Upload folder using huggingface_hub
Browse files- config.json +135 -0
- model.safetensors +3 -0
- optimizer.pt +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- special_tokens_map.json +7 -0
- tokenizer.json +0 -0
- tokenizer_config.json +59 -0
- trainer_state.json +1498 -0
- training_args.bin +3 -0
- vocab.txt +0 -0
config.json
ADDED
|
@@ -0,0 +1,135 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"BertForTokenClassification"
|
| 4 |
+
],
|
| 5 |
+
"attention_probs_dropout_prob": 0.1,
|
| 6 |
+
"classifier_dropout": null,
|
| 7 |
+
"gradient_checkpointing": false,
|
| 8 |
+
"hidden_act": "gelu",
|
| 9 |
+
"hidden_dropout_prob": 0.1,
|
| 10 |
+
"hidden_size": 768,
|
| 11 |
+
"id2label": {
|
| 12 |
+
"0": "O",
|
| 13 |
+
"1": "B-IP_ADDRESS",
|
| 14 |
+
"2": "I-IP_ADDRESS",
|
| 15 |
+
"3": "B-SERVER_NAME",
|
| 16 |
+
"4": "I-SERVER_NAME",
|
| 17 |
+
"5": "B-DOMAIN",
|
| 18 |
+
"6": "I-DOMAIN",
|
| 19 |
+
"7": "B-PERSON",
|
| 20 |
+
"8": "I-PERSON",
|
| 21 |
+
"9": "B-PHONE_NUMBER",
|
| 22 |
+
"10": "I-PHONE_NUMBER",
|
| 23 |
+
"11": "B-EMAIL",
|
| 24 |
+
"12": "I-EMAIL",
|
| 25 |
+
"13": "B-IBAN_ACCOUNT",
|
| 26 |
+
"14": "I-IBAN_ACCOUNT",
|
| 27 |
+
"15": "B-BANK_ACCOUNT",
|
| 28 |
+
"16": "I-BANK_ACCOUNT",
|
| 29 |
+
"17": "B-BANK_NAME",
|
| 30 |
+
"18": "I-BANK_NAME",
|
| 31 |
+
"19": "B-AMOUNT",
|
| 32 |
+
"20": "I-AMOUNT",
|
| 33 |
+
"21": "B-CREDIT_CARD",
|
| 34 |
+
"22": "I-CREDIT_CARD",
|
| 35 |
+
"23": "B-GEO_UNIT_NUMBER",
|
| 36 |
+
"24": "I-GEO_UNIT_NUMBER",
|
| 37 |
+
"25": "B-BANK_BRANCH_NUMBER",
|
| 38 |
+
"26": "I-BANK_BRANCH_NUMBER",
|
| 39 |
+
"27": "B-STREET_ADDRESS",
|
| 40 |
+
"28": "I-STREET_ADDRESS",
|
| 41 |
+
"29": "B-CITY",
|
| 42 |
+
"30": "I-CITY",
|
| 43 |
+
"31": "B-PROVINCE_STATE",
|
| 44 |
+
"32": "I-PROVINCE_STATE",
|
| 45 |
+
"33": "B-ISRAEL_ID",
|
| 46 |
+
"34": "I-ISRAEL_ID",
|
| 47 |
+
"35": "B-PASSPORT",
|
| 48 |
+
"36": "I-PASSPORT",
|
| 49 |
+
"37": "B-COMPANY_NUMBER",
|
| 50 |
+
"38": "I-COMPANY_NUMBER",
|
| 51 |
+
"39": "B-COMPANY_NAME",
|
| 52 |
+
"40": "I-COMPANY_NAME",
|
| 53 |
+
"41": "B-JWT",
|
| 54 |
+
"42": "I-JWT",
|
| 55 |
+
"43": "B-USERNAME",
|
| 56 |
+
"44": "I-USERNAME",
|
| 57 |
+
"45": "B-PASSWORD",
|
| 58 |
+
"46": "I-PASSWORD",
|
| 59 |
+
"47": "B-SECRET",
|
| 60 |
+
"48": "I-SECRET",
|
| 61 |
+
"49": "B-API_KEY",
|
| 62 |
+
"50": "I-API_KEY",
|
| 63 |
+
"51": "B-CERTIFICATE",
|
| 64 |
+
"52": "I-CERTIFICATE"
|
| 65 |
+
},
|
| 66 |
+
"initializer_range": 0.02,
|
| 67 |
+
"intermediate_size": 3072,
|
| 68 |
+
"label2id": {
|
| 69 |
+
"B-AMOUNT": 19,
|
| 70 |
+
"B-API_KEY": 49,
|
| 71 |
+
"B-BANK_ACCOUNT": 15,
|
| 72 |
+
"B-BANK_BRANCH_NUMBER": 25,
|
| 73 |
+
"B-BANK_NAME": 17,
|
| 74 |
+
"B-CERTIFICATE": 51,
|
| 75 |
+
"B-CITY": 29,
|
| 76 |
+
"B-COMPANY_NAME": 39,
|
| 77 |
+
"B-COMPANY_NUMBER": 37,
|
| 78 |
+
"B-CREDIT_CARD": 21,
|
| 79 |
+
"B-DOMAIN": 5,
|
| 80 |
+
"B-EMAIL": 11,
|
| 81 |
+
"B-GEO_UNIT_NUMBER": 23,
|
| 82 |
+
"B-IBAN_ACCOUNT": 13,
|
| 83 |
+
"B-IP_ADDRESS": 1,
|
| 84 |
+
"B-ISRAEL_ID": 33,
|
| 85 |
+
"B-JWT": 41,
|
| 86 |
+
"B-PASSPORT": 35,
|
| 87 |
+
"B-PASSWORD": 45,
|
| 88 |
+
"B-PERSON": 7,
|
| 89 |
+
"B-PHONE_NUMBER": 9,
|
| 90 |
+
"B-PROVINCE_STATE": 31,
|
| 91 |
+
"B-SECRET": 47,
|
| 92 |
+
"B-SERVER_NAME": 3,
|
| 93 |
+
"B-STREET_ADDRESS": 27,
|
| 94 |
+
"B-USERNAME": 43,
|
| 95 |
+
"I-AMOUNT": 20,
|
| 96 |
+
"I-API_KEY": 50,
|
| 97 |
+
"I-BANK_ACCOUNT": 16,
|
| 98 |
+
"I-BANK_BRANCH_NUMBER": 26,
|
| 99 |
+
"I-BANK_NAME": 18,
|
| 100 |
+
"I-CERTIFICATE": 52,
|
| 101 |
+
"I-CITY": 30,
|
| 102 |
+
"I-COMPANY_NAME": 40,
|
| 103 |
+
"I-COMPANY_NUMBER": 38,
|
| 104 |
+
"I-CREDIT_CARD": 22,
|
| 105 |
+
"I-DOMAIN": 6,
|
| 106 |
+
"I-EMAIL": 12,
|
| 107 |
+
"I-GEO_UNIT_NUMBER": 24,
|
| 108 |
+
"I-IBAN_ACCOUNT": 14,
|
| 109 |
+
"I-IP_ADDRESS": 2,
|
| 110 |
+
"I-ISRAEL_ID": 34,
|
| 111 |
+
"I-JWT": 42,
|
| 112 |
+
"I-PASSPORT": 36,
|
| 113 |
+
"I-PASSWORD": 46,
|
| 114 |
+
"I-PERSON": 8,
|
| 115 |
+
"I-PHONE_NUMBER": 10,
|
| 116 |
+
"I-PROVINCE_STATE": 32,
|
| 117 |
+
"I-SECRET": 48,
|
| 118 |
+
"I-SERVER_NAME": 4,
|
| 119 |
+
"I-STREET_ADDRESS": 28,
|
| 120 |
+
"I-USERNAME": 44,
|
| 121 |
+
"O": 0
|
| 122 |
+
},
|
| 123 |
+
"layer_norm_eps": 1e-12,
|
| 124 |
+
"max_position_embeddings": 512,
|
| 125 |
+
"model_type": "bert",
|
| 126 |
+
"num_attention_heads": 12,
|
| 127 |
+
"num_hidden_layers": 12,
|
| 128 |
+
"pad_token_id": 0,
|
| 129 |
+
"position_embedding_type": "absolute",
|
| 130 |
+
"torch_dtype": "float32",
|
| 131 |
+
"transformers_version": "4.54.1",
|
| 132 |
+
"type_vocab_size": 1,
|
| 133 |
+
"use_cache": true,
|
| 134 |
+
"vocab_size": 52000
|
| 135 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:12d079162172f13d0fe7f6f2249c2c23bd1475023642fba552eb9bdef7ef9b27
|
| 3 |
+
size 501730324
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:60fef75e6f7fe835f7b68b636887c38793bd33989a1025bf4ce08876c514f1c6
|
| 3 |
+
size 1003580346
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:31468f08e7a9f95f89f9d0b4ad33a8298bf66b13e3b505fc5c1c46943b2cd3f1
|
| 3 |
+
size 14244
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c202dacfc4d7001609152c6fc324bee0b3cf258928bb66e63d80186a00991fb0
|
| 3 |
+
size 1064
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"cls_token": "[CLS]",
|
| 3 |
+
"mask_token": "[MASK]",
|
| 4 |
+
"pad_token": "[PAD]",
|
| 5 |
+
"sep_token": "[SEP]",
|
| 6 |
+
"unk_token": "[UNK]"
|
| 7 |
+
}
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,59 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "[MASK]",
|
| 50 |
+
"max_len": 512,
|
| 51 |
+
"model_max_length": 512,
|
| 52 |
+
"never_split": null,
|
| 53 |
+
"pad_token": "[PAD]",
|
| 54 |
+
"sep_token": "[SEP]",
|
| 55 |
+
"strip_accents": null,
|
| 56 |
+
"tokenize_chinese_chars": true,
|
| 57 |
+
"tokenizer_class": "BertTokenizer",
|
| 58 |
+
"unk_token": "[UNK]"
|
| 59 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1498 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 20475,
|
| 3 |
+
"best_metric": 1.0,
|
| 4 |
+
"best_model_checkpoint": "./ner_model_logs/checkpoint-20475",
|
| 5 |
+
"epoch": 3.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 20475,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.014652014652014652,
|
| 14 |
+
"grad_norm": 2.3198535442352295,
|
| 15 |
+
"learning_rate": 1.9903296703296703e-05,
|
| 16 |
+
"loss": 0.721,
|
| 17 |
+
"step": 100
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.029304029304029304,
|
| 21 |
+
"grad_norm": 0.227557972073555,
|
| 22 |
+
"learning_rate": 1.9805616605616607e-05,
|
| 23 |
+
"loss": 0.0878,
|
| 24 |
+
"step": 200
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.04395604395604396,
|
| 28 |
+
"grad_norm": 3.8474526405334473,
|
| 29 |
+
"learning_rate": 1.970793650793651e-05,
|
| 30 |
+
"loss": 0.0133,
|
| 31 |
+
"step": 300
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.05860805860805861,
|
| 35 |
+
"grad_norm": 0.06616735458374023,
|
| 36 |
+
"learning_rate": 1.961025641025641e-05,
|
| 37 |
+
"loss": 0.0051,
|
| 38 |
+
"step": 400
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.07326007326007326,
|
| 42 |
+
"grad_norm": 0.02466263622045517,
|
| 43 |
+
"learning_rate": 1.9512576312576314e-05,
|
| 44 |
+
"loss": 0.0046,
|
| 45 |
+
"step": 500
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.08791208791208792,
|
| 49 |
+
"grad_norm": 0.3532762825489044,
|
| 50 |
+
"learning_rate": 1.9414896214896218e-05,
|
| 51 |
+
"loss": 0.0033,
|
| 52 |
+
"step": 600
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.10256410256410256,
|
| 56 |
+
"grad_norm": 1.2159192562103271,
|
| 57 |
+
"learning_rate": 1.931721611721612e-05,
|
| 58 |
+
"loss": 0.0023,
|
| 59 |
+
"step": 700
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 0.11721611721611722,
|
| 63 |
+
"grad_norm": 0.07953804731369019,
|
| 64 |
+
"learning_rate": 1.9219536019536022e-05,
|
| 65 |
+
"loss": 0.0036,
|
| 66 |
+
"step": 800
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 0.13186813186813187,
|
| 70 |
+
"grad_norm": 0.022895004600286484,
|
| 71 |
+
"learning_rate": 1.9121855921855922e-05,
|
| 72 |
+
"loss": 0.0018,
|
| 73 |
+
"step": 900
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 0.14652014652014653,
|
| 77 |
+
"grad_norm": 0.0714658722281456,
|
| 78 |
+
"learning_rate": 1.9024175824175826e-05,
|
| 79 |
+
"loss": 0.0013,
|
| 80 |
+
"step": 1000
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.16117216117216118,
|
| 84 |
+
"grad_norm": 0.07990960776805878,
|
| 85 |
+
"learning_rate": 1.892649572649573e-05,
|
| 86 |
+
"loss": 0.0011,
|
| 87 |
+
"step": 1100
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.17582417582417584,
|
| 91 |
+
"grad_norm": 0.005112498998641968,
|
| 92 |
+
"learning_rate": 1.882881562881563e-05,
|
| 93 |
+
"loss": 0.0013,
|
| 94 |
+
"step": 1200
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"epoch": 0.19047619047619047,
|
| 98 |
+
"grad_norm": 0.004775651730597019,
|
| 99 |
+
"learning_rate": 1.8731135531135533e-05,
|
| 100 |
+
"loss": 0.0006,
|
| 101 |
+
"step": 1300
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 0.20512820512820512,
|
| 105 |
+
"grad_norm": 0.003744883695617318,
|
| 106 |
+
"learning_rate": 1.8633455433455434e-05,
|
| 107 |
+
"loss": 0.0004,
|
| 108 |
+
"step": 1400
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 0.21978021978021978,
|
| 112 |
+
"grad_norm": 0.16701260209083557,
|
| 113 |
+
"learning_rate": 1.8535775335775337e-05,
|
| 114 |
+
"loss": 0.0007,
|
| 115 |
+
"step": 1500
|
| 116 |
+
},
|
| 117 |
+
{
|
| 118 |
+
"epoch": 0.23443223443223443,
|
| 119 |
+
"grad_norm": 0.012116856873035431,
|
| 120 |
+
"learning_rate": 1.843809523809524e-05,
|
| 121 |
+
"loss": 0.0019,
|
| 122 |
+
"step": 1600
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.2490842490842491,
|
| 126 |
+
"grad_norm": 0.03328242152929306,
|
| 127 |
+
"learning_rate": 1.834041514041514e-05,
|
| 128 |
+
"loss": 0.0006,
|
| 129 |
+
"step": 1700
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.26373626373626374,
|
| 133 |
+
"grad_norm": 0.48128026723861694,
|
| 134 |
+
"learning_rate": 1.8242735042735045e-05,
|
| 135 |
+
"loss": 0.0022,
|
| 136 |
+
"step": 1800
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"epoch": 0.2783882783882784,
|
| 140 |
+
"grad_norm": 0.008382058702409267,
|
| 141 |
+
"learning_rate": 1.814505494505495e-05,
|
| 142 |
+
"loss": 0.001,
|
| 143 |
+
"step": 1900
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 0.29304029304029305,
|
| 147 |
+
"grad_norm": 0.0023408152628690004,
|
| 148 |
+
"learning_rate": 1.804737484737485e-05,
|
| 149 |
+
"loss": 0.0009,
|
| 150 |
+
"step": 2000
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"epoch": 0.3076923076923077,
|
| 154 |
+
"grad_norm": 0.003858278738334775,
|
| 155 |
+
"learning_rate": 1.794969474969475e-05,
|
| 156 |
+
"loss": 0.0075,
|
| 157 |
+
"step": 2100
|
| 158 |
+
},
|
| 159 |
+
{
|
| 160 |
+
"epoch": 0.32234432234432236,
|
| 161 |
+
"grad_norm": 0.005588823929429054,
|
| 162 |
+
"learning_rate": 1.7852014652014653e-05,
|
| 163 |
+
"loss": 0.0025,
|
| 164 |
+
"step": 2200
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.336996336996337,
|
| 168 |
+
"grad_norm": 0.002319753635674715,
|
| 169 |
+
"learning_rate": 1.7754334554334556e-05,
|
| 170 |
+
"loss": 0.0004,
|
| 171 |
+
"step": 2300
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.3516483516483517,
|
| 175 |
+
"grad_norm": 0.002465345896780491,
|
| 176 |
+
"learning_rate": 1.765665445665446e-05,
|
| 177 |
+
"loss": 0.001,
|
| 178 |
+
"step": 2400
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"epoch": 0.3663003663003663,
|
| 182 |
+
"grad_norm": 0.014657393097877502,
|
| 183 |
+
"learning_rate": 1.755897435897436e-05,
|
| 184 |
+
"loss": 0.0009,
|
| 185 |
+
"step": 2500
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 0.38095238095238093,
|
| 189 |
+
"grad_norm": 0.0048686908558011055,
|
| 190 |
+
"learning_rate": 1.7461294261294264e-05,
|
| 191 |
+
"loss": 0.0013,
|
| 192 |
+
"step": 2600
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 0.3956043956043956,
|
| 196 |
+
"grad_norm": 0.0015726798446848989,
|
| 197 |
+
"learning_rate": 1.7363614163614164e-05,
|
| 198 |
+
"loss": 0.0004,
|
| 199 |
+
"step": 2700
|
| 200 |
+
},
|
| 201 |
+
{
|
| 202 |
+
"epoch": 0.41025641025641024,
|
| 203 |
+
"grad_norm": 0.002173542743548751,
|
| 204 |
+
"learning_rate": 1.7265934065934068e-05,
|
| 205 |
+
"loss": 0.0002,
|
| 206 |
+
"step": 2800
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.4249084249084249,
|
| 210 |
+
"grad_norm": 0.0014299682807177305,
|
| 211 |
+
"learning_rate": 1.7168253968253968e-05,
|
| 212 |
+
"loss": 0.0001,
|
| 213 |
+
"step": 2900
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 0.43956043956043955,
|
| 217 |
+
"grad_norm": 0.001351997023448348,
|
| 218 |
+
"learning_rate": 1.707057387057387e-05,
|
| 219 |
+
"loss": 0.0003,
|
| 220 |
+
"step": 3000
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"epoch": 0.4542124542124542,
|
| 224 |
+
"grad_norm": 0.04424109309911728,
|
| 225 |
+
"learning_rate": 1.6972893772893775e-05,
|
| 226 |
+
"loss": 0.0004,
|
| 227 |
+
"step": 3100
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 0.46886446886446886,
|
| 231 |
+
"grad_norm": 0.0023619807325303555,
|
| 232 |
+
"learning_rate": 1.687521367521368e-05,
|
| 233 |
+
"loss": 0.0018,
|
| 234 |
+
"step": 3200
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 0.4835164835164835,
|
| 238 |
+
"grad_norm": 0.0014780177734792233,
|
| 239 |
+
"learning_rate": 1.677753357753358e-05,
|
| 240 |
+
"loss": 0.0003,
|
| 241 |
+
"step": 3300
|
| 242 |
+
},
|
| 243 |
+
{
|
| 244 |
+
"epoch": 0.4981684981684982,
|
| 245 |
+
"grad_norm": 0.0007553008617833257,
|
| 246 |
+
"learning_rate": 1.667985347985348e-05,
|
| 247 |
+
"loss": 0.0001,
|
| 248 |
+
"step": 3400
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.5128205128205128,
|
| 252 |
+
"grad_norm": 0.0011919466778635979,
|
| 253 |
+
"learning_rate": 1.6582173382173383e-05,
|
| 254 |
+
"loss": 0.0005,
|
| 255 |
+
"step": 3500
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 0.5274725274725275,
|
| 259 |
+
"grad_norm": 0.0008271373226307333,
|
| 260 |
+
"learning_rate": 1.6484493284493287e-05,
|
| 261 |
+
"loss": 0.0005,
|
| 262 |
+
"step": 3600
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"epoch": 0.5421245421245421,
|
| 266 |
+
"grad_norm": 0.0020716034341603518,
|
| 267 |
+
"learning_rate": 1.6386813186813187e-05,
|
| 268 |
+
"loss": 0.0001,
|
| 269 |
+
"step": 3700
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 0.5567765567765568,
|
| 273 |
+
"grad_norm": 0.0011554955272004008,
|
| 274 |
+
"learning_rate": 1.628913308913309e-05,
|
| 275 |
+
"loss": 0.0007,
|
| 276 |
+
"step": 3800
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 0.5714285714285714,
|
| 280 |
+
"grad_norm": 0.0009056870476342738,
|
| 281 |
+
"learning_rate": 1.6191452991452994e-05,
|
| 282 |
+
"loss": 0.0001,
|
| 283 |
+
"step": 3900
|
| 284 |
+
},
|
| 285 |
+
{
|
| 286 |
+
"epoch": 0.5860805860805861,
|
| 287 |
+
"grad_norm": 0.00144649064168334,
|
| 288 |
+
"learning_rate": 1.6093772893772894e-05,
|
| 289 |
+
"loss": 0.0002,
|
| 290 |
+
"step": 4000
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.6007326007326007,
|
| 294 |
+
"grad_norm": 0.0012368283933028579,
|
| 295 |
+
"learning_rate": 1.5996092796092798e-05,
|
| 296 |
+
"loss": 0.0012,
|
| 297 |
+
"step": 4100
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.6153846153846154,
|
| 301 |
+
"grad_norm": 0.032919418066740036,
|
| 302 |
+
"learning_rate": 1.58984126984127e-05,
|
| 303 |
+
"loss": 0.0013,
|
| 304 |
+
"step": 4200
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 0.63003663003663,
|
| 308 |
+
"grad_norm": 0.0019616656936705112,
|
| 309 |
+
"learning_rate": 1.5800732600732602e-05,
|
| 310 |
+
"loss": 0.0013,
|
| 311 |
+
"step": 4300
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 0.6446886446886447,
|
| 315 |
+
"grad_norm": 0.006232458166778088,
|
| 316 |
+
"learning_rate": 1.5703052503052506e-05,
|
| 317 |
+
"loss": 0.0007,
|
| 318 |
+
"step": 4400
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"epoch": 0.6593406593406593,
|
| 322 |
+
"grad_norm": 0.00093706080224365,
|
| 323 |
+
"learning_rate": 1.5605372405372406e-05,
|
| 324 |
+
"loss": 0.0005,
|
| 325 |
+
"step": 4500
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 0.673992673992674,
|
| 329 |
+
"grad_norm": 0.001097803469747305,
|
| 330 |
+
"learning_rate": 1.550769230769231e-05,
|
| 331 |
+
"loss": 0.0001,
|
| 332 |
+
"step": 4600
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.6886446886446886,
|
| 336 |
+
"grad_norm": 0.0011469683377072215,
|
| 337 |
+
"learning_rate": 1.541001221001221e-05,
|
| 338 |
+
"loss": 0.0001,
|
| 339 |
+
"step": 4700
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.7032967032967034,
|
| 343 |
+
"grad_norm": 0.001127622788771987,
|
| 344 |
+
"learning_rate": 1.5312332112332114e-05,
|
| 345 |
+
"loss": 0.0006,
|
| 346 |
+
"step": 4800
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 0.717948717948718,
|
| 350 |
+
"grad_norm": 0.0017739522736519575,
|
| 351 |
+
"learning_rate": 1.5214652014652015e-05,
|
| 352 |
+
"loss": 0.0006,
|
| 353 |
+
"step": 4900
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 0.7326007326007326,
|
| 357 |
+
"grad_norm": 0.00775282084941864,
|
| 358 |
+
"learning_rate": 1.5116971916971919e-05,
|
| 359 |
+
"loss": 0.0001,
|
| 360 |
+
"step": 5000
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 0.7472527472527473,
|
| 364 |
+
"grad_norm": 0.0007407303201034665,
|
| 365 |
+
"learning_rate": 1.5019291819291821e-05,
|
| 366 |
+
"loss": 0.0004,
|
| 367 |
+
"step": 5100
|
| 368 |
+
},
|
| 369 |
+
{
|
| 370 |
+
"epoch": 0.7619047619047619,
|
| 371 |
+
"grad_norm": 0.0013244397705420852,
|
| 372 |
+
"learning_rate": 1.4921611721611721e-05,
|
| 373 |
+
"loss": 0.0002,
|
| 374 |
+
"step": 5200
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.7765567765567766,
|
| 378 |
+
"grad_norm": 0.0006274768384173512,
|
| 379 |
+
"learning_rate": 1.4823931623931625e-05,
|
| 380 |
+
"loss": 0.0001,
|
| 381 |
+
"step": 5300
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 0.7912087912087912,
|
| 385 |
+
"grad_norm": 0.001210793387144804,
|
| 386 |
+
"learning_rate": 1.4726251526251527e-05,
|
| 387 |
+
"loss": 0.0003,
|
| 388 |
+
"step": 5400
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 0.8058608058608059,
|
| 392 |
+
"grad_norm": 0.002480907831341028,
|
| 393 |
+
"learning_rate": 1.462857142857143e-05,
|
| 394 |
+
"loss": 0.0005,
|
| 395 |
+
"step": 5500
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 0.8205128205128205,
|
| 399 |
+
"grad_norm": 0.0005772067815996706,
|
| 400 |
+
"learning_rate": 1.4530891330891333e-05,
|
| 401 |
+
"loss": 0.0008,
|
| 402 |
+
"step": 5600
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 0.8351648351648352,
|
| 406 |
+
"grad_norm": 0.002250727266073227,
|
| 407 |
+
"learning_rate": 1.4433211233211234e-05,
|
| 408 |
+
"loss": 0.0023,
|
| 409 |
+
"step": 5700
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 0.8498168498168498,
|
| 413 |
+
"grad_norm": 0.001360761933028698,
|
| 414 |
+
"learning_rate": 1.4335531135531138e-05,
|
| 415 |
+
"loss": 0.0001,
|
| 416 |
+
"step": 5800
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.8644688644688645,
|
| 420 |
+
"grad_norm": 0.004648945759981871,
|
| 421 |
+
"learning_rate": 1.4237851037851038e-05,
|
| 422 |
+
"loss": 0.0001,
|
| 423 |
+
"step": 5900
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 0.8791208791208791,
|
| 427 |
+
"grad_norm": 0.0005541480495594442,
|
| 428 |
+
"learning_rate": 1.414017094017094e-05,
|
| 429 |
+
"loss": 0.0001,
|
| 430 |
+
"step": 6000
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 0.8937728937728938,
|
| 434 |
+
"grad_norm": 0.0015395252266898751,
|
| 435 |
+
"learning_rate": 1.4042490842490844e-05,
|
| 436 |
+
"loss": 0.0003,
|
| 437 |
+
"step": 6100
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 0.9084249084249084,
|
| 441 |
+
"grad_norm": 0.0007218205719254911,
|
| 442 |
+
"learning_rate": 1.3944810744810746e-05,
|
| 443 |
+
"loss": 0.0001,
|
| 444 |
+
"step": 6200
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 0.9230769230769231,
|
| 448 |
+
"grad_norm": 0.0005430618184618652,
|
| 449 |
+
"learning_rate": 1.384713064713065e-05,
|
| 450 |
+
"loss": 0.0,
|
| 451 |
+
"step": 6300
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 0.9377289377289377,
|
| 455 |
+
"grad_norm": 0.0004149100568611175,
|
| 456 |
+
"learning_rate": 1.3749450549450552e-05,
|
| 457 |
+
"loss": 0.0,
|
| 458 |
+
"step": 6400
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 0.9523809523809523,
|
| 462 |
+
"grad_norm": 0.0004269062774255872,
|
| 463 |
+
"learning_rate": 1.3651770451770452e-05,
|
| 464 |
+
"loss": 0.0,
|
| 465 |
+
"step": 6500
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 0.967032967032967,
|
| 469 |
+
"grad_norm": 0.0002754240995272994,
|
| 470 |
+
"learning_rate": 1.3554090354090354e-05,
|
| 471 |
+
"loss": 0.0001,
|
| 472 |
+
"step": 6600
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 0.9816849816849816,
|
| 476 |
+
"grad_norm": 0.0005386194679886103,
|
| 477 |
+
"learning_rate": 1.3456410256410257e-05,
|
| 478 |
+
"loss": 0.001,
|
| 479 |
+
"step": 6700
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 0.9963369963369964,
|
| 483 |
+
"grad_norm": 0.0007499135099351406,
|
| 484 |
+
"learning_rate": 1.335873015873016e-05,
|
| 485 |
+
"loss": 0.0006,
|
| 486 |
+
"step": 6800
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 1.0,
|
| 490 |
+
"eval_accuracy": 0.9999860922583947,
|
| 491 |
+
"eval_f1": 0.9996811563396747,
|
| 492 |
+
"eval_loss": 0.0001409229007549584,
|
| 493 |
+
"eval_precision": 0.9996356238423466,
|
| 494 |
+
"eval_recall": 0.99972669298512,
|
| 495 |
+
"eval_runtime": 299.7618,
|
| 496 |
+
"eval_samples_per_second": 91.072,
|
| 497 |
+
"eval_steps_per_second": 5.695,
|
| 498 |
+
"step": 6825
|
| 499 |
+
},
|
| 500 |
+
{
|
| 501 |
+
"epoch": 1.010989010989011,
|
| 502 |
+
"grad_norm": 0.00030950666405260563,
|
| 503 |
+
"learning_rate": 1.3261050061050063e-05,
|
| 504 |
+
"loss": 0.0002,
|
| 505 |
+
"step": 6900
|
| 506 |
+
},
|
| 507 |
+
{
|
| 508 |
+
"epoch": 1.0256410256410255,
|
| 509 |
+
"grad_norm": 0.0019652473274618387,
|
| 510 |
+
"learning_rate": 1.3163369963369965e-05,
|
| 511 |
+
"loss": 0.0012,
|
| 512 |
+
"step": 7000
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 1.0402930402930404,
|
| 516 |
+
"grad_norm": 0.00197768397629261,
|
| 517 |
+
"learning_rate": 1.3065689865689869e-05,
|
| 518 |
+
"loss": 0.0009,
|
| 519 |
+
"step": 7100
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 1.054945054945055,
|
| 523 |
+
"grad_norm": 0.0035088348668068647,
|
| 524 |
+
"learning_rate": 1.2968009768009769e-05,
|
| 525 |
+
"loss": 0.0004,
|
| 526 |
+
"step": 7200
|
| 527 |
+
},
|
| 528 |
+
{
|
| 529 |
+
"epoch": 1.0695970695970696,
|
| 530 |
+
"grad_norm": 0.0010668466566130519,
|
| 531 |
+
"learning_rate": 1.287032967032967e-05,
|
| 532 |
+
"loss": 0.0001,
|
| 533 |
+
"step": 7300
|
| 534 |
+
},
|
| 535 |
+
{
|
| 536 |
+
"epoch": 1.0842490842490842,
|
| 537 |
+
"grad_norm": 0.0011741893831640482,
|
| 538 |
+
"learning_rate": 1.2772649572649573e-05,
|
| 539 |
+
"loss": 0.0001,
|
| 540 |
+
"step": 7400
|
| 541 |
+
},
|
| 542 |
+
{
|
| 543 |
+
"epoch": 1.098901098901099,
|
| 544 |
+
"grad_norm": 0.0011687855003401637,
|
| 545 |
+
"learning_rate": 1.2674969474969476e-05,
|
| 546 |
+
"loss": 0.0002,
|
| 547 |
+
"step": 7500
|
| 548 |
+
},
|
| 549 |
+
{
|
| 550 |
+
"epoch": 1.1135531135531136,
|
| 551 |
+
"grad_norm": 0.0013387134531512856,
|
| 552 |
+
"learning_rate": 1.2577289377289378e-05,
|
| 553 |
+
"loss": 0.0009,
|
| 554 |
+
"step": 7600
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 1.1282051282051282,
|
| 558 |
+
"grad_norm": 0.004518842324614525,
|
| 559 |
+
"learning_rate": 1.2479609279609282e-05,
|
| 560 |
+
"loss": 0.0,
|
| 561 |
+
"step": 7700
|
| 562 |
+
},
|
| 563 |
+
{
|
| 564 |
+
"epoch": 1.1428571428571428,
|
| 565 |
+
"grad_norm": 0.007567571010440588,
|
| 566 |
+
"learning_rate": 1.2381929181929182e-05,
|
| 567 |
+
"loss": 0.0001,
|
| 568 |
+
"step": 7800
|
| 569 |
+
},
|
| 570 |
+
{
|
| 571 |
+
"epoch": 1.1575091575091574,
|
| 572 |
+
"grad_norm": 0.000706327729858458,
|
| 573 |
+
"learning_rate": 1.2284249084249084e-05,
|
| 574 |
+
"loss": 0.0002,
|
| 575 |
+
"step": 7900
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"epoch": 1.1721611721611722,
|
| 579 |
+
"grad_norm": 0.0004430621338542551,
|
| 580 |
+
"learning_rate": 1.2186568986568988e-05,
|
| 581 |
+
"loss": 0.0004,
|
| 582 |
+
"step": 8000
|
| 583 |
+
},
|
| 584 |
+
{
|
| 585 |
+
"epoch": 1.1868131868131868,
|
| 586 |
+
"grad_norm": 0.003274950897321105,
|
| 587 |
+
"learning_rate": 1.208888888888889e-05,
|
| 588 |
+
"loss": 0.0,
|
| 589 |
+
"step": 8100
|
| 590 |
+
},
|
| 591 |
+
{
|
| 592 |
+
"epoch": 1.2014652014652014,
|
| 593 |
+
"grad_norm": 0.009472182020545006,
|
| 594 |
+
"learning_rate": 1.1991208791208793e-05,
|
| 595 |
+
"loss": 0.0,
|
| 596 |
+
"step": 8200
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 1.2161172161172162,
|
| 600 |
+
"grad_norm": 0.000377336866222322,
|
| 601 |
+
"learning_rate": 1.1893528693528695e-05,
|
| 602 |
+
"loss": 0.0001,
|
| 603 |
+
"step": 8300
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"epoch": 1.2307692307692308,
|
| 607 |
+
"grad_norm": 0.00038757469155825675,
|
| 608 |
+
"learning_rate": 1.1795848595848596e-05,
|
| 609 |
+
"loss": 0.0002,
|
| 610 |
+
"step": 8400
|
| 611 |
+
},
|
| 612 |
+
{
|
| 613 |
+
"epoch": 1.2454212454212454,
|
| 614 |
+
"grad_norm": 0.000478828267659992,
|
| 615 |
+
"learning_rate": 1.1698168498168498e-05,
|
| 616 |
+
"loss": 0.0,
|
| 617 |
+
"step": 8500
|
| 618 |
+
},
|
| 619 |
+
{
|
| 620 |
+
"epoch": 1.26007326007326,
|
| 621 |
+
"grad_norm": 0.0003117730957455933,
|
| 622 |
+
"learning_rate": 1.1600488400488401e-05,
|
| 623 |
+
"loss": 0.0002,
|
| 624 |
+
"step": 8600
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"epoch": 1.2747252747252746,
|
| 628 |
+
"grad_norm": 0.0005039616953581572,
|
| 629 |
+
"learning_rate": 1.1502808302808303e-05,
|
| 630 |
+
"loss": 0.0001,
|
| 631 |
+
"step": 8700
|
| 632 |
+
},
|
| 633 |
+
{
|
| 634 |
+
"epoch": 1.2893772893772895,
|
| 635 |
+
"grad_norm": 0.00023192950175143778,
|
| 636 |
+
"learning_rate": 1.1405128205128207e-05,
|
| 637 |
+
"loss": 0.0,
|
| 638 |
+
"step": 8800
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 1.304029304029304,
|
| 642 |
+
"grad_norm": 0.0033253557048738003,
|
| 643 |
+
"learning_rate": 1.1307448107448109e-05,
|
| 644 |
+
"loss": 0.0004,
|
| 645 |
+
"step": 8900
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 1.3186813186813187,
|
| 649 |
+
"grad_norm": 0.00022548828565049917,
|
| 650 |
+
"learning_rate": 1.1209768009768013e-05,
|
| 651 |
+
"loss": 0.0,
|
| 652 |
+
"step": 9000
|
| 653 |
+
},
|
| 654 |
+
{
|
| 655 |
+
"epoch": 1.3333333333333333,
|
| 656 |
+
"grad_norm": 0.00022915085719432682,
|
| 657 |
+
"learning_rate": 1.1112087912087913e-05,
|
| 658 |
+
"loss": 0.0,
|
| 659 |
+
"step": 9100
|
| 660 |
+
},
|
| 661 |
+
{
|
| 662 |
+
"epoch": 1.347985347985348,
|
| 663 |
+
"grad_norm": 0.00038897068588994443,
|
| 664 |
+
"learning_rate": 1.1014407814407815e-05,
|
| 665 |
+
"loss": 0.0003,
|
| 666 |
+
"step": 9200
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"epoch": 1.3626373626373627,
|
| 670 |
+
"grad_norm": 0.00022964017989579588,
|
| 671 |
+
"learning_rate": 1.0916727716727717e-05,
|
| 672 |
+
"loss": 0.0,
|
| 673 |
+
"step": 9300
|
| 674 |
+
},
|
| 675 |
+
{
|
| 676 |
+
"epoch": 1.3772893772893773,
|
| 677 |
+
"grad_norm": 0.00019336077093612403,
|
| 678 |
+
"learning_rate": 1.081904761904762e-05,
|
| 679 |
+
"loss": 0.0,
|
| 680 |
+
"step": 9400
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 1.3919413919413919,
|
| 684 |
+
"grad_norm": 0.00016649049939587712,
|
| 685 |
+
"learning_rate": 1.0721367521367522e-05,
|
| 686 |
+
"loss": 0.0,
|
| 687 |
+
"step": 9500
|
| 688 |
+
},
|
| 689 |
+
{
|
| 690 |
+
"epoch": 1.4065934065934065,
|
| 691 |
+
"grad_norm": 0.0029115676879882812,
|
| 692 |
+
"learning_rate": 1.0623687423687426e-05,
|
| 693 |
+
"loss": 0.0,
|
| 694 |
+
"step": 9600
|
| 695 |
+
},
|
| 696 |
+
{
|
| 697 |
+
"epoch": 1.4212454212454213,
|
| 698 |
+
"grad_norm": 0.00016056567255873233,
|
| 699 |
+
"learning_rate": 1.0526007326007326e-05,
|
| 700 |
+
"loss": 0.0,
|
| 701 |
+
"step": 9700
|
| 702 |
+
},
|
| 703 |
+
{
|
| 704 |
+
"epoch": 1.435897435897436,
|
| 705 |
+
"grad_norm": 0.0007812806870788336,
|
| 706 |
+
"learning_rate": 1.0428327228327228e-05,
|
| 707 |
+
"loss": 0.0015,
|
| 708 |
+
"step": 9800
|
| 709 |
+
},
|
| 710 |
+
{
|
| 711 |
+
"epoch": 1.4505494505494505,
|
| 712 |
+
"grad_norm": 0.0008388872374780476,
|
| 713 |
+
"learning_rate": 1.0330647130647132e-05,
|
| 714 |
+
"loss": 0.0003,
|
| 715 |
+
"step": 9900
|
| 716 |
+
},
|
| 717 |
+
{
|
| 718 |
+
"epoch": 1.4652014652014653,
|
| 719 |
+
"grad_norm": 0.0003454253019299358,
|
| 720 |
+
"learning_rate": 1.0232967032967034e-05,
|
| 721 |
+
"loss": 0.0001,
|
| 722 |
+
"step": 10000
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 1.47985347985348,
|
| 726 |
+
"grad_norm": 0.00045976179535500705,
|
| 727 |
+
"learning_rate": 1.0135286935286936e-05,
|
| 728 |
+
"loss": 0.0,
|
| 729 |
+
"step": 10100
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 1.4945054945054945,
|
| 733 |
+
"grad_norm": 0.00019197350775357336,
|
| 734 |
+
"learning_rate": 1.003760683760684e-05,
|
| 735 |
+
"loss": 0.0,
|
| 736 |
+
"step": 10200
|
| 737 |
+
},
|
| 738 |
+
{
|
| 739 |
+
"epoch": 1.5091575091575091,
|
| 740 |
+
"grad_norm": 0.00016801034507807344,
|
| 741 |
+
"learning_rate": 9.939926739926741e-06,
|
| 742 |
+
"loss": 0.0001,
|
| 743 |
+
"step": 10300
|
| 744 |
+
},
|
| 745 |
+
{
|
| 746 |
+
"epoch": 1.5238095238095237,
|
| 747 |
+
"grad_norm": 1.131913423538208,
|
| 748 |
+
"learning_rate": 9.842246642246643e-06,
|
| 749 |
+
"loss": 0.0001,
|
| 750 |
+
"step": 10400
|
| 751 |
+
},
|
| 752 |
+
{
|
| 753 |
+
"epoch": 1.5384615384615383,
|
| 754 |
+
"grad_norm": 0.0002565563772805035,
|
| 755 |
+
"learning_rate": 9.744566544566545e-06,
|
| 756 |
+
"loss": 0.0,
|
| 757 |
+
"step": 10500
|
| 758 |
+
},
|
| 759 |
+
{
|
| 760 |
+
"epoch": 1.5531135531135531,
|
| 761 |
+
"grad_norm": 0.00042528833728283644,
|
| 762 |
+
"learning_rate": 9.646886446886447e-06,
|
| 763 |
+
"loss": 0.0002,
|
| 764 |
+
"step": 10600
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 1.5677655677655677,
|
| 768 |
+
"grad_norm": 0.012502134777605534,
|
| 769 |
+
"learning_rate": 9.54920634920635e-06,
|
| 770 |
+
"loss": 0.0003,
|
| 771 |
+
"step": 10700
|
| 772 |
+
},
|
| 773 |
+
{
|
| 774 |
+
"epoch": 1.5824175824175826,
|
| 775 |
+
"grad_norm": 0.0005918039241805673,
|
| 776 |
+
"learning_rate": 9.451526251526251e-06,
|
| 777 |
+
"loss": 0.0,
|
| 778 |
+
"step": 10800
|
| 779 |
+
},
|
| 780 |
+
{
|
| 781 |
+
"epoch": 1.5970695970695972,
|
| 782 |
+
"grad_norm": 0.0017294063000008464,
|
| 783 |
+
"learning_rate": 9.353846153846155e-06,
|
| 784 |
+
"loss": 0.0,
|
| 785 |
+
"step": 10900
|
| 786 |
+
},
|
| 787 |
+
{
|
| 788 |
+
"epoch": 1.6117216117216118,
|
| 789 |
+
"grad_norm": 0.00014640063454862684,
|
| 790 |
+
"learning_rate": 9.256166056166057e-06,
|
| 791 |
+
"loss": 0.0,
|
| 792 |
+
"step": 11000
|
| 793 |
+
},
|
| 794 |
+
{
|
| 795 |
+
"epoch": 1.6263736263736264,
|
| 796 |
+
"grad_norm": 0.0004110069421585649,
|
| 797 |
+
"learning_rate": 9.15848595848596e-06,
|
| 798 |
+
"loss": 0.0,
|
| 799 |
+
"step": 11100
|
| 800 |
+
},
|
| 801 |
+
{
|
| 802 |
+
"epoch": 1.641025641025641,
|
| 803 |
+
"grad_norm": 0.00027677713660523295,
|
| 804 |
+
"learning_rate": 9.06080586080586e-06,
|
| 805 |
+
"loss": 0.0001,
|
| 806 |
+
"step": 11200
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 1.6556776556776556,
|
| 810 |
+
"grad_norm": 0.0006185442907735705,
|
| 811 |
+
"learning_rate": 8.963125763125764e-06,
|
| 812 |
+
"loss": 0.0,
|
| 813 |
+
"step": 11300
|
| 814 |
+
},
|
| 815 |
+
{
|
| 816 |
+
"epoch": 1.6703296703296702,
|
| 817 |
+
"grad_norm": 0.00041590031469240785,
|
| 818 |
+
"learning_rate": 8.865445665445666e-06,
|
| 819 |
+
"loss": 0.0,
|
| 820 |
+
"step": 11400
|
| 821 |
+
},
|
| 822 |
+
{
|
| 823 |
+
"epoch": 1.684981684981685,
|
| 824 |
+
"grad_norm": 0.00022950036509428173,
|
| 825 |
+
"learning_rate": 8.767765567765568e-06,
|
| 826 |
+
"loss": 0.0,
|
| 827 |
+
"step": 11500
|
| 828 |
+
},
|
| 829 |
+
{
|
| 830 |
+
"epoch": 1.6996336996336996,
|
| 831 |
+
"grad_norm": 0.00020648095232900232,
|
| 832 |
+
"learning_rate": 8.67008547008547e-06,
|
| 833 |
+
"loss": 0.0,
|
| 834 |
+
"step": 11600
|
| 835 |
+
},
|
| 836 |
+
{
|
| 837 |
+
"epoch": 1.7142857142857144,
|
| 838 |
+
"grad_norm": 8.724381041247398e-05,
|
| 839 |
+
"learning_rate": 8.572405372405374e-06,
|
| 840 |
+
"loss": 0.0001,
|
| 841 |
+
"step": 11700
|
| 842 |
+
},
|
| 843 |
+
{
|
| 844 |
+
"epoch": 1.728937728937729,
|
| 845 |
+
"grad_norm": 0.0001912677544169128,
|
| 846 |
+
"learning_rate": 8.474725274725276e-06,
|
| 847 |
+
"loss": 0.0,
|
| 848 |
+
"step": 11800
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 1.7435897435897436,
|
| 852 |
+
"grad_norm": 0.00011383296805433929,
|
| 853 |
+
"learning_rate": 8.377045177045178e-06,
|
| 854 |
+
"loss": 0.0,
|
| 855 |
+
"step": 11900
|
| 856 |
+
},
|
| 857 |
+
{
|
| 858 |
+
"epoch": 1.7582417582417582,
|
| 859 |
+
"grad_norm": 0.00030712466104887426,
|
| 860 |
+
"learning_rate": 8.27936507936508e-06,
|
| 861 |
+
"loss": 0.0002,
|
| 862 |
+
"step": 12000
|
| 863 |
+
},
|
| 864 |
+
{
|
| 865 |
+
"epoch": 1.7728937728937728,
|
| 866 |
+
"grad_norm": 0.00013500441855285317,
|
| 867 |
+
"learning_rate": 8.181684981684982e-06,
|
| 868 |
+
"loss": 0.0,
|
| 869 |
+
"step": 12100
|
| 870 |
+
},
|
| 871 |
+
{
|
| 872 |
+
"epoch": 1.7875457875457874,
|
| 873 |
+
"grad_norm": 0.00015925474872346967,
|
| 874 |
+
"learning_rate": 8.084004884004885e-06,
|
| 875 |
+
"loss": 0.0007,
|
| 876 |
+
"step": 12200
|
| 877 |
+
},
|
| 878 |
+
{
|
| 879 |
+
"epoch": 1.8021978021978022,
|
| 880 |
+
"grad_norm": 0.0006131925620138645,
|
| 881 |
+
"learning_rate": 7.986324786324787e-06,
|
| 882 |
+
"loss": 0.0,
|
| 883 |
+
"step": 12300
|
| 884 |
+
},
|
| 885 |
+
{
|
| 886 |
+
"epoch": 1.8168498168498168,
|
| 887 |
+
"grad_norm": 0.001735904486849904,
|
| 888 |
+
"learning_rate": 7.888644688644689e-06,
|
| 889 |
+
"loss": 0.0002,
|
| 890 |
+
"step": 12400
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 1.8315018315018317,
|
| 894 |
+
"grad_norm": 0.0001556806091684848,
|
| 895 |
+
"learning_rate": 7.790964590964591e-06,
|
| 896 |
+
"loss": 0.0,
|
| 897 |
+
"step": 12500
|
| 898 |
+
},
|
| 899 |
+
{
|
| 900 |
+
"epoch": 1.8461538461538463,
|
| 901 |
+
"grad_norm": 1.0768227577209473,
|
| 902 |
+
"learning_rate": 7.693284493284495e-06,
|
| 903 |
+
"loss": 0.0003,
|
| 904 |
+
"step": 12600
|
| 905 |
+
},
|
| 906 |
+
{
|
| 907 |
+
"epoch": 1.8608058608058609,
|
| 908 |
+
"grad_norm": 0.00020022221724502742,
|
| 909 |
+
"learning_rate": 7.595604395604397e-06,
|
| 910 |
+
"loss": 0.0001,
|
| 911 |
+
"step": 12700
|
| 912 |
+
},
|
| 913 |
+
{
|
| 914 |
+
"epoch": 1.8754578754578755,
|
| 915 |
+
"grad_norm": 0.00016349930956494063,
|
| 916 |
+
"learning_rate": 7.497924297924299e-06,
|
| 917 |
+
"loss": 0.0,
|
| 918 |
+
"step": 12800
|
| 919 |
+
},
|
| 920 |
+
{
|
| 921 |
+
"epoch": 1.89010989010989,
|
| 922 |
+
"grad_norm": 0.0001262535952264443,
|
| 923 |
+
"learning_rate": 7.400244200244201e-06,
|
| 924 |
+
"loss": 0.0,
|
| 925 |
+
"step": 12900
|
| 926 |
+
},
|
| 927 |
+
{
|
| 928 |
+
"epoch": 1.9047619047619047,
|
| 929 |
+
"grad_norm": 0.003379235276952386,
|
| 930 |
+
"learning_rate": 7.302564102564103e-06,
|
| 931 |
+
"loss": 0.0003,
|
| 932 |
+
"step": 13000
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 1.9194139194139193,
|
| 936 |
+
"grad_norm": 0.00012544514902401716,
|
| 937 |
+
"learning_rate": 7.204884004884005e-06,
|
| 938 |
+
"loss": 0.0,
|
| 939 |
+
"step": 13100
|
| 940 |
+
},
|
| 941 |
+
{
|
| 942 |
+
"epoch": 1.934065934065934,
|
| 943 |
+
"grad_norm": 0.0004016385355498642,
|
| 944 |
+
"learning_rate": 7.107203907203908e-06,
|
| 945 |
+
"loss": 0.0,
|
| 946 |
+
"step": 13200
|
| 947 |
+
},
|
| 948 |
+
{
|
| 949 |
+
"epoch": 1.9487179487179487,
|
| 950 |
+
"grad_norm": 0.0003825027379207313,
|
| 951 |
+
"learning_rate": 7.00952380952381e-06,
|
| 952 |
+
"loss": 0.0003,
|
| 953 |
+
"step": 13300
|
| 954 |
+
},
|
| 955 |
+
{
|
| 956 |
+
"epoch": 1.9633699633699635,
|
| 957 |
+
"grad_norm": 0.00017085819854401052,
|
| 958 |
+
"learning_rate": 6.911843711843712e-06,
|
| 959 |
+
"loss": 0.0,
|
| 960 |
+
"step": 13400
|
| 961 |
+
},
|
| 962 |
+
{
|
| 963 |
+
"epoch": 1.978021978021978,
|
| 964 |
+
"grad_norm": 0.0001403139322064817,
|
| 965 |
+
"learning_rate": 6.814163614163615e-06,
|
| 966 |
+
"loss": 0.0,
|
| 967 |
+
"step": 13500
|
| 968 |
+
},
|
| 969 |
+
{
|
| 970 |
+
"epoch": 1.9926739926739927,
|
| 971 |
+
"grad_norm": 0.003122469875961542,
|
| 972 |
+
"learning_rate": 6.716483516483518e-06,
|
| 973 |
+
"loss": 0.0,
|
| 974 |
+
"step": 13600
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 2.0,
|
| 978 |
+
"eval_accuracy": 0.9999990728172263,
|
| 979 |
+
"eval_f1": 0.9999544495224792,
|
| 980 |
+
"eval_loss": 1.237483047589194e-05,
|
| 981 |
+
"eval_precision": 0.999939266952112,
|
| 982 |
+
"eval_recall": 0.9999696325539023,
|
| 983 |
+
"eval_runtime": 267.2853,
|
| 984 |
+
"eval_samples_per_second": 102.138,
|
| 985 |
+
"eval_steps_per_second": 6.386,
|
| 986 |
+
"step": 13650
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 2.0073260073260073,
|
| 990 |
+
"grad_norm": 6.217395275598392e-05,
|
| 991 |
+
"learning_rate": 6.618803418803419e-06,
|
| 992 |
+
"loss": 0.0,
|
| 993 |
+
"step": 13700
|
| 994 |
+
},
|
| 995 |
+
{
|
| 996 |
+
"epoch": 2.021978021978022,
|
| 997 |
+
"grad_norm": 0.0001487318950239569,
|
| 998 |
+
"learning_rate": 6.5211233211233216e-06,
|
| 999 |
+
"loss": 0.0003,
|
| 1000 |
+
"step": 13800
|
| 1001 |
+
},
|
| 1002 |
+
{
|
| 1003 |
+
"epoch": 2.0366300366300365,
|
| 1004 |
+
"grad_norm": 0.00045171970850788057,
|
| 1005 |
+
"learning_rate": 6.423443223443224e-06,
|
| 1006 |
+
"loss": 0.0001,
|
| 1007 |
+
"step": 13900
|
| 1008 |
+
},
|
| 1009 |
+
{
|
| 1010 |
+
"epoch": 2.051282051282051,
|
| 1011 |
+
"grad_norm": 0.00022699052351526916,
|
| 1012 |
+
"learning_rate": 6.3257631257631255e-06,
|
| 1013 |
+
"loss": 0.0,
|
| 1014 |
+
"step": 14000
|
| 1015 |
+
},
|
| 1016 |
+
{
|
| 1017 |
+
"epoch": 2.065934065934066,
|
| 1018 |
+
"grad_norm": 9.758808300830424e-05,
|
| 1019 |
+
"learning_rate": 6.228083028083028e-06,
|
| 1020 |
+
"loss": 0.0,
|
| 1021 |
+
"step": 14100
|
| 1022 |
+
},
|
| 1023 |
+
{
|
| 1024 |
+
"epoch": 2.0805860805860807,
|
| 1025 |
+
"grad_norm": 0.00048689660616219044,
|
| 1026 |
+
"learning_rate": 6.130402930402931e-06,
|
| 1027 |
+
"loss": 0.0,
|
| 1028 |
+
"step": 14200
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 2.0952380952380953,
|
| 1032 |
+
"grad_norm": 7.219218969112262e-05,
|
| 1033 |
+
"learning_rate": 6.032722832722834e-06,
|
| 1034 |
+
"loss": 0.0,
|
| 1035 |
+
"step": 14300
|
| 1036 |
+
},
|
| 1037 |
+
{
|
| 1038 |
+
"epoch": 2.10989010989011,
|
| 1039 |
+
"grad_norm": 0.00010942752123810351,
|
| 1040 |
+
"learning_rate": 5.935042735042735e-06,
|
| 1041 |
+
"loss": 0.0002,
|
| 1042 |
+
"step": 14400
|
| 1043 |
+
},
|
| 1044 |
+
{
|
| 1045 |
+
"epoch": 2.1245421245421245,
|
| 1046 |
+
"grad_norm": 0.00013398531882558018,
|
| 1047 |
+
"learning_rate": 5.837362637362638e-06,
|
| 1048 |
+
"loss": 0.0,
|
| 1049 |
+
"step": 14500
|
| 1050 |
+
},
|
| 1051 |
+
{
|
| 1052 |
+
"epoch": 2.139194139194139,
|
| 1053 |
+
"grad_norm": 0.00013351649977266788,
|
| 1054 |
+
"learning_rate": 5.739682539682541e-06,
|
| 1055 |
+
"loss": 0.0,
|
| 1056 |
+
"step": 14600
|
| 1057 |
+
},
|
| 1058 |
+
{
|
| 1059 |
+
"epoch": 2.1538461538461537,
|
| 1060 |
+
"grad_norm": 8.359822095371783e-05,
|
| 1061 |
+
"learning_rate": 5.6420024420024425e-06,
|
| 1062 |
+
"loss": 0.0,
|
| 1063 |
+
"step": 14700
|
| 1064 |
+
},
|
| 1065 |
+
{
|
| 1066 |
+
"epoch": 2.1684981684981683,
|
| 1067 |
+
"grad_norm": 7.686048775212839e-05,
|
| 1068 |
+
"learning_rate": 5.5443223443223445e-06,
|
| 1069 |
+
"loss": 0.0,
|
| 1070 |
+
"step": 14800
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 2.183150183150183,
|
| 1074 |
+
"grad_norm": 0.00011812122829724103,
|
| 1075 |
+
"learning_rate": 5.446642246642247e-06,
|
| 1076 |
+
"loss": 0.0,
|
| 1077 |
+
"step": 14900
|
| 1078 |
+
},
|
| 1079 |
+
{
|
| 1080 |
+
"epoch": 2.197802197802198,
|
| 1081 |
+
"grad_norm": 0.00044185685692355037,
|
| 1082 |
+
"learning_rate": 5.348962148962149e-06,
|
| 1083 |
+
"loss": 0.0001,
|
| 1084 |
+
"step": 15000
|
| 1085 |
+
},
|
| 1086 |
+
{
|
| 1087 |
+
"epoch": 2.2124542124542126,
|
| 1088 |
+
"grad_norm": 0.00013721364666707814,
|
| 1089 |
+
"learning_rate": 5.251282051282052e-06,
|
| 1090 |
+
"loss": 0.0001,
|
| 1091 |
+
"step": 15100
|
| 1092 |
+
},
|
| 1093 |
+
{
|
| 1094 |
+
"epoch": 2.227106227106227,
|
| 1095 |
+
"grad_norm": 0.00018953319522552192,
|
| 1096 |
+
"learning_rate": 5.153601953601954e-06,
|
| 1097 |
+
"loss": 0.0,
|
| 1098 |
+
"step": 15200
|
| 1099 |
+
},
|
| 1100 |
+
{
|
| 1101 |
+
"epoch": 2.241758241758242,
|
| 1102 |
+
"grad_norm": 0.0003042246389668435,
|
| 1103 |
+
"learning_rate": 5.055921855921856e-06,
|
| 1104 |
+
"loss": 0.0,
|
| 1105 |
+
"step": 15300
|
| 1106 |
+
},
|
| 1107 |
+
{
|
| 1108 |
+
"epoch": 2.2564102564102564,
|
| 1109 |
+
"grad_norm": 7.083150558173656e-05,
|
| 1110 |
+
"learning_rate": 4.958241758241759e-06,
|
| 1111 |
+
"loss": 0.0,
|
| 1112 |
+
"step": 15400
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 2.271062271062271,
|
| 1116 |
+
"grad_norm": 8.48570343805477e-05,
|
| 1117 |
+
"learning_rate": 4.8605616605616616e-06,
|
| 1118 |
+
"loss": 0.0,
|
| 1119 |
+
"step": 15500
|
| 1120 |
+
},
|
| 1121 |
+
{
|
| 1122 |
+
"epoch": 2.2857142857142856,
|
| 1123 |
+
"grad_norm": 7.620136602781713e-05,
|
| 1124 |
+
"learning_rate": 4.7628815628815635e-06,
|
| 1125 |
+
"loss": 0.0,
|
| 1126 |
+
"step": 15600
|
| 1127 |
+
},
|
| 1128 |
+
{
|
| 1129 |
+
"epoch": 2.3003663003663,
|
| 1130 |
+
"grad_norm": 6.34704774711281e-05,
|
| 1131 |
+
"learning_rate": 4.6652014652014655e-06,
|
| 1132 |
+
"loss": 0.0,
|
| 1133 |
+
"step": 15700
|
| 1134 |
+
},
|
| 1135 |
+
{
|
| 1136 |
+
"epoch": 2.315018315018315,
|
| 1137 |
+
"grad_norm": 8.851837628753856e-05,
|
| 1138 |
+
"learning_rate": 4.567521367521368e-06,
|
| 1139 |
+
"loss": 0.0001,
|
| 1140 |
+
"step": 15800
|
| 1141 |
+
},
|
| 1142 |
+
{
|
| 1143 |
+
"epoch": 2.32967032967033,
|
| 1144 |
+
"grad_norm": 0.0001343141047982499,
|
| 1145 |
+
"learning_rate": 4.46984126984127e-06,
|
| 1146 |
+
"loss": 0.0001,
|
| 1147 |
+
"step": 15900
|
| 1148 |
+
},
|
| 1149 |
+
{
|
| 1150 |
+
"epoch": 2.3443223443223444,
|
| 1151 |
+
"grad_norm": 0.0001232538343174383,
|
| 1152 |
+
"learning_rate": 4.372161172161172e-06,
|
| 1153 |
+
"loss": 0.0,
|
| 1154 |
+
"step": 16000
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 2.358974358974359,
|
| 1158 |
+
"grad_norm": 5.023027551942505e-05,
|
| 1159 |
+
"learning_rate": 4.274481074481075e-06,
|
| 1160 |
+
"loss": 0.0,
|
| 1161 |
+
"step": 16100
|
| 1162 |
+
},
|
| 1163 |
+
{
|
| 1164 |
+
"epoch": 2.3736263736263736,
|
| 1165 |
+
"grad_norm": 8.723604696569964e-05,
|
| 1166 |
+
"learning_rate": 4.176800976800977e-06,
|
| 1167 |
+
"loss": 0.0,
|
| 1168 |
+
"step": 16200
|
| 1169 |
+
},
|
| 1170 |
+
{
|
| 1171 |
+
"epoch": 2.3882783882783882,
|
| 1172 |
+
"grad_norm": 5.652988329529762e-05,
|
| 1173 |
+
"learning_rate": 4.07912087912088e-06,
|
| 1174 |
+
"loss": 0.0,
|
| 1175 |
+
"step": 16300
|
| 1176 |
+
},
|
| 1177 |
+
{
|
| 1178 |
+
"epoch": 2.402930402930403,
|
| 1179 |
+
"grad_norm": 6.024859976605512e-05,
|
| 1180 |
+
"learning_rate": 3.981440781440782e-06,
|
| 1181 |
+
"loss": 0.0,
|
| 1182 |
+
"step": 16400
|
| 1183 |
+
},
|
| 1184 |
+
{
|
| 1185 |
+
"epoch": 2.4175824175824174,
|
| 1186 |
+
"grad_norm": 4.5512519136536866e-05,
|
| 1187 |
+
"learning_rate": 3.883760683760684e-06,
|
| 1188 |
+
"loss": 0.0,
|
| 1189 |
+
"step": 16500
|
| 1190 |
+
},
|
| 1191 |
+
{
|
| 1192 |
+
"epoch": 2.4322344322344325,
|
| 1193 |
+
"grad_norm": 6.305933493422344e-05,
|
| 1194 |
+
"learning_rate": 3.7860805860805864e-06,
|
| 1195 |
+
"loss": 0.0,
|
| 1196 |
+
"step": 16600
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 2.446886446886447,
|
| 1200 |
+
"grad_norm": 0.00021723458485212177,
|
| 1201 |
+
"learning_rate": 3.688400488400489e-06,
|
| 1202 |
+
"loss": 0.0,
|
| 1203 |
+
"step": 16700
|
| 1204 |
+
},
|
| 1205 |
+
{
|
| 1206 |
+
"epoch": 2.4615384615384617,
|
| 1207 |
+
"grad_norm": 0.00010556123743299395,
|
| 1208 |
+
"learning_rate": 3.5907203907203908e-06,
|
| 1209 |
+
"loss": 0.0,
|
| 1210 |
+
"step": 16800
|
| 1211 |
+
},
|
| 1212 |
+
{
|
| 1213 |
+
"epoch": 2.4761904761904763,
|
| 1214 |
+
"grad_norm": 4.880682536168024e-05,
|
| 1215 |
+
"learning_rate": 3.4930402930402936e-06,
|
| 1216 |
+
"loss": 0.0,
|
| 1217 |
+
"step": 16900
|
| 1218 |
+
},
|
| 1219 |
+
{
|
| 1220 |
+
"epoch": 2.490842490842491,
|
| 1221 |
+
"grad_norm": 4.9467933422420174e-05,
|
| 1222 |
+
"learning_rate": 3.3953601953601955e-06,
|
| 1223 |
+
"loss": 0.0,
|
| 1224 |
+
"step": 17000
|
| 1225 |
+
},
|
| 1226 |
+
{
|
| 1227 |
+
"epoch": 2.5054945054945055,
|
| 1228 |
+
"grad_norm": 6.394098454620689e-05,
|
| 1229 |
+
"learning_rate": 3.2976800976800983e-06,
|
| 1230 |
+
"loss": 0.0,
|
| 1231 |
+
"step": 17100
|
| 1232 |
+
},
|
| 1233 |
+
{
|
| 1234 |
+
"epoch": 2.52014652014652,
|
| 1235 |
+
"grad_norm": 0.0005223629996180534,
|
| 1236 |
+
"learning_rate": 3.2000000000000003e-06,
|
| 1237 |
+
"loss": 0.0,
|
| 1238 |
+
"step": 17200
|
| 1239 |
+
},
|
| 1240 |
+
{
|
| 1241 |
+
"epoch": 2.5347985347985347,
|
| 1242 |
+
"grad_norm": 3.2207102776737884e-05,
|
| 1243 |
+
"learning_rate": 3.1023199023199022e-06,
|
| 1244 |
+
"loss": 0.0,
|
| 1245 |
+
"step": 17300
|
| 1246 |
+
},
|
| 1247 |
+
{
|
| 1248 |
+
"epoch": 2.5494505494505493,
|
| 1249 |
+
"grad_norm": 6.0004018450854346e-05,
|
| 1250 |
+
"learning_rate": 3.004639804639805e-06,
|
| 1251 |
+
"loss": 0.0001,
|
| 1252 |
+
"step": 17400
|
| 1253 |
+
},
|
| 1254 |
+
{
|
| 1255 |
+
"epoch": 2.564102564102564,
|
| 1256 |
+
"grad_norm": 3.9327616832451895e-05,
|
| 1257 |
+
"learning_rate": 2.906959706959707e-06,
|
| 1258 |
+
"loss": 0.0,
|
| 1259 |
+
"step": 17500
|
| 1260 |
+
},
|
| 1261 |
+
{
|
| 1262 |
+
"epoch": 2.578754578754579,
|
| 1263 |
+
"grad_norm": 4.118070864933543e-05,
|
| 1264 |
+
"learning_rate": 2.8092796092796094e-06,
|
| 1265 |
+
"loss": 0.0,
|
| 1266 |
+
"step": 17600
|
| 1267 |
+
},
|
| 1268 |
+
{
|
| 1269 |
+
"epoch": 2.5934065934065935,
|
| 1270 |
+
"grad_norm": 4.89596750412602e-05,
|
| 1271 |
+
"learning_rate": 2.711599511599512e-06,
|
| 1272 |
+
"loss": 0.0001,
|
| 1273 |
+
"step": 17700
|
| 1274 |
+
},
|
| 1275 |
+
{
|
| 1276 |
+
"epoch": 2.608058608058608,
|
| 1277 |
+
"grad_norm": 5.990030331304297e-05,
|
| 1278 |
+
"learning_rate": 2.613919413919414e-06,
|
| 1279 |
+
"loss": 0.0,
|
| 1280 |
+
"step": 17800
|
| 1281 |
+
},
|
| 1282 |
+
{
|
| 1283 |
+
"epoch": 2.6227106227106227,
|
| 1284 |
+
"grad_norm": 3.487396315904334e-05,
|
| 1285 |
+
"learning_rate": 2.516239316239317e-06,
|
| 1286 |
+
"loss": 0.0002,
|
| 1287 |
+
"step": 17900
|
| 1288 |
+
},
|
| 1289 |
+
{
|
| 1290 |
+
"epoch": 2.6373626373626373,
|
| 1291 |
+
"grad_norm": 2.6920677555608563e-05,
|
| 1292 |
+
"learning_rate": 2.418559218559219e-06,
|
| 1293 |
+
"loss": 0.0001,
|
| 1294 |
+
"step": 18000
|
| 1295 |
+
},
|
| 1296 |
+
{
|
| 1297 |
+
"epoch": 2.652014652014652,
|
| 1298 |
+
"grad_norm": 5.215009514358826e-05,
|
| 1299 |
+
"learning_rate": 2.3208791208791213e-06,
|
| 1300 |
+
"loss": 0.0,
|
| 1301 |
+
"step": 18100
|
| 1302 |
+
},
|
| 1303 |
+
{
|
| 1304 |
+
"epoch": 2.6666666666666665,
|
| 1305 |
+
"grad_norm": 4.3928706872975454e-05,
|
| 1306 |
+
"learning_rate": 2.223199023199023e-06,
|
| 1307 |
+
"loss": 0.0,
|
| 1308 |
+
"step": 18200
|
| 1309 |
+
},
|
| 1310 |
+
{
|
| 1311 |
+
"epoch": 2.6813186813186816,
|
| 1312 |
+
"grad_norm": 0.004979077726602554,
|
| 1313 |
+
"learning_rate": 2.1255189255189256e-06,
|
| 1314 |
+
"loss": 0.0,
|
| 1315 |
+
"step": 18300
|
| 1316 |
+
},
|
| 1317 |
+
{
|
| 1318 |
+
"epoch": 2.695970695970696,
|
| 1319 |
+
"grad_norm": 4.079126665601507e-05,
|
| 1320 |
+
"learning_rate": 2.027838827838828e-06,
|
| 1321 |
+
"loss": 0.0,
|
| 1322 |
+
"step": 18400
|
| 1323 |
+
},
|
| 1324 |
+
{
|
| 1325 |
+
"epoch": 2.7106227106227108,
|
| 1326 |
+
"grad_norm": 3.755389479920268e-05,
|
| 1327 |
+
"learning_rate": 1.9301587301587303e-06,
|
| 1328 |
+
"loss": 0.0,
|
| 1329 |
+
"step": 18500
|
| 1330 |
+
},
|
| 1331 |
+
{
|
| 1332 |
+
"epoch": 2.7252747252747254,
|
| 1333 |
+
"grad_norm": 3.5347176890354604e-05,
|
| 1334 |
+
"learning_rate": 1.8324786324786325e-06,
|
| 1335 |
+
"loss": 0.0,
|
| 1336 |
+
"step": 18600
|
| 1337 |
+
},
|
| 1338 |
+
{
|
| 1339 |
+
"epoch": 2.73992673992674,
|
| 1340 |
+
"grad_norm": 6.854772072983906e-05,
|
| 1341 |
+
"learning_rate": 1.7347985347985349e-06,
|
| 1342 |
+
"loss": 0.0,
|
| 1343 |
+
"step": 18700
|
| 1344 |
+
},
|
| 1345 |
+
{
|
| 1346 |
+
"epoch": 2.7545787545787546,
|
| 1347 |
+
"grad_norm": 5.984567178529687e-05,
|
| 1348 |
+
"learning_rate": 1.6371184371184373e-06,
|
| 1349 |
+
"loss": 0.0,
|
| 1350 |
+
"step": 18800
|
| 1351 |
+
},
|
| 1352 |
+
{
|
| 1353 |
+
"epoch": 2.769230769230769,
|
| 1354 |
+
"grad_norm": 6.427949119824916e-05,
|
| 1355 |
+
"learning_rate": 1.5394383394383396e-06,
|
| 1356 |
+
"loss": 0.0001,
|
| 1357 |
+
"step": 18900
|
| 1358 |
+
},
|
| 1359 |
+
{
|
| 1360 |
+
"epoch": 2.7838827838827838,
|
| 1361 |
+
"grad_norm": 2.7727375709218904e-05,
|
| 1362 |
+
"learning_rate": 1.4417582417582418e-06,
|
| 1363 |
+
"loss": 0.0,
|
| 1364 |
+
"step": 19000
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 2.7985347985347984,
|
| 1368 |
+
"grad_norm": 3.604817175073549e-05,
|
| 1369 |
+
"learning_rate": 1.3440781440781442e-06,
|
| 1370 |
+
"loss": 0.0,
|
| 1371 |
+
"step": 19100
|
| 1372 |
+
},
|
| 1373 |
+
{
|
| 1374 |
+
"epoch": 2.813186813186813,
|
| 1375 |
+
"grad_norm": 9.17447468964383e-05,
|
| 1376 |
+
"learning_rate": 1.2463980463980466e-06,
|
| 1377 |
+
"loss": 0.0,
|
| 1378 |
+
"step": 19200
|
| 1379 |
+
},
|
| 1380 |
+
{
|
| 1381 |
+
"epoch": 2.8278388278388276,
|
| 1382 |
+
"grad_norm": 3.2978157832985744e-05,
|
| 1383 |
+
"learning_rate": 1.1487179487179487e-06,
|
| 1384 |
+
"loss": 0.0,
|
| 1385 |
+
"step": 19300
|
| 1386 |
+
},
|
| 1387 |
+
{
|
| 1388 |
+
"epoch": 2.8424908424908426,
|
| 1389 |
+
"grad_norm": 4.4227072066860273e-05,
|
| 1390 |
+
"learning_rate": 1.051037851037851e-06,
|
| 1391 |
+
"loss": 0.0,
|
| 1392 |
+
"step": 19400
|
| 1393 |
+
},
|
| 1394 |
+
{
|
| 1395 |
+
"epoch": 2.857142857142857,
|
| 1396 |
+
"grad_norm": 3.573105641407892e-05,
|
| 1397 |
+
"learning_rate": 9.533577533577535e-07,
|
| 1398 |
+
"loss": 0.0,
|
| 1399 |
+
"step": 19500
|
| 1400 |
+
},
|
| 1401 |
+
{
|
| 1402 |
+
"epoch": 2.871794871794872,
|
| 1403 |
+
"grad_norm": 5.604741090792231e-05,
|
| 1404 |
+
"learning_rate": 8.556776556776559e-07,
|
| 1405 |
+
"loss": 0.0,
|
| 1406 |
+
"step": 19600
|
| 1407 |
+
},
|
| 1408 |
+
{
|
| 1409 |
+
"epoch": 2.8864468864468864,
|
| 1410 |
+
"grad_norm": 6.343449058476835e-05,
|
| 1411 |
+
"learning_rate": 7.57997557997558e-07,
|
| 1412 |
+
"loss": 0.0,
|
| 1413 |
+
"step": 19700
|
| 1414 |
+
},
|
| 1415 |
+
{
|
| 1416 |
+
"epoch": 2.901098901098901,
|
| 1417 |
+
"grad_norm": 2.3594573576701805e-05,
|
| 1418 |
+
"learning_rate": 6.603174603174604e-07,
|
| 1419 |
+
"loss": 0.0,
|
| 1420 |
+
"step": 19800
|
| 1421 |
+
},
|
| 1422 |
+
{
|
| 1423 |
+
"epoch": 2.9157509157509156,
|
| 1424 |
+
"grad_norm": 4.237998291500844e-05,
|
| 1425 |
+
"learning_rate": 5.626373626373627e-07,
|
| 1426 |
+
"loss": 0.0,
|
| 1427 |
+
"step": 19900
|
| 1428 |
+
},
|
| 1429 |
+
{
|
| 1430 |
+
"epoch": 2.9304029304029307,
|
| 1431 |
+
"grad_norm": 4.4342101318761706e-05,
|
| 1432 |
+
"learning_rate": 4.64957264957265e-07,
|
| 1433 |
+
"loss": 0.0,
|
| 1434 |
+
"step": 20000
|
| 1435 |
+
},
|
| 1436 |
+
{
|
| 1437 |
+
"epoch": 2.9450549450549453,
|
| 1438 |
+
"grad_norm": 4.6577501052524894e-05,
|
| 1439 |
+
"learning_rate": 3.672771672771673e-07,
|
| 1440 |
+
"loss": 0.0,
|
| 1441 |
+
"step": 20100
|
| 1442 |
+
},
|
| 1443 |
+
{
|
| 1444 |
+
"epoch": 2.95970695970696,
|
| 1445 |
+
"grad_norm": 3.366880628163926e-05,
|
| 1446 |
+
"learning_rate": 2.695970695970696e-07,
|
| 1447 |
+
"loss": 0.0,
|
| 1448 |
+
"step": 20200
|
| 1449 |
+
},
|
| 1450 |
+
{
|
| 1451 |
+
"epoch": 2.9743589743589745,
|
| 1452 |
+
"grad_norm": 4.93158950121142e-05,
|
| 1453 |
+
"learning_rate": 1.7191697191697192e-07,
|
| 1454 |
+
"loss": 0.0,
|
| 1455 |
+
"step": 20300
|
| 1456 |
+
},
|
| 1457 |
+
{
|
| 1458 |
+
"epoch": 2.989010989010989,
|
| 1459 |
+
"grad_norm": 3.799773912760429e-05,
|
| 1460 |
+
"learning_rate": 7.423687423687424e-08,
|
| 1461 |
+
"loss": 0.0,
|
| 1462 |
+
"step": 20400
|
| 1463 |
+
},
|
| 1464 |
+
{
|
| 1465 |
+
"epoch": 3.0,
|
| 1466 |
+
"eval_accuracy": 1.0,
|
| 1467 |
+
"eval_f1": 1.0,
|
| 1468 |
+
"eval_loss": 1.2949521988048218e-06,
|
| 1469 |
+
"eval_precision": 1.0,
|
| 1470 |
+
"eval_recall": 1.0,
|
| 1471 |
+
"eval_runtime": 266.3604,
|
| 1472 |
+
"eval_samples_per_second": 102.493,
|
| 1473 |
+
"eval_steps_per_second": 6.409,
|
| 1474 |
+
"step": 20475
|
| 1475 |
+
}
|
| 1476 |
+
],
|
| 1477 |
+
"logging_steps": 100,
|
| 1478 |
+
"max_steps": 20475,
|
| 1479 |
+
"num_input_tokens_seen": 0,
|
| 1480 |
+
"num_train_epochs": 3,
|
| 1481 |
+
"save_steps": 500,
|
| 1482 |
+
"stateful_callbacks": {
|
| 1483 |
+
"TrainerControl": {
|
| 1484 |
+
"args": {
|
| 1485 |
+
"should_epoch_stop": false,
|
| 1486 |
+
"should_evaluate": false,
|
| 1487 |
+
"should_log": false,
|
| 1488 |
+
"should_save": true,
|
| 1489 |
+
"should_training_stop": true
|
| 1490 |
+
},
|
| 1491 |
+
"attributes": {}
|
| 1492 |
+
}
|
| 1493 |
+
},
|
| 1494 |
+
"total_flos": 2.5970716030629024e+16,
|
| 1495 |
+
"train_batch_size": 16,
|
| 1496 |
+
"trial_name": null,
|
| 1497 |
+
"trial_params": null
|
| 1498 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fdc84b0e9c686c061974cd287d3c24839f53f2c3972ac34b3199981e7073d4c6
|
| 3 |
+
size 5368
|
vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|