
|
| 2459 |
+
"step": 350
|
| 2460 |
+
},
|
| 2461 |
+
{
|
| 2462 |
+
"epoch": 0.88,
|
| 2463 |
+
"grad_norm": 9.630943298339844,
|
| 2464 |
+
"learning_rate": 2.5579345088161213e-05,
|
| 2465 |
+
"loss": 0.3686,
|
| 2466 |
+
"step": 351
|
| 2467 |
+
},
|
| 2468 |
+
{
|
| 2469 |
+
"epoch": 0.89,
|
| 2470 |
+
"grad_norm": 7.153581142425537,
|
| 2471 |
+
"learning_rate": 2.556675062972292e-05,
|
| 2472 |
+
"loss": 0.468,
|
| 2473 |
+
"step": 352
|
| 2474 |
+
},
|
| 2475 |
+
{
|
| 2476 |
+
"epoch": 0.89,
|
| 2477 |
+
"grad_norm": 13.286303520202637,
|
| 2478 |
+
"learning_rate": 2.5554156171284634e-05,
|
| 2479 |
+
"loss": 0.6772,
|
| 2480 |
+
"step": 353
|
| 2481 |
+
},
|
| 2482 |
+
{
|
| 2483 |
+
"epoch": 0.89,
|
| 2484 |
+
"grad_norm": 7.783265113830566,
|
| 2485 |
+
"learning_rate": 2.554156171284635e-05,
|
| 2486 |
+
"loss": 0.2495,
|
| 2487 |
+
"step": 354
|
| 2488 |
+
},
|
| 2489 |
+
{
|
| 2490 |
+
"epoch": 0.89,
|
| 2491 |
+
"grad_norm": 9.463069915771484,
|
| 2492 |
+
"learning_rate": 2.5528967254408062e-05,
|
| 2493 |
+
"loss": 0.393,
|
| 2494 |
+
"step": 355
|
| 2495 |
+
},
|
| 2496 |
+
{
|
| 2497 |
+
"epoch": 0.9,
|
| 2498 |
+
"grad_norm": 7.801197052001953,
|
| 2499 |
+
"learning_rate": 2.551637279596977e-05,
|
| 2500 |
+
"loss": 0.2972,
|
| 2501 |
+
"step": 356
|
| 2502 |
+
},
|
| 2503 |
+
{
|
| 2504 |
+
"epoch": 0.9,
|
| 2505 |
+
"grad_norm": 8.433084487915039,
|
| 2506 |
+
"learning_rate": 2.5503778337531487e-05,
|
| 2507 |
+
"loss": 0.4681,
|
| 2508 |
+
"step": 357
|
| 2509 |
+
},
|
| 2510 |
+
{
|
| 2511 |
+
"epoch": 0.9,
|
| 2512 |
+
"grad_norm": 5.388092041015625,
|
| 2513 |
+
"learning_rate": 2.54911838790932e-05,
|
| 2514 |
+
"loss": 0.513,
|
| 2515 |
+
"step": 358
|
| 2516 |
+
},
|
| 2517 |
+
{
|
| 2518 |
+
"epoch": 0.9,
|
| 2519 |
+
"grad_norm": 5.502450942993164,
|
| 2520 |
+
"learning_rate": 2.5478589420654912e-05,
|
| 2521 |
+
"loss": 0.1494,
|
| 2522 |
+
"step": 359
|
| 2523 |
+
},
|
| 2524 |
+
{
|
| 2525 |
+
"epoch": 0.91,
|
| 2526 |
+
"grad_norm": 4.962391376495361,
|
| 2527 |
+
"learning_rate": 2.5465994962216628e-05,
|
| 2528 |
+
"loss": 0.1157,
|
| 2529 |
+
"step": 360
|
| 2530 |
+
},
|
| 2531 |
+
{
|
| 2532 |
+
"epoch": 0.91,
|
| 2533 |
+
"grad_norm": 6.759737014770508,
|
| 2534 |
+
"learning_rate": 2.5453400503778337e-05,
|
| 2535 |
+
"loss": 0.1665,
|
| 2536 |
+
"step": 361
|
| 2537 |
+
},
|
| 2538 |
+
{
|
| 2539 |
+
"epoch": 0.91,
|
| 2540 |
+
"grad_norm": 7.5388407707214355,
|
| 2541 |
+
"learning_rate": 2.544080604534005e-05,
|
| 2542 |
+
"loss": 0.3471,
|
| 2543 |
+
"step": 362
|
| 2544 |
+
},
|
| 2545 |
+
{
|
| 2546 |
+
"epoch": 0.91,
|
| 2547 |
+
"grad_norm": 8.41981315612793,
|
| 2548 |
+
"learning_rate": 2.5428211586901765e-05,
|
| 2549 |
+
"loss": 0.4093,
|
| 2550 |
+
"step": 363
|
| 2551 |
+
},
|
| 2552 |
+
{
|
| 2553 |
+
"epoch": 0.92,
|
| 2554 |
+
"grad_norm": 9.60474681854248,
|
| 2555 |
+
"learning_rate": 2.5415617128463477e-05,
|
| 2556 |
+
"loss": 0.576,
|
| 2557 |
+
"step": 364
|
| 2558 |
+
},
|
| 2559 |
+
{
|
| 2560 |
+
"epoch": 0.92,
|
| 2561 |
+
"grad_norm": 6.068057537078857,
|
| 2562 |
+
"learning_rate": 2.540302267002519e-05,
|
| 2563 |
+
"loss": 0.203,
|
| 2564 |
+
"step": 365
|
| 2565 |
+
},
|
| 2566 |
+
{
|
| 2567 |
+
"epoch": 0.92,
|
| 2568 |
+
"grad_norm": 4.424688339233398,
|
| 2569 |
+
"learning_rate": 2.5390428211586902e-05,
|
| 2570 |
+
"loss": 0.3454,
|
| 2571 |
+
"step": 366
|
| 2572 |
+
},
|
| 2573 |
+
{
|
| 2574 |
+
"epoch": 0.92,
|
| 2575 |
+
"grad_norm": 11.978801727294922,
|
| 2576 |
+
"learning_rate": 2.5377833753148615e-05,
|
| 2577 |
+
"loss": 0.5684,
|
| 2578 |
+
"step": 367
|
| 2579 |
+
},
|
| 2580 |
+
{
|
| 2581 |
+
"epoch": 0.93,
|
| 2582 |
+
"grad_norm": 6.773107051849365,
|
| 2583 |
+
"learning_rate": 2.5365239294710327e-05,
|
| 2584 |
+
"loss": 0.176,
|
| 2585 |
+
"step": 368
|
| 2586 |
+
},
|
| 2587 |
+
{
|
| 2588 |
+
"epoch": 0.93,
|
| 2589 |
+
"grad_norm": 19.249605178833008,
|
| 2590 |
+
"learning_rate": 2.5352644836272043e-05,
|
| 2591 |
+
"loss": 0.586,
|
| 2592 |
+
"step": 369
|
| 2593 |
+
},
|
| 2594 |
+
{
|
| 2595 |
+
"epoch": 0.93,
|
| 2596 |
+
"grad_norm": 15.166516304016113,
|
| 2597 |
+
"learning_rate": 2.5340050377833752e-05,
|
| 2598 |
+
"loss": 1.1046,
|
| 2599 |
+
"step": 370
|
| 2600 |
+
},
|
| 2601 |
+
{
|
| 2602 |
+
"epoch": 0.93,
|
| 2603 |
+
"grad_norm": 7.654723167419434,
|
| 2604 |
+
"learning_rate": 2.5327455919395464e-05,
|
| 2605 |
+
"loss": 0.5403,
|
| 2606 |
+
"step": 371
|
| 2607 |
+
},
|
| 2608 |
+
{
|
| 2609 |
+
"epoch": 0.94,
|
| 2610 |
+
"grad_norm": 18.302127838134766,
|
| 2611 |
+
"learning_rate": 2.531486146095718e-05,
|
| 2612 |
+
"loss": 0.9826,
|
| 2613 |
+
"step": 372
|
| 2614 |
+
},
|
| 2615 |
+
{
|
| 2616 |
+
"epoch": 0.94,
|
| 2617 |
+
"grad_norm": 11.380560874938965,
|
| 2618 |
+
"learning_rate": 2.5302267002518893e-05,
|
| 2619 |
+
"loss": 0.8694,
|
| 2620 |
+
"step": 373
|
| 2621 |
+
},
|
| 2622 |
+
{
|
| 2623 |
+
"epoch": 0.94,
|
| 2624 |
+
"grad_norm": 12.14782428741455,
|
| 2625 |
+
"learning_rate": 2.5289672544080605e-05,
|
| 2626 |
+
"loss": 0.6133,
|
| 2627 |
+
"step": 374
|
| 2628 |
+
},
|
| 2629 |
+
{
|
| 2630 |
+
"epoch": 0.94,
|
| 2631 |
+
"grad_norm": 5.418036460876465,
|
| 2632 |
+
"learning_rate": 2.5277078085642318e-05,
|
| 2633 |
+
"loss": 0.3352,
|
| 2634 |
+
"step": 375
|
| 2635 |
+
},
|
| 2636 |
+
{
|
| 2637 |
+
"epoch": 0.95,
|
| 2638 |
+
"grad_norm": 14.057409286499023,
|
| 2639 |
+
"learning_rate": 2.526448362720403e-05,
|
| 2640 |
+
"loss": 0.5738,
|
| 2641 |
+
"step": 376
|
| 2642 |
+
},
|
| 2643 |
+
{
|
| 2644 |
+
"epoch": 0.95,
|
| 2645 |
+
"grad_norm": 6.224894046783447,
|
| 2646 |
+
"learning_rate": 2.5251889168765742e-05,
|
| 2647 |
+
"loss": 0.2167,
|
| 2648 |
+
"step": 377
|
| 2649 |
+
},
|
| 2650 |
+
{
|
| 2651 |
+
"epoch": 0.95,
|
| 2652 |
+
"grad_norm": 5.565800189971924,
|
| 2653 |
+
"learning_rate": 2.5239294710327458e-05,
|
| 2654 |
+
"loss": 0.2301,
|
| 2655 |
+
"step": 378
|
| 2656 |
+
},
|
| 2657 |
+
{
|
| 2658 |
+
"epoch": 0.95,
|
| 2659 |
+
"grad_norm": 5.572880744934082,
|
| 2660 |
+
"learning_rate": 2.522670025188917e-05,
|
| 2661 |
+
"loss": 0.2257,
|
| 2662 |
+
"step": 379
|
| 2663 |
+
},
|
| 2664 |
+
{
|
| 2665 |
+
"epoch": 0.96,
|
| 2666 |
+
"grad_norm": 9.148365020751953,
|
| 2667 |
+
"learning_rate": 2.521410579345088e-05,
|
| 2668 |
+
"loss": 0.7268,
|
| 2669 |
+
"step": 380
|
| 2670 |
+
},
|
| 2671 |
+
{
|
| 2672 |
+
"epoch": 0.96,
|
| 2673 |
+
"grad_norm": 5.175827980041504,
|
| 2674 |
+
"learning_rate": 2.5201511335012596e-05,
|
| 2675 |
+
"loss": 0.2781,
|
| 2676 |
+
"step": 381
|
| 2677 |
+
},
|
| 2678 |
+
{
|
| 2679 |
+
"epoch": 0.96,
|
| 2680 |
+
"grad_norm": 10.978497505187988,
|
| 2681 |
+
"learning_rate": 2.5188916876574308e-05,
|
| 2682 |
+
"loss": 0.9962,
|
| 2683 |
+
"step": 382
|
| 2684 |
+
},
|
| 2685 |
+
{
|
| 2686 |
+
"epoch": 0.96,
|
| 2687 |
+
"grad_norm": 6.896300792694092,
|
| 2688 |
+
"learning_rate": 2.517632241813602e-05,
|
| 2689 |
+
"loss": 0.7353,
|
| 2690 |
+
"step": 383
|
| 2691 |
+
},
|
| 2692 |
+
{
|
| 2693 |
+
"epoch": 0.97,
|
| 2694 |
+
"grad_norm": 6.292906284332275,
|
| 2695 |
+
"learning_rate": 2.5163727959697733e-05,
|
| 2696 |
+
"loss": 0.3912,
|
| 2697 |
+
"step": 384
|
| 2698 |
+
},
|
| 2699 |
+
{
|
| 2700 |
+
"epoch": 0.97,
|
| 2701 |
+
"grad_norm": 5.655113697052002,
|
| 2702 |
+
"learning_rate": 2.5151133501259445e-05,
|
| 2703 |
+
"loss": 0.2604,
|
| 2704 |
+
"step": 385
|
| 2705 |
+
},
|
| 2706 |
+
{
|
| 2707 |
+
"epoch": 0.97,
|
| 2708 |
+
"grad_norm": 16.326297760009766,
|
| 2709 |
+
"learning_rate": 2.5138539042821158e-05,
|
| 2710 |
+
"loss": 1.0246,
|
| 2711 |
+
"step": 386
|
| 2712 |
+
},
|
| 2713 |
+
{
|
| 2714 |
+
"epoch": 0.97,
|
| 2715 |
+
"grad_norm": 8.862824440002441,
|
| 2716 |
+
"learning_rate": 2.5125944584382874e-05,
|
| 2717 |
+
"loss": 0.287,
|
| 2718 |
+
"step": 387
|
| 2719 |
+
},
|
| 2720 |
+
{
|
| 2721 |
+
"epoch": 0.98,
|
| 2722 |
+
"grad_norm": 12.7918119430542,
|
| 2723 |
+
"learning_rate": 2.5113350125944586e-05,
|
| 2724 |
+
"loss": 0.5495,
|
| 2725 |
+
"step": 388
|
| 2726 |
+
},
|
| 2727 |
+
{
|
| 2728 |
+
"epoch": 0.98,
|
| 2729 |
+
"grad_norm": 8.119661331176758,
|
| 2730 |
+
"learning_rate": 2.5100755667506295e-05,
|
| 2731 |
+
"loss": 0.6277,
|
| 2732 |
+
"step": 389
|
| 2733 |
+
},
|
| 2734 |
+
{
|
| 2735 |
+
"epoch": 0.98,
|
| 2736 |
+
"grad_norm": 7.078393459320068,
|
| 2737 |
+
"learning_rate": 2.508816120906801e-05,
|
| 2738 |
+
"loss": 0.4774,
|
| 2739 |
+
"step": 390
|
| 2740 |
+
},
|
| 2741 |
+
{
|
| 2742 |
+
"epoch": 0.98,
|
| 2743 |
+
"grad_norm": 8.347442626953125,
|
| 2744 |
+
"learning_rate": 2.5075566750629723e-05,
|
| 2745 |
+
"loss": 0.6561,
|
| 2746 |
+
"step": 391
|
| 2747 |
+
},
|
| 2748 |
+
{
|
| 2749 |
+
"epoch": 0.99,
|
| 2750 |
+
"grad_norm": 5.844377040863037,
|
| 2751 |
+
"learning_rate": 2.5062972292191436e-05,
|
| 2752 |
+
"loss": 0.3074,
|
| 2753 |
+
"step": 392
|
| 2754 |
+
},
|
| 2755 |
+
{
|
| 2756 |
+
"epoch": 0.99,
|
| 2757 |
+
"grad_norm": 10.436945915222168,
|
| 2758 |
+
"learning_rate": 2.505037783375315e-05,
|
| 2759 |
+
"loss": 0.5096,
|
| 2760 |
+
"step": 393
|
| 2761 |
+
},
|
| 2762 |
+
{
|
| 2763 |
+
"epoch": 0.99,
|
| 2764 |
+
"grad_norm": 7.574183464050293,
|
| 2765 |
+
"learning_rate": 2.503778337531486e-05,
|
| 2766 |
+
"loss": 0.2668,
|
| 2767 |
+
"step": 394
|
| 2768 |
+
},
|
| 2769 |
+
{
|
| 2770 |
+
"epoch": 0.99,
|
| 2771 |
+
"grad_norm": 6.011397361755371,
|
| 2772 |
+
"learning_rate": 2.5025188916876573e-05,
|
| 2773 |
+
"loss": 0.4125,
|
| 2774 |
+
"step": 395
|
| 2775 |
+
},
|
| 2776 |
+
{
|
| 2777 |
+
"epoch": 1.0,
|
| 2778 |
+
"grad_norm": 5.520245552062988,
|
| 2779 |
+
"learning_rate": 2.501259445843829e-05,
|
| 2780 |
+
"loss": 0.5079,
|
| 2781 |
+
"step": 396
|
| 2782 |
+
},
|
| 2783 |
+
{
|
| 2784 |
+
"epoch": 1.0,
|
| 2785 |
+
"grad_norm": 4.188296794891357,
|
| 2786 |
+
"learning_rate": 2.5e-05,
|
| 2787 |
+
"loss": 0.0709,
|
| 2788 |
+
"step": 397
|
| 2789 |
+
},
|
| 2790 |
+
{
|
| 2791 |
+
"epoch": 1.0,
|
| 2792 |
+
"eval_accuracy": 0.7905362776025237,
|
| 2793 |
+
"eval_f1": 0.7903191872567364,
|
| 2794 |
+
"eval_loss": 0.4741531312465668,
|
| 2795 |
+
"eval_precision": 0.7992890189359668,
|
| 2796 |
+
"eval_recall": 0.8039643127134329,
|
| 2797 |
+
"eval_runtime": 1566.2518,
|
| 2798 |
+
"eval_samples_per_second": 1.012,
|
| 2799 |
+
"eval_steps_per_second": 0.064,
|
| 2800 |
+
"step": 397
|
| 2801 |
+
}
|
| 2802 |
+
],
|
| 2803 |
+
"logging_steps": 1,
|
| 2804 |
+
"max_steps": 2382,
|
| 2805 |
+
"num_input_tokens_seen": 0,
|
| 2806 |
+
"num_train_epochs": 6,
|
| 2807 |
+
"save_steps": 500,
|
| 2808 |
+
"total_flos": 833667378908160.0,
|
| 2809 |
+
"train_batch_size": 16,
|
| 2810 |
+
"trial_name": null,
|
| 2811 |
+
"trial_params": null
|
| 2812 |
+
}
|
tmp-checkpoint-397/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:57e67fc3ab87e1ca24b63448da544d25db23ca61b336ee5e110ad3e1b44c668b
|
| 3 |
+
size 4984
|
tmp-checkpoint-397/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,57 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "[PAD]",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "[UNK]",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "[CLS]",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "[SEP]",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"4": {
|
| 36 |
+
"content": "[MASK]",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"clean_up_tokenization_spaces": true,
|
| 45 |
+
"cls_token": "[CLS]",
|
| 46 |
+
"do_basic_tokenize": true,
|
| 47 |
+
"do_lower_case": true,
|
| 48 |
+
"mask_token": "[MASK]",
|
| 49 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 50 |
+
"never_split": null,
|
| 51 |
+
"pad_token": "[PAD]",
|
| 52 |
+
"sep_token": "[SEP]",
|
| 53 |
+
"strip_accents": null,
|
| 54 |
+
"tokenize_chinese_chars": true,
|
| 55 |
+
"tokenizer_class": "BertTokenizer",
|
| 56 |
+
"unk_token": "[UNK]"
|
| 57 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:57e67fc3ab87e1ca24b63448da544d25db23ca61b336ee5e110ad3e1b44c668b
|
| 3 |
+
size 4984
|
vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|