Commit
·
a9f0fc3
1
Parent(s):
9e35bf4
reset
Browse files- README.md +0 -3
- added_tokens.json +0 -3
- checkpoint-6562/added_tokens.json +0 -3
- checkpoint-6562/config.json +0 -69
- checkpoint-6562/special_tokens_map.json +0 -15
- checkpoint-6562/tokenizer.json +0 -0
- checkpoint-6562/tokenizer_config.json +0 -59
- checkpoint-6562/trainer_state.json +0 -951
- checkpoint-6948/added_tokens.json +0 -3
- checkpoint-6948/config.json +0 -69
- checkpoint-6948/special_tokens_map.json +0 -15
- checkpoint-6948/tokenizer.json +0 -0
- checkpoint-6948/tokenizer_config.json +0 -59
- checkpoint-6948/trainer_state.json +0 -1000
- checkpoint-7334/added_tokens.json +0 -3
- checkpoint-7334/config.json +0 -69
- checkpoint-7334/special_tokens_map.json +0 -15
- checkpoint-7334/tokenizer.json +0 -0
- checkpoint-7334/tokenizer_config.json +0 -59
- checkpoint-7334/trainer_state.json +0 -1056
- config.json +0 -69
- special_tokens_map.json +0 -15
- tokenizer.json +0 -0
- tokenizer_config.json +0 -59
README.md
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
---
|
| 2 |
-
license: mit
|
| 3 |
-
---
|
|
|
|
|
|
|
|
|
|
|
|
added_tokens.json
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"[MASK]": 128000
|
| 3 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
checkpoint-6562/added_tokens.json
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"[MASK]": 128000
|
| 3 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
checkpoint-6562/config.json
DELETED
|
@@ -1,69 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"architectures": [
|
| 3 |
-
"DebertaV2ForTokenClassification"
|
| 4 |
-
],
|
| 5 |
-
"attention_probs_dropout_prob": 0.1,
|
| 6 |
-
"hidden_act": "gelu",
|
| 7 |
-
"hidden_dropout_prob": 0.1,
|
| 8 |
-
"hidden_size": 768,
|
| 9 |
-
"id2label": {
|
| 10 |
-
"0": "O",
|
| 11 |
-
"1": "B-CHA",
|
| 12 |
-
"2": "I-CHA",
|
| 13 |
-
"3": "B-LOC",
|
| 14 |
-
"4": "I-LOC",
|
| 15 |
-
"5": "B-FAC",
|
| 16 |
-
"6": "I-FAC",
|
| 17 |
-
"7": "B-OBJ",
|
| 18 |
-
"8": "I-OBJ",
|
| 19 |
-
"9": "B-EVT",
|
| 20 |
-
"10": "I-EVT",
|
| 21 |
-
"11": "B-ORG",
|
| 22 |
-
"12": "I-ORG",
|
| 23 |
-
"13": "B-MISC",
|
| 24 |
-
"14": "I-MISC"
|
| 25 |
-
},
|
| 26 |
-
"initializer_range": 0.02,
|
| 27 |
-
"intermediate_size": 3072,
|
| 28 |
-
"label2id": {
|
| 29 |
-
"B-CHA": 1,
|
| 30 |
-
"B-EVT": 9,
|
| 31 |
-
"B-FAC": 5,
|
| 32 |
-
"B-LOC": 3,
|
| 33 |
-
"B-MISC": 13,
|
| 34 |
-
"B-OBJ": 7,
|
| 35 |
-
"B-ORG": 11,
|
| 36 |
-
"I-CHA": 2,
|
| 37 |
-
"I-EVT": 10,
|
| 38 |
-
"I-FAC": 6,
|
| 39 |
-
"I-LOC": 4,
|
| 40 |
-
"I-MISC": 14,
|
| 41 |
-
"I-OBJ": 8,
|
| 42 |
-
"I-ORG": 12,
|
| 43 |
-
"O": 0
|
| 44 |
-
},
|
| 45 |
-
"layer_norm_eps": 1e-07,
|
| 46 |
-
"legacy": true,
|
| 47 |
-
"max_position_embeddings": 512,
|
| 48 |
-
"max_relative_positions": -1,
|
| 49 |
-
"model_type": "deberta-v2",
|
| 50 |
-
"norm_rel_ebd": "layer_norm",
|
| 51 |
-
"num_attention_heads": 12,
|
| 52 |
-
"num_hidden_layers": 12,
|
| 53 |
-
"pad_token_id": 0,
|
| 54 |
-
"pooler_dropout": 0,
|
| 55 |
-
"pooler_hidden_act": "gelu",
|
| 56 |
-
"pooler_hidden_size": 768,
|
| 57 |
-
"pos_att_type": [
|
| 58 |
-
"p2c",
|
| 59 |
-
"c2p"
|
| 60 |
-
],
|
| 61 |
-
"position_biased_input": false,
|
| 62 |
-
"position_buckets": 256,
|
| 63 |
-
"relative_attention": true,
|
| 64 |
-
"share_att_key": true,
|
| 65 |
-
"torch_dtype": "float32",
|
| 66 |
-
"transformers_version": "4.55.4",
|
| 67 |
-
"type_vocab_size": 0,
|
| 68 |
-
"vocab_size": 128100
|
| 69 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
checkpoint-6562/special_tokens_map.json
DELETED
|
@@ -1,15 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"bos_token": "[CLS]",
|
| 3 |
-
"cls_token": "[CLS]",
|
| 4 |
-
"eos_token": "[SEP]",
|
| 5 |
-
"mask_token": "[MASK]",
|
| 6 |
-
"pad_token": "[PAD]",
|
| 7 |
-
"sep_token": "[SEP]",
|
| 8 |
-
"unk_token": {
|
| 9 |
-
"content": "[UNK]",
|
| 10 |
-
"lstrip": false,
|
| 11 |
-
"normalized": true,
|
| 12 |
-
"rstrip": false,
|
| 13 |
-
"single_word": false
|
| 14 |
-
}
|
| 15 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
checkpoint-6562/tokenizer.json
DELETED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-6562/tokenizer_config.json
DELETED
|
@@ -1,59 +0,0 @@
|
|
| 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": "[CLS]",
|
| 13 |
-
"lstrip": false,
|
| 14 |
-
"normalized": false,
|
| 15 |
-
"rstrip": false,
|
| 16 |
-
"single_word": false,
|
| 17 |
-
"special": true
|
| 18 |
-
},
|
| 19 |
-
"2": {
|
| 20 |
-
"content": "[SEP]",
|
| 21 |
-
"lstrip": false,
|
| 22 |
-
"normalized": false,
|
| 23 |
-
"rstrip": false,
|
| 24 |
-
"single_word": false,
|
| 25 |
-
"special": true
|
| 26 |
-
},
|
| 27 |
-
"3": {
|
| 28 |
-
"content": "[UNK]",
|
| 29 |
-
"lstrip": false,
|
| 30 |
-
"normalized": true,
|
| 31 |
-
"rstrip": false,
|
| 32 |
-
"single_word": false,
|
| 33 |
-
"special": true
|
| 34 |
-
},
|
| 35 |
-
"128000": {
|
| 36 |
-
"content": "[MASK]",
|
| 37 |
-
"lstrip": false,
|
| 38 |
-
"normalized": false,
|
| 39 |
-
"rstrip": false,
|
| 40 |
-
"single_word": false,
|
| 41 |
-
"special": true
|
| 42 |
-
}
|
| 43 |
-
},
|
| 44 |
-
"bos_token": "[CLS]",
|
| 45 |
-
"clean_up_tokenization_spaces": false,
|
| 46 |
-
"cls_token": "[CLS]",
|
| 47 |
-
"do_lower_case": false,
|
| 48 |
-
"eos_token": "[SEP]",
|
| 49 |
-
"extra_special_tokens": {},
|
| 50 |
-
"mask_token": "[MASK]",
|
| 51 |
-
"model_max_length": 1000000000000000019884624838656,
|
| 52 |
-
"pad_token": "[PAD]",
|
| 53 |
-
"sep_token": "[SEP]",
|
| 54 |
-
"sp_model_kwargs": {},
|
| 55 |
-
"split_by_punct": false,
|
| 56 |
-
"tokenizer_class": "DebertaV2Tokenizer",
|
| 57 |
-
"unk_token": "[UNK]",
|
| 58 |
-
"vocab_type": "spm"
|
| 59 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
checkpoint-6562/trainer_state.json
DELETED
|
@@ -1,951 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"best_global_step": null,
|
| 3 |
-
"best_metric": null,
|
| 4 |
-
"best_model_checkpoint": null,
|
| 5 |
-
"epoch": 0.8520971302428256,
|
| 6 |
-
"eval_steps": 500,
|
| 7 |
-
"global_step": 6562,
|
| 8 |
-
"is_hyper_param_search": false,
|
| 9 |
-
"is_local_process_zero": true,
|
| 10 |
-
"is_world_process_zero": true,
|
| 11 |
-
"log_history": [
|
| 12 |
-
{
|
| 13 |
-
"epoch": 0.006492663290481756,
|
| 14 |
-
"grad_norm": 23.6250057220459,
|
| 15 |
-
"learning_rate": 1.9600000000000003e-06,
|
| 16 |
-
"loss": 3.3492,
|
| 17 |
-
"step": 50
|
| 18 |
-
},
|
| 19 |
-
{
|
| 20 |
-
"epoch": 0.012985326580963512,
|
| 21 |
-
"grad_norm": 1.5745655298233032,
|
| 22 |
-
"learning_rate": 3.96e-06,
|
| 23 |
-
"loss": 1.3837,
|
| 24 |
-
"step": 100
|
| 25 |
-
},
|
| 26 |
-
{
|
| 27 |
-
"epoch": 0.019477989871445268,
|
| 28 |
-
"grad_norm": 1.3241935968399048,
|
| 29 |
-
"learning_rate": 5.9600000000000005e-06,
|
| 30 |
-
"loss": 0.6673,
|
| 31 |
-
"step": 150
|
| 32 |
-
},
|
| 33 |
-
{
|
| 34 |
-
"epoch": 0.025970653161927024,
|
| 35 |
-
"grad_norm": 1.068557620048523,
|
| 36 |
-
"learning_rate": 7.960000000000002e-06,
|
| 37 |
-
"loss": 0.4426,
|
| 38 |
-
"step": 200
|
| 39 |
-
},
|
| 40 |
-
{
|
| 41 |
-
"epoch": 0.032463316452408776,
|
| 42 |
-
"grad_norm": 1.2953921556472778,
|
| 43 |
-
"learning_rate": 9.960000000000001e-06,
|
| 44 |
-
"loss": 0.363,
|
| 45 |
-
"step": 250
|
| 46 |
-
},
|
| 47 |
-
{
|
| 48 |
-
"epoch": 0.038955979742890535,
|
| 49 |
-
"grad_norm": 1.4753903150558472,
|
| 50 |
-
"learning_rate": 1.196e-05,
|
| 51 |
-
"loss": 0.3397,
|
| 52 |
-
"step": 300
|
| 53 |
-
},
|
| 54 |
-
{
|
| 55 |
-
"epoch": 0.04544864303337229,
|
| 56 |
-
"grad_norm": 1.245597243309021,
|
| 57 |
-
"learning_rate": 1.396e-05,
|
| 58 |
-
"loss": 0.3125,
|
| 59 |
-
"step": 350
|
| 60 |
-
},
|
| 61 |
-
{
|
| 62 |
-
"epoch": 0.05194130632385405,
|
| 63 |
-
"grad_norm": 1.3421887159347534,
|
| 64 |
-
"learning_rate": 1.5960000000000003e-05,
|
| 65 |
-
"loss": 0.3086,
|
| 66 |
-
"step": 400
|
| 67 |
-
},
|
| 68 |
-
{
|
| 69 |
-
"epoch": 0.0584339696143358,
|
| 70 |
-
"grad_norm": 1.7852529287338257,
|
| 71 |
-
"learning_rate": 1.796e-05,
|
| 72 |
-
"loss": 0.2883,
|
| 73 |
-
"step": 450
|
| 74 |
-
},
|
| 75 |
-
{
|
| 76 |
-
"epoch": 0.06492663290481755,
|
| 77 |
-
"grad_norm": 1.086064100265503,
|
| 78 |
-
"learning_rate": 1.9960000000000002e-05,
|
| 79 |
-
"loss": 0.2947,
|
| 80 |
-
"step": 500
|
| 81 |
-
},
|
| 82 |
-
{
|
| 83 |
-
"epoch": 0.07141929619529931,
|
| 84 |
-
"grad_norm": 0.8753414750099182,
|
| 85 |
-
"learning_rate": 1.9974213919221156e-05,
|
| 86 |
-
"loss": 0.2955,
|
| 87 |
-
"step": 550
|
| 88 |
-
},
|
| 89 |
-
{
|
| 90 |
-
"epoch": 0.07791195948578107,
|
| 91 |
-
"grad_norm": 0.7875685095787048,
|
| 92 |
-
"learning_rate": 1.9947901591895804e-05,
|
| 93 |
-
"loss": 0.2773,
|
| 94 |
-
"step": 600
|
| 95 |
-
},
|
| 96 |
-
{
|
| 97 |
-
"epoch": 0.08440462277626282,
|
| 98 |
-
"grad_norm": 1.1609506607055664,
|
| 99 |
-
"learning_rate": 1.9921589264570452e-05,
|
| 100 |
-
"loss": 0.2696,
|
| 101 |
-
"step": 650
|
| 102 |
-
},
|
| 103 |
-
{
|
| 104 |
-
"epoch": 0.09089728606674458,
|
| 105 |
-
"grad_norm": 0.9162603616714478,
|
| 106 |
-
"learning_rate": 1.98952769372451e-05,
|
| 107 |
-
"loss": 0.2667,
|
| 108 |
-
"step": 700
|
| 109 |
-
},
|
| 110 |
-
{
|
| 111 |
-
"epoch": 0.09738994935722634,
|
| 112 |
-
"grad_norm": 1.0398370027542114,
|
| 113 |
-
"learning_rate": 1.986896460991975e-05,
|
| 114 |
-
"loss": 0.277,
|
| 115 |
-
"step": 750
|
| 116 |
-
},
|
| 117 |
-
{
|
| 118 |
-
"epoch": 0.1038826126477081,
|
| 119 |
-
"grad_norm": 0.9404798150062561,
|
| 120 |
-
"learning_rate": 1.9842652282594397e-05,
|
| 121 |
-
"loss": 0.2584,
|
| 122 |
-
"step": 800
|
| 123 |
-
},
|
| 124 |
-
{
|
| 125 |
-
"epoch": 0.11037527593818984,
|
| 126 |
-
"grad_norm": 1.0353909730911255,
|
| 127 |
-
"learning_rate": 1.9816339955269045e-05,
|
| 128 |
-
"loss": 0.2533,
|
| 129 |
-
"step": 850
|
| 130 |
-
},
|
| 131 |
-
{
|
| 132 |
-
"epoch": 0.1168679392286716,
|
| 133 |
-
"grad_norm": 1.485606074333191,
|
| 134 |
-
"learning_rate": 1.9790027627943694e-05,
|
| 135 |
-
"loss": 0.2506,
|
| 136 |
-
"step": 900
|
| 137 |
-
},
|
| 138 |
-
{
|
| 139 |
-
"epoch": 0.12336060251915336,
|
| 140 |
-
"grad_norm": 0.9874151945114136,
|
| 141 |
-
"learning_rate": 1.9763715300618342e-05,
|
| 142 |
-
"loss": 0.2515,
|
| 143 |
-
"step": 950
|
| 144 |
-
},
|
| 145 |
-
{
|
| 146 |
-
"epoch": 0.1298532658096351,
|
| 147 |
-
"grad_norm": 1.036007285118103,
|
| 148 |
-
"learning_rate": 1.973740297329299e-05,
|
| 149 |
-
"loss": 0.244,
|
| 150 |
-
"step": 1000
|
| 151 |
-
},
|
| 152 |
-
{
|
| 153 |
-
"epoch": 0.13634592910011686,
|
| 154 |
-
"grad_norm": 0.946956217288971,
|
| 155 |
-
"learning_rate": 1.9711090645967635e-05,
|
| 156 |
-
"loss": 0.2518,
|
| 157 |
-
"step": 1050
|
| 158 |
-
},
|
| 159 |
-
{
|
| 160 |
-
"epoch": 0.14283859239059862,
|
| 161 |
-
"grad_norm": 0.9096735119819641,
|
| 162 |
-
"learning_rate": 1.9684778318642287e-05,
|
| 163 |
-
"loss": 0.2543,
|
| 164 |
-
"step": 1100
|
| 165 |
-
},
|
| 166 |
-
{
|
| 167 |
-
"epoch": 0.14933125568108038,
|
| 168 |
-
"grad_norm": 0.9371875524520874,
|
| 169 |
-
"learning_rate": 1.965846599131693e-05,
|
| 170 |
-
"loss": 0.2605,
|
| 171 |
-
"step": 1150
|
| 172 |
-
},
|
| 173 |
-
{
|
| 174 |
-
"epoch": 0.15582391897156214,
|
| 175 |
-
"grad_norm": 0.8072580695152283,
|
| 176 |
-
"learning_rate": 1.9632153663991583e-05,
|
| 177 |
-
"loss": 0.2423,
|
| 178 |
-
"step": 1200
|
| 179 |
-
},
|
| 180 |
-
{
|
| 181 |
-
"epoch": 0.1623165822620439,
|
| 182 |
-
"grad_norm": 0.6907565593719482,
|
| 183 |
-
"learning_rate": 1.9605841336666228e-05,
|
| 184 |
-
"loss": 0.2315,
|
| 185 |
-
"step": 1250
|
| 186 |
-
},
|
| 187 |
-
{
|
| 188 |
-
"epoch": 0.16880924555252563,
|
| 189 |
-
"grad_norm": 0.6676374673843384,
|
| 190 |
-
"learning_rate": 1.957952900934088e-05,
|
| 191 |
-
"loss": 0.2363,
|
| 192 |
-
"step": 1300
|
| 193 |
-
},
|
| 194 |
-
{
|
| 195 |
-
"epoch": 0.1753019088430074,
|
| 196 |
-
"grad_norm": 0.7281391024589539,
|
| 197 |
-
"learning_rate": 1.9553216682015525e-05,
|
| 198 |
-
"loss": 0.2399,
|
| 199 |
-
"step": 1350
|
| 200 |
-
},
|
| 201 |
-
{
|
| 202 |
-
"epoch": 0.18179457213348915,
|
| 203 |
-
"grad_norm": 0.8105588555335999,
|
| 204 |
-
"learning_rate": 1.9526904354690177e-05,
|
| 205 |
-
"loss": 0.2448,
|
| 206 |
-
"step": 1400
|
| 207 |
-
},
|
| 208 |
-
{
|
| 209 |
-
"epoch": 0.1882872354239709,
|
| 210 |
-
"grad_norm": 0.7466333508491516,
|
| 211 |
-
"learning_rate": 1.950059202736482e-05,
|
| 212 |
-
"loss": 0.2388,
|
| 213 |
-
"step": 1450
|
| 214 |
-
},
|
| 215 |
-
{
|
| 216 |
-
"epoch": 0.19477989871445267,
|
| 217 |
-
"grad_norm": 0.5949985384941101,
|
| 218 |
-
"learning_rate": 1.947427970003947e-05,
|
| 219 |
-
"loss": 0.233,
|
| 220 |
-
"step": 1500
|
| 221 |
-
},
|
| 222 |
-
{
|
| 223 |
-
"epoch": 0.20127256200493443,
|
| 224 |
-
"grad_norm": 0.7439960241317749,
|
| 225 |
-
"learning_rate": 1.9447967372714118e-05,
|
| 226 |
-
"loss": 0.2448,
|
| 227 |
-
"step": 1550
|
| 228 |
-
},
|
| 229 |
-
{
|
| 230 |
-
"epoch": 0.2077652252954162,
|
| 231 |
-
"grad_norm": 0.9508784413337708,
|
| 232 |
-
"learning_rate": 1.9421655045388766e-05,
|
| 233 |
-
"loss": 0.2353,
|
| 234 |
-
"step": 1600
|
| 235 |
-
},
|
| 236 |
-
{
|
| 237 |
-
"epoch": 0.21425788858589795,
|
| 238 |
-
"grad_norm": 0.8785332441329956,
|
| 239 |
-
"learning_rate": 1.9395342718063415e-05,
|
| 240 |
-
"loss": 0.2249,
|
| 241 |
-
"step": 1650
|
| 242 |
-
},
|
| 243 |
-
{
|
| 244 |
-
"epoch": 0.22075055187637968,
|
| 245 |
-
"grad_norm": 0.7568134069442749,
|
| 246 |
-
"learning_rate": 1.9369030390738063e-05,
|
| 247 |
-
"loss": 0.2307,
|
| 248 |
-
"step": 1700
|
| 249 |
-
},
|
| 250 |
-
{
|
| 251 |
-
"epoch": 0.22724321516686144,
|
| 252 |
-
"grad_norm": 0.6832641363143921,
|
| 253 |
-
"learning_rate": 1.934271806341271e-05,
|
| 254 |
-
"loss": 0.2263,
|
| 255 |
-
"step": 1750
|
| 256 |
-
},
|
| 257 |
-
{
|
| 258 |
-
"epoch": 0.2337358784573432,
|
| 259 |
-
"grad_norm": 0.9721575379371643,
|
| 260 |
-
"learning_rate": 1.931640573608736e-05,
|
| 261 |
-
"loss": 0.2287,
|
| 262 |
-
"step": 1800
|
| 263 |
-
},
|
| 264 |
-
{
|
| 265 |
-
"epoch": 0.24022854174782496,
|
| 266 |
-
"grad_norm": 0.620299220085144,
|
| 267 |
-
"learning_rate": 1.9290093408762008e-05,
|
| 268 |
-
"loss": 0.226,
|
| 269 |
-
"step": 1850
|
| 270 |
-
},
|
| 271 |
-
{
|
| 272 |
-
"epoch": 0.24672120503830672,
|
| 273 |
-
"grad_norm": 0.6656680703163147,
|
| 274 |
-
"learning_rate": 1.9263781081436656e-05,
|
| 275 |
-
"loss": 0.2223,
|
| 276 |
-
"step": 1900
|
| 277 |
-
},
|
| 278 |
-
{
|
| 279 |
-
"epoch": 0.25321386832878845,
|
| 280 |
-
"grad_norm": 0.8536450266838074,
|
| 281 |
-
"learning_rate": 1.9237468754111304e-05,
|
| 282 |
-
"loss": 0.2332,
|
| 283 |
-
"step": 1950
|
| 284 |
-
},
|
| 285 |
-
{
|
| 286 |
-
"epoch": 0.2597065316192702,
|
| 287 |
-
"grad_norm": 0.9754176139831543,
|
| 288 |
-
"learning_rate": 1.921115642678595e-05,
|
| 289 |
-
"loss": 0.2211,
|
| 290 |
-
"step": 2000
|
| 291 |
-
},
|
| 292 |
-
{
|
| 293 |
-
"epoch": 0.26619919490975197,
|
| 294 |
-
"grad_norm": 0.6166796684265137,
|
| 295 |
-
"learning_rate": 1.9184844099460598e-05,
|
| 296 |
-
"loss": 0.2201,
|
| 297 |
-
"step": 2050
|
| 298 |
-
},
|
| 299 |
-
{
|
| 300 |
-
"epoch": 0.2726918582002337,
|
| 301 |
-
"grad_norm": 0.8182518482208252,
|
| 302 |
-
"learning_rate": 1.9158531772135246e-05,
|
| 303 |
-
"loss": 0.2213,
|
| 304 |
-
"step": 2100
|
| 305 |
-
},
|
| 306 |
-
{
|
| 307 |
-
"epoch": 0.2791845214907155,
|
| 308 |
-
"grad_norm": 0.8245829939842224,
|
| 309 |
-
"learning_rate": 1.9132219444809894e-05,
|
| 310 |
-
"loss": 0.2208,
|
| 311 |
-
"step": 2150
|
| 312 |
-
},
|
| 313 |
-
{
|
| 314 |
-
"epoch": 0.28567718478119725,
|
| 315 |
-
"grad_norm": 0.6423109173774719,
|
| 316 |
-
"learning_rate": 1.9105907117484542e-05,
|
| 317 |
-
"loss": 0.2308,
|
| 318 |
-
"step": 2200
|
| 319 |
-
},
|
| 320 |
-
{
|
| 321 |
-
"epoch": 0.292169848071679,
|
| 322 |
-
"grad_norm": 0.7164100408554077,
|
| 323 |
-
"learning_rate": 1.907959479015919e-05,
|
| 324 |
-
"loss": 0.2193,
|
| 325 |
-
"step": 2250
|
| 326 |
-
},
|
| 327 |
-
{
|
| 328 |
-
"epoch": 0.29866251136216077,
|
| 329 |
-
"grad_norm": 0.7910536527633667,
|
| 330 |
-
"learning_rate": 1.905328246283384e-05,
|
| 331 |
-
"loss": 0.2292,
|
| 332 |
-
"step": 2300
|
| 333 |
-
},
|
| 334 |
-
{
|
| 335 |
-
"epoch": 0.3051551746526425,
|
| 336 |
-
"grad_norm": 0.9699934720993042,
|
| 337 |
-
"learning_rate": 1.9026970135508487e-05,
|
| 338 |
-
"loss": 0.2175,
|
| 339 |
-
"step": 2350
|
| 340 |
-
},
|
| 341 |
-
{
|
| 342 |
-
"epoch": 0.3116478379431243,
|
| 343 |
-
"grad_norm": 1.0665416717529297,
|
| 344 |
-
"learning_rate": 1.9000657808183136e-05,
|
| 345 |
-
"loss": 0.2252,
|
| 346 |
-
"step": 2400
|
| 347 |
-
},
|
| 348 |
-
{
|
| 349 |
-
"epoch": 0.31814050123360604,
|
| 350 |
-
"grad_norm": 0.9915699362754822,
|
| 351 |
-
"learning_rate": 1.8974345480857784e-05,
|
| 352 |
-
"loss": 0.215,
|
| 353 |
-
"step": 2450
|
| 354 |
-
},
|
| 355 |
-
{
|
| 356 |
-
"epoch": 0.3246331645240878,
|
| 357 |
-
"grad_norm": 0.7021859288215637,
|
| 358 |
-
"learning_rate": 1.8948033153532432e-05,
|
| 359 |
-
"loss": 0.2123,
|
| 360 |
-
"step": 2500
|
| 361 |
-
},
|
| 362 |
-
{
|
| 363 |
-
"epoch": 0.33112582781456956,
|
| 364 |
-
"grad_norm": 0.6402600407600403,
|
| 365 |
-
"learning_rate": 1.892172082620708e-05,
|
| 366 |
-
"loss": 0.2176,
|
| 367 |
-
"step": 2550
|
| 368 |
-
},
|
| 369 |
-
{
|
| 370 |
-
"epoch": 0.33761849110505127,
|
| 371 |
-
"grad_norm": 0.9925078749656677,
|
| 372 |
-
"learning_rate": 1.8895408498881725e-05,
|
| 373 |
-
"loss": 0.223,
|
| 374 |
-
"step": 2600
|
| 375 |
-
},
|
| 376 |
-
{
|
| 377 |
-
"epoch": 0.344111154395533,
|
| 378 |
-
"grad_norm": 0.5188687443733215,
|
| 379 |
-
"learning_rate": 1.8869096171556377e-05,
|
| 380 |
-
"loss": 0.2174,
|
| 381 |
-
"step": 2650
|
| 382 |
-
},
|
| 383 |
-
{
|
| 384 |
-
"epoch": 0.3506038176860148,
|
| 385 |
-
"grad_norm": 0.7427687644958496,
|
| 386 |
-
"learning_rate": 1.8842783844231022e-05,
|
| 387 |
-
"loss": 0.2181,
|
| 388 |
-
"step": 2700
|
| 389 |
-
},
|
| 390 |
-
{
|
| 391 |
-
"epoch": 0.35709648097649654,
|
| 392 |
-
"grad_norm": 0.6055501103401184,
|
| 393 |
-
"learning_rate": 1.8816471516905674e-05,
|
| 394 |
-
"loss": 0.2097,
|
| 395 |
-
"step": 2750
|
| 396 |
-
},
|
| 397 |
-
{
|
| 398 |
-
"epoch": 0.3635891442669783,
|
| 399 |
-
"grad_norm": 0.6529932022094727,
|
| 400 |
-
"learning_rate": 1.879015918958032e-05,
|
| 401 |
-
"loss": 0.2133,
|
| 402 |
-
"step": 2800
|
| 403 |
-
},
|
| 404 |
-
{
|
| 405 |
-
"epoch": 0.37008180755746006,
|
| 406 |
-
"grad_norm": 0.9499914646148682,
|
| 407 |
-
"learning_rate": 1.8763846862254967e-05,
|
| 408 |
-
"loss": 0.2189,
|
| 409 |
-
"step": 2850
|
| 410 |
-
},
|
| 411 |
-
{
|
| 412 |
-
"epoch": 0.3765744708479418,
|
| 413 |
-
"grad_norm": 0.8217771053314209,
|
| 414 |
-
"learning_rate": 1.8737534534929615e-05,
|
| 415 |
-
"loss": 0.2172,
|
| 416 |
-
"step": 2900
|
| 417 |
-
},
|
| 418 |
-
{
|
| 419 |
-
"epoch": 0.3830671341384236,
|
| 420 |
-
"grad_norm": 1.031895399093628,
|
| 421 |
-
"learning_rate": 1.8711222207604263e-05,
|
| 422 |
-
"loss": 0.2171,
|
| 423 |
-
"step": 2950
|
| 424 |
-
},
|
| 425 |
-
{
|
| 426 |
-
"epoch": 0.38955979742890534,
|
| 427 |
-
"grad_norm": 0.7179540395736694,
|
| 428 |
-
"learning_rate": 1.8684909880278912e-05,
|
| 429 |
-
"loss": 0.2179,
|
| 430 |
-
"step": 3000
|
| 431 |
-
},
|
| 432 |
-
{
|
| 433 |
-
"epoch": 0.3960524607193871,
|
| 434 |
-
"grad_norm": 0.8115535974502563,
|
| 435 |
-
"learning_rate": 1.865859755295356e-05,
|
| 436 |
-
"loss": 0.2131,
|
| 437 |
-
"step": 3050
|
| 438 |
-
},
|
| 439 |
-
{
|
| 440 |
-
"epoch": 0.40254512400986886,
|
| 441 |
-
"grad_norm": 0.7659513354301453,
|
| 442 |
-
"learning_rate": 1.863228522562821e-05,
|
| 443 |
-
"loss": 0.2189,
|
| 444 |
-
"step": 3100
|
| 445 |
-
},
|
| 446 |
-
{
|
| 447 |
-
"epoch": 0.4090377873003506,
|
| 448 |
-
"grad_norm": 0.8645811080932617,
|
| 449 |
-
"learning_rate": 1.8605972898302857e-05,
|
| 450 |
-
"loss": 0.2453,
|
| 451 |
-
"step": 3150
|
| 452 |
-
},
|
| 453 |
-
{
|
| 454 |
-
"epoch": 0.4155304505908324,
|
| 455 |
-
"grad_norm": 0.8571991920471191,
|
| 456 |
-
"learning_rate": 1.8579660570977505e-05,
|
| 457 |
-
"loss": 0.2076,
|
| 458 |
-
"step": 3200
|
| 459 |
-
},
|
| 460 |
-
{
|
| 461 |
-
"epoch": 0.42202311388131414,
|
| 462 |
-
"grad_norm": 1.222774863243103,
|
| 463 |
-
"learning_rate": 1.8553348243652153e-05,
|
| 464 |
-
"loss": 0.2094,
|
| 465 |
-
"step": 3250
|
| 466 |
-
},
|
| 467 |
-
{
|
| 468 |
-
"epoch": 0.4285157771717959,
|
| 469 |
-
"grad_norm": 0.9631436467170715,
|
| 470 |
-
"learning_rate": 1.85270359163268e-05,
|
| 471 |
-
"loss": 0.2123,
|
| 472 |
-
"step": 3300
|
| 473 |
-
},
|
| 474 |
-
{
|
| 475 |
-
"epoch": 0.4350084404622776,
|
| 476 |
-
"grad_norm": 0.5590534806251526,
|
| 477 |
-
"learning_rate": 1.8500723589001446e-05,
|
| 478 |
-
"loss": 0.2028,
|
| 479 |
-
"step": 3350
|
| 480 |
-
},
|
| 481 |
-
{
|
| 482 |
-
"epoch": 0.44150110375275936,
|
| 483 |
-
"grad_norm": 0.7615678310394287,
|
| 484 |
-
"learning_rate": 1.8474411261676098e-05,
|
| 485 |
-
"loss": 0.2125,
|
| 486 |
-
"step": 3400
|
| 487 |
-
},
|
| 488 |
-
{
|
| 489 |
-
"epoch": 0.4479937670432411,
|
| 490 |
-
"grad_norm": 0.7124540209770203,
|
| 491 |
-
"learning_rate": 1.8448098934350743e-05,
|
| 492 |
-
"loss": 0.206,
|
| 493 |
-
"step": 3450
|
| 494 |
-
},
|
| 495 |
-
{
|
| 496 |
-
"epoch": 0.4544864303337229,
|
| 497 |
-
"grad_norm": 0.8492136001586914,
|
| 498 |
-
"learning_rate": 1.8421786607025395e-05,
|
| 499 |
-
"loss": 0.2054,
|
| 500 |
-
"step": 3500
|
| 501 |
-
},
|
| 502 |
-
{
|
| 503 |
-
"epoch": 0.46097909362420464,
|
| 504 |
-
"grad_norm": 1.0388613939285278,
|
| 505 |
-
"learning_rate": 1.839547427970004e-05,
|
| 506 |
-
"loss": 0.2117,
|
| 507 |
-
"step": 3550
|
| 508 |
-
},
|
| 509 |
-
{
|
| 510 |
-
"epoch": 0.4674717569146864,
|
| 511 |
-
"grad_norm": 0.8087108731269836,
|
| 512 |
-
"learning_rate": 1.836916195237469e-05,
|
| 513 |
-
"loss": 0.2192,
|
| 514 |
-
"step": 3600
|
| 515 |
-
},
|
| 516 |
-
{
|
| 517 |
-
"epoch": 0.47396442020516816,
|
| 518 |
-
"grad_norm": 0.7446919679641724,
|
| 519 |
-
"learning_rate": 1.8342849625049336e-05,
|
| 520 |
-
"loss": 0.2112,
|
| 521 |
-
"step": 3650
|
| 522 |
-
},
|
| 523 |
-
{
|
| 524 |
-
"epoch": 0.4804570834956499,
|
| 525 |
-
"grad_norm": 0.6669062972068787,
|
| 526 |
-
"learning_rate": 1.8316537297723984e-05,
|
| 527 |
-
"loss": 0.2167,
|
| 528 |
-
"step": 3700
|
| 529 |
-
},
|
| 530 |
-
{
|
| 531 |
-
"epoch": 0.4869497467861317,
|
| 532 |
-
"grad_norm": 0.519498884677887,
|
| 533 |
-
"learning_rate": 1.8290224970398633e-05,
|
| 534 |
-
"loss": 0.2069,
|
| 535 |
-
"step": 3750
|
| 536 |
-
},
|
| 537 |
-
{
|
| 538 |
-
"epoch": 0.49344241007661344,
|
| 539 |
-
"grad_norm": 0.5735522508621216,
|
| 540 |
-
"learning_rate": 1.826391264307328e-05,
|
| 541 |
-
"loss": 0.2102,
|
| 542 |
-
"step": 3800
|
| 543 |
-
},
|
| 544 |
-
{
|
| 545 |
-
"epoch": 0.4999350733670952,
|
| 546 |
-
"grad_norm": 0.6683741211891174,
|
| 547 |
-
"learning_rate": 1.823760031574793e-05,
|
| 548 |
-
"loss": 0.2141,
|
| 549 |
-
"step": 3850
|
| 550 |
-
},
|
| 551 |
-
{
|
| 552 |
-
"epoch": 0.5064277366575769,
|
| 553 |
-
"grad_norm": 0.6239739060401917,
|
| 554 |
-
"learning_rate": 1.8211287988422578e-05,
|
| 555 |
-
"loss": 0.1975,
|
| 556 |
-
"step": 3900
|
| 557 |
-
},
|
| 558 |
-
{
|
| 559 |
-
"epoch": 0.5129203999480587,
|
| 560 |
-
"grad_norm": 0.6327545046806335,
|
| 561 |
-
"learning_rate": 1.8184975661097226e-05,
|
| 562 |
-
"loss": 0.2165,
|
| 563 |
-
"step": 3950
|
| 564 |
-
},
|
| 565 |
-
{
|
| 566 |
-
"epoch": 0.5194130632385404,
|
| 567 |
-
"grad_norm": 0.6098693609237671,
|
| 568 |
-
"learning_rate": 1.8158663333771874e-05,
|
| 569 |
-
"loss": 0.2034,
|
| 570 |
-
"step": 4000
|
| 571 |
-
},
|
| 572 |
-
{
|
| 573 |
-
"epoch": 0.5259057265290222,
|
| 574 |
-
"grad_norm": 0.7356197237968445,
|
| 575 |
-
"learning_rate": 1.8132351006446522e-05,
|
| 576 |
-
"loss": 0.2041,
|
| 577 |
-
"step": 4050
|
| 578 |
-
},
|
| 579 |
-
{
|
| 580 |
-
"epoch": 0.5323983898195039,
|
| 581 |
-
"grad_norm": 0.7616349458694458,
|
| 582 |
-
"learning_rate": 1.810603867912117e-05,
|
| 583 |
-
"loss": 0.2121,
|
| 584 |
-
"step": 4100
|
| 585 |
-
},
|
| 586 |
-
{
|
| 587 |
-
"epoch": 0.5388910531099858,
|
| 588 |
-
"grad_norm": 0.6561347842216492,
|
| 589 |
-
"learning_rate": 1.807972635179582e-05,
|
| 590 |
-
"loss": 0.2002,
|
| 591 |
-
"step": 4150
|
| 592 |
-
},
|
| 593 |
-
{
|
| 594 |
-
"epoch": 0.5453837164004675,
|
| 595 |
-
"grad_norm": 0.6744963526725769,
|
| 596 |
-
"learning_rate": 1.8053414024470467e-05,
|
| 597 |
-
"loss": 0.2029,
|
| 598 |
-
"step": 4200
|
| 599 |
-
},
|
| 600 |
-
{
|
| 601 |
-
"epoch": 0.5518763796909493,
|
| 602 |
-
"grad_norm": 0.7836194038391113,
|
| 603 |
-
"learning_rate": 1.8027101697145112e-05,
|
| 604 |
-
"loss": 0.2055,
|
| 605 |
-
"step": 4250
|
| 606 |
-
},
|
| 607 |
-
{
|
| 608 |
-
"epoch": 0.558369042981431,
|
| 609 |
-
"grad_norm": 0.6983553767204285,
|
| 610 |
-
"learning_rate": 1.800078936981976e-05,
|
| 611 |
-
"loss": 0.2139,
|
| 612 |
-
"step": 4300
|
| 613 |
-
},
|
| 614 |
-
{
|
| 615 |
-
"epoch": 0.5648617062719128,
|
| 616 |
-
"grad_norm": 0.6315485239028931,
|
| 617 |
-
"learning_rate": 1.797447704249441e-05,
|
| 618 |
-
"loss": 0.1978,
|
| 619 |
-
"step": 4350
|
| 620 |
-
},
|
| 621 |
-
{
|
| 622 |
-
"epoch": 0.5713543695623945,
|
| 623 |
-
"grad_norm": 0.5928835868835449,
|
| 624 |
-
"learning_rate": 1.7948164715169057e-05,
|
| 625 |
-
"loss": 0.2027,
|
| 626 |
-
"step": 4400
|
| 627 |
-
},
|
| 628 |
-
{
|
| 629 |
-
"epoch": 0.5778470328528762,
|
| 630 |
-
"grad_norm": 0.8622831702232361,
|
| 631 |
-
"learning_rate": 1.7921852387843705e-05,
|
| 632 |
-
"loss": 0.2104,
|
| 633 |
-
"step": 4450
|
| 634 |
-
},
|
| 635 |
-
{
|
| 636 |
-
"epoch": 0.584339696143358,
|
| 637 |
-
"grad_norm": 0.7967308759689331,
|
| 638 |
-
"learning_rate": 1.7895540060518354e-05,
|
| 639 |
-
"loss": 0.2328,
|
| 640 |
-
"step": 4500
|
| 641 |
-
},
|
| 642 |
-
{
|
| 643 |
-
"epoch": 0.5908323594338397,
|
| 644 |
-
"grad_norm": 0.8289620280265808,
|
| 645 |
-
"learning_rate": 1.7869227733193002e-05,
|
| 646 |
-
"loss": 0.2099,
|
| 647 |
-
"step": 4550
|
| 648 |
-
},
|
| 649 |
-
{
|
| 650 |
-
"epoch": 0.5973250227243215,
|
| 651 |
-
"grad_norm": 0.7185404300689697,
|
| 652 |
-
"learning_rate": 1.784291540586765e-05,
|
| 653 |
-
"loss": 0.2045,
|
| 654 |
-
"step": 4600
|
| 655 |
-
},
|
| 656 |
-
{
|
| 657 |
-
"epoch": 0.6038176860148032,
|
| 658 |
-
"grad_norm": 0.6304630637168884,
|
| 659 |
-
"learning_rate": 1.78166030785423e-05,
|
| 660 |
-
"loss": 0.2055,
|
| 661 |
-
"step": 4650
|
| 662 |
-
},
|
| 663 |
-
{
|
| 664 |
-
"epoch": 0.610310349305285,
|
| 665 |
-
"grad_norm": 0.614983856678009,
|
| 666 |
-
"learning_rate": 1.7790290751216947e-05,
|
| 667 |
-
"loss": 0.1979,
|
| 668 |
-
"step": 4700
|
| 669 |
-
},
|
| 670 |
-
{
|
| 671 |
-
"epoch": 0.6168030125957668,
|
| 672 |
-
"grad_norm": 0.568352460861206,
|
| 673 |
-
"learning_rate": 1.7763978423891595e-05,
|
| 674 |
-
"loss": 0.197,
|
| 675 |
-
"step": 4750
|
| 676 |
-
},
|
| 677 |
-
{
|
| 678 |
-
"epoch": 0.6232956758862486,
|
| 679 |
-
"grad_norm": 0.8283701539039612,
|
| 680 |
-
"learning_rate": 1.773766609656624e-05,
|
| 681 |
-
"loss": 0.1986,
|
| 682 |
-
"step": 4800
|
| 683 |
-
},
|
| 684 |
-
{
|
| 685 |
-
"epoch": 0.6297883391767303,
|
| 686 |
-
"grad_norm": 0.888967752456665,
|
| 687 |
-
"learning_rate": 1.7711353769240892e-05,
|
| 688 |
-
"loss": 0.2271,
|
| 689 |
-
"step": 4850
|
| 690 |
-
},
|
| 691 |
-
{
|
| 692 |
-
"epoch": 0.6362810024672121,
|
| 693 |
-
"grad_norm": 0.5292450189590454,
|
| 694 |
-
"learning_rate": 1.7685041441915537e-05,
|
| 695 |
-
"loss": 0.1971,
|
| 696 |
-
"step": 4900
|
| 697 |
-
},
|
| 698 |
-
{
|
| 699 |
-
"epoch": 0.6427736657576938,
|
| 700 |
-
"grad_norm": 0.5274189114570618,
|
| 701 |
-
"learning_rate": 1.765872911459019e-05,
|
| 702 |
-
"loss": 0.1963,
|
| 703 |
-
"step": 4950
|
| 704 |
-
},
|
| 705 |
-
{
|
| 706 |
-
"epoch": 0.6492663290481756,
|
| 707 |
-
"grad_norm": 0.5699307322502136,
|
| 708 |
-
"learning_rate": 1.7632416787264833e-05,
|
| 709 |
-
"loss": 0.1984,
|
| 710 |
-
"step": 5000
|
| 711 |
-
},
|
| 712 |
-
{
|
| 713 |
-
"epoch": 0.6557589923386573,
|
| 714 |
-
"grad_norm": 0.8022367358207703,
|
| 715 |
-
"learning_rate": 1.7606104459939485e-05,
|
| 716 |
-
"loss": 0.2053,
|
| 717 |
-
"step": 5050
|
| 718 |
-
},
|
| 719 |
-
{
|
| 720 |
-
"epoch": 0.6622516556291391,
|
| 721 |
-
"grad_norm": 0.6432430148124695,
|
| 722 |
-
"learning_rate": 1.757979213261413e-05,
|
| 723 |
-
"loss": 0.1991,
|
| 724 |
-
"step": 5100
|
| 725 |
-
},
|
| 726 |
-
{
|
| 727 |
-
"epoch": 0.6687443189196208,
|
| 728 |
-
"grad_norm": 0.7671304941177368,
|
| 729 |
-
"learning_rate": 1.755347980528878e-05,
|
| 730 |
-
"loss": 0.2064,
|
| 731 |
-
"step": 5150
|
| 732 |
-
},
|
| 733 |
-
{
|
| 734 |
-
"epoch": 0.6752369822101025,
|
| 735 |
-
"grad_norm": 0.5107030272483826,
|
| 736 |
-
"learning_rate": 1.7527167477963426e-05,
|
| 737 |
-
"loss": 0.2049,
|
| 738 |
-
"step": 5200
|
| 739 |
-
},
|
| 740 |
-
{
|
| 741 |
-
"epoch": 0.6817296455005843,
|
| 742 |
-
"grad_norm": 0.7239235639572144,
|
| 743 |
-
"learning_rate": 1.7500855150638075e-05,
|
| 744 |
-
"loss": 0.2001,
|
| 745 |
-
"step": 5250
|
| 746 |
-
},
|
| 747 |
-
{
|
| 748 |
-
"epoch": 0.688222308791066,
|
| 749 |
-
"grad_norm": 0.6116129755973816,
|
| 750 |
-
"learning_rate": 1.7474542823312723e-05,
|
| 751 |
-
"loss": 0.2193,
|
| 752 |
-
"step": 5300
|
| 753 |
-
},
|
| 754 |
-
{
|
| 755 |
-
"epoch": 0.6947149720815479,
|
| 756 |
-
"grad_norm": 0.5425911545753479,
|
| 757 |
-
"learning_rate": 1.744823049598737e-05,
|
| 758 |
-
"loss": 0.2001,
|
| 759 |
-
"step": 5350
|
| 760 |
-
},
|
| 761 |
-
{
|
| 762 |
-
"epoch": 0.7012076353720296,
|
| 763 |
-
"grad_norm": 0.6464748382568359,
|
| 764 |
-
"learning_rate": 1.742191816866202e-05,
|
| 765 |
-
"loss": 0.1963,
|
| 766 |
-
"step": 5400
|
| 767 |
-
},
|
| 768 |
-
{
|
| 769 |
-
"epoch": 0.7077002986625114,
|
| 770 |
-
"grad_norm": 0.8812252879142761,
|
| 771 |
-
"learning_rate": 1.7395605841336668e-05,
|
| 772 |
-
"loss": 0.196,
|
| 773 |
-
"step": 5450
|
| 774 |
-
},
|
| 775 |
-
{
|
| 776 |
-
"epoch": 0.7141929619529931,
|
| 777 |
-
"grad_norm": 0.6928241848945618,
|
| 778 |
-
"learning_rate": 1.7369293514011316e-05,
|
| 779 |
-
"loss": 0.2011,
|
| 780 |
-
"step": 5500
|
| 781 |
-
},
|
| 782 |
-
{
|
| 783 |
-
"epoch": 0.7206856252434749,
|
| 784 |
-
"grad_norm": 0.6892450451850891,
|
| 785 |
-
"learning_rate": 1.7342981186685965e-05,
|
| 786 |
-
"loss": 0.1949,
|
| 787 |
-
"step": 5550
|
| 788 |
-
},
|
| 789 |
-
{
|
| 790 |
-
"epoch": 0.7271782885339566,
|
| 791 |
-
"grad_norm": 0.4782065749168396,
|
| 792 |
-
"learning_rate": 1.7316668859360613e-05,
|
| 793 |
-
"loss": 0.194,
|
| 794 |
-
"step": 5600
|
| 795 |
-
},
|
| 796 |
-
{
|
| 797 |
-
"epoch": 0.7336709518244384,
|
| 798 |
-
"grad_norm": 0.6438505053520203,
|
| 799 |
-
"learning_rate": 1.729035653203526e-05,
|
| 800 |
-
"loss": 0.1967,
|
| 801 |
-
"step": 5650
|
| 802 |
-
},
|
| 803 |
-
{
|
| 804 |
-
"epoch": 0.7401636151149201,
|
| 805 |
-
"grad_norm": 0.5797818899154663,
|
| 806 |
-
"learning_rate": 1.726404420470991e-05,
|
| 807 |
-
"loss": 0.2185,
|
| 808 |
-
"step": 5700
|
| 809 |
-
},
|
| 810 |
-
{
|
| 811 |
-
"epoch": 0.7466562784054019,
|
| 812 |
-
"grad_norm": 0.6884586811065674,
|
| 813 |
-
"learning_rate": 1.7237731877384554e-05,
|
| 814 |
-
"loss": 0.1977,
|
| 815 |
-
"step": 5750
|
| 816 |
-
},
|
| 817 |
-
{
|
| 818 |
-
"epoch": 0.7531489416958836,
|
| 819 |
-
"grad_norm": 0.648883581161499,
|
| 820 |
-
"learning_rate": 1.7211419550059206e-05,
|
| 821 |
-
"loss": 0.1964,
|
| 822 |
-
"step": 5800
|
| 823 |
-
},
|
| 824 |
-
{
|
| 825 |
-
"epoch": 0.7596416049863655,
|
| 826 |
-
"grad_norm": 0.6440086960792542,
|
| 827 |
-
"learning_rate": 1.718510722273385e-05,
|
| 828 |
-
"loss": 0.2014,
|
| 829 |
-
"step": 5850
|
| 830 |
-
},
|
| 831 |
-
{
|
| 832 |
-
"epoch": 0.7661342682768472,
|
| 833 |
-
"grad_norm": 0.5619300007820129,
|
| 834 |
-
"learning_rate": 1.71587948954085e-05,
|
| 835 |
-
"loss": 0.1909,
|
| 836 |
-
"step": 5900
|
| 837 |
-
},
|
| 838 |
-
{
|
| 839 |
-
"epoch": 0.7726269315673289,
|
| 840 |
-
"grad_norm": 0.6859204769134521,
|
| 841 |
-
"learning_rate": 1.7132482568083147e-05,
|
| 842 |
-
"loss": 0.2049,
|
| 843 |
-
"step": 5950
|
| 844 |
-
},
|
| 845 |
-
{
|
| 846 |
-
"epoch": 0.7791195948578107,
|
| 847 |
-
"grad_norm": 0.6132592558860779,
|
| 848 |
-
"learning_rate": 1.7106170240757796e-05,
|
| 849 |
-
"loss": 0.2,
|
| 850 |
-
"step": 6000
|
| 851 |
-
},
|
| 852 |
-
{
|
| 853 |
-
"epoch": 0.7856122581482924,
|
| 854 |
-
"grad_norm": 0.7050901055335999,
|
| 855 |
-
"learning_rate": 1.7079857913432444e-05,
|
| 856 |
-
"loss": 0.189,
|
| 857 |
-
"step": 6050
|
| 858 |
-
},
|
| 859 |
-
{
|
| 860 |
-
"epoch": 0.7921049214387742,
|
| 861 |
-
"grad_norm": 0.6752614974975586,
|
| 862 |
-
"learning_rate": 1.7053545586107092e-05,
|
| 863 |
-
"loss": 0.2248,
|
| 864 |
-
"step": 6100
|
| 865 |
-
},
|
| 866 |
-
{
|
| 867 |
-
"epoch": 0.7985975847292559,
|
| 868 |
-
"grad_norm": 0.7186923623085022,
|
| 869 |
-
"learning_rate": 1.702723325878174e-05,
|
| 870 |
-
"loss": 0.1903,
|
| 871 |
-
"step": 6150
|
| 872 |
-
},
|
| 873 |
-
{
|
| 874 |
-
"epoch": 0.8050902480197377,
|
| 875 |
-
"grad_norm": 0.5991400480270386,
|
| 876 |
-
"learning_rate": 1.700092093145639e-05,
|
| 877 |
-
"loss": 0.197,
|
| 878 |
-
"step": 6200
|
| 879 |
-
},
|
| 880 |
-
{
|
| 881 |
-
"epoch": 0.8115829113102194,
|
| 882 |
-
"grad_norm": 0.9522245526313782,
|
| 883 |
-
"learning_rate": 1.6974608604131037e-05,
|
| 884 |
-
"loss": 0.1962,
|
| 885 |
-
"step": 6250
|
| 886 |
-
},
|
| 887 |
-
{
|
| 888 |
-
"epoch": 0.8180755746007012,
|
| 889 |
-
"grad_norm": 0.8645381927490234,
|
| 890 |
-
"learning_rate": 1.6948296276805686e-05,
|
| 891 |
-
"loss": 0.1901,
|
| 892 |
-
"step": 6300
|
| 893 |
-
},
|
| 894 |
-
{
|
| 895 |
-
"epoch": 0.8245682378911829,
|
| 896 |
-
"grad_norm": 0.5243034958839417,
|
| 897 |
-
"learning_rate": 1.6921983949480334e-05,
|
| 898 |
-
"loss": 0.194,
|
| 899 |
-
"step": 6350
|
| 900 |
-
},
|
| 901 |
-
{
|
| 902 |
-
"epoch": 0.8310609011816648,
|
| 903 |
-
"grad_norm": 0.5842151641845703,
|
| 904 |
-
"learning_rate": 1.6895671622154982e-05,
|
| 905 |
-
"loss": 0.1917,
|
| 906 |
-
"step": 6400
|
| 907 |
-
},
|
| 908 |
-
{
|
| 909 |
-
"epoch": 0.8375535644721465,
|
| 910 |
-
"grad_norm": 0.6111485362052917,
|
| 911 |
-
"learning_rate": 1.6869359294829627e-05,
|
| 912 |
-
"loss": 0.2042,
|
| 913 |
-
"step": 6450
|
| 914 |
-
},
|
| 915 |
-
{
|
| 916 |
-
"epoch": 0.8440462277626283,
|
| 917 |
-
"grad_norm": 0.6515288949012756,
|
| 918 |
-
"learning_rate": 1.684304696750428e-05,
|
| 919 |
-
"loss": 0.2032,
|
| 920 |
-
"step": 6500
|
| 921 |
-
},
|
| 922 |
-
{
|
| 923 |
-
"epoch": 0.85053889105311,
|
| 924 |
-
"grad_norm": 0.7596396207809448,
|
| 925 |
-
"learning_rate": 1.6816734640178924e-05,
|
| 926 |
-
"loss": 0.193,
|
| 927 |
-
"step": 6550
|
| 928 |
-
}
|
| 929 |
-
],
|
| 930 |
-
"logging_steps": 50,
|
| 931 |
-
"max_steps": 38505,
|
| 932 |
-
"num_input_tokens_seen": 0,
|
| 933 |
-
"num_train_epochs": 5,
|
| 934 |
-
"save_steps": 386,
|
| 935 |
-
"stateful_callbacks": {
|
| 936 |
-
"TrainerControl": {
|
| 937 |
-
"args": {
|
| 938 |
-
"should_epoch_stop": false,
|
| 939 |
-
"should_evaluate": false,
|
| 940 |
-
"should_log": false,
|
| 941 |
-
"should_save": true,
|
| 942 |
-
"should_training_stop": false
|
| 943 |
-
},
|
| 944 |
-
"attributes": {}
|
| 945 |
-
}
|
| 946 |
-
},
|
| 947 |
-
"total_flos": 5.487557747765412e+17,
|
| 948 |
-
"train_batch_size": 40,
|
| 949 |
-
"trial_name": null,
|
| 950 |
-
"trial_params": null
|
| 951 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
checkpoint-6948/added_tokens.json
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"[MASK]": 128000
|
| 3 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
checkpoint-6948/config.json
DELETED
|
@@ -1,69 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"architectures": [
|
| 3 |
-
"DebertaV2ForTokenClassification"
|
| 4 |
-
],
|
| 5 |
-
"attention_probs_dropout_prob": 0.1,
|
| 6 |
-
"hidden_act": "gelu",
|
| 7 |
-
"hidden_dropout_prob": 0.1,
|
| 8 |
-
"hidden_size": 768,
|
| 9 |
-
"id2label": {
|
| 10 |
-
"0": "O",
|
| 11 |
-
"1": "B-CHA",
|
| 12 |
-
"2": "I-CHA",
|
| 13 |
-
"3": "B-LOC",
|
| 14 |
-
"4": "I-LOC",
|
| 15 |
-
"5": "B-FAC",
|
| 16 |
-
"6": "I-FAC",
|
| 17 |
-
"7": "B-OBJ",
|
| 18 |
-
"8": "I-OBJ",
|
| 19 |
-
"9": "B-EVT",
|
| 20 |
-
"10": "I-EVT",
|
| 21 |
-
"11": "B-ORG",
|
| 22 |
-
"12": "I-ORG",
|
| 23 |
-
"13": "B-MISC",
|
| 24 |
-
"14": "I-MISC"
|
| 25 |
-
},
|
| 26 |
-
"initializer_range": 0.02,
|
| 27 |
-
"intermediate_size": 3072,
|
| 28 |
-
"label2id": {
|
| 29 |
-
"B-CHA": 1,
|
| 30 |
-
"B-EVT": 9,
|
| 31 |
-
"B-FAC": 5,
|
| 32 |
-
"B-LOC": 3,
|
| 33 |
-
"B-MISC": 13,
|
| 34 |
-
"B-OBJ": 7,
|
| 35 |
-
"B-ORG": 11,
|
| 36 |
-
"I-CHA": 2,
|
| 37 |
-
"I-EVT": 10,
|
| 38 |
-
"I-FAC": 6,
|
| 39 |
-
"I-LOC": 4,
|
| 40 |
-
"I-MISC": 14,
|
| 41 |
-
"I-OBJ": 8,
|
| 42 |
-
"I-ORG": 12,
|
| 43 |
-
"O": 0
|
| 44 |
-
},
|
| 45 |
-
"layer_norm_eps": 1e-07,
|
| 46 |
-
"legacy": true,
|
| 47 |
-
"max_position_embeddings": 512,
|
| 48 |
-
"max_relative_positions": -1,
|
| 49 |
-
"model_type": "deberta-v2",
|
| 50 |
-
"norm_rel_ebd": "layer_norm",
|
| 51 |
-
"num_attention_heads": 12,
|
| 52 |
-
"num_hidden_layers": 12,
|
| 53 |
-
"pad_token_id": 0,
|
| 54 |
-
"pooler_dropout": 0,
|
| 55 |
-
"pooler_hidden_act": "gelu",
|
| 56 |
-
"pooler_hidden_size": 768,
|
| 57 |
-
"pos_att_type": [
|
| 58 |
-
"p2c",
|
| 59 |
-
"c2p"
|
| 60 |
-
],
|
| 61 |
-
"position_biased_input": false,
|
| 62 |
-
"position_buckets": 256,
|
| 63 |
-
"relative_attention": true,
|
| 64 |
-
"share_att_key": true,
|
| 65 |
-
"torch_dtype": "float32",
|
| 66 |
-
"transformers_version": "4.55.4",
|
| 67 |
-
"type_vocab_size": 0,
|
| 68 |
-
"vocab_size": 128100
|
| 69 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
checkpoint-6948/special_tokens_map.json
DELETED
|
@@ -1,15 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"bos_token": "[CLS]",
|
| 3 |
-
"cls_token": "[CLS]",
|
| 4 |
-
"eos_token": "[SEP]",
|
| 5 |
-
"mask_token": "[MASK]",
|
| 6 |
-
"pad_token": "[PAD]",
|
| 7 |
-
"sep_token": "[SEP]",
|
| 8 |
-
"unk_token": {
|
| 9 |
-
"content": "[UNK]",
|
| 10 |
-
"lstrip": false,
|
| 11 |
-
"normalized": true,
|
| 12 |
-
"rstrip": false,
|
| 13 |
-
"single_word": false
|
| 14 |
-
}
|
| 15 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
checkpoint-6948/tokenizer.json
DELETED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-6948/tokenizer_config.json
DELETED
|
@@ -1,59 +0,0 @@
|
|
| 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": "[CLS]",
|
| 13 |
-
"lstrip": false,
|
| 14 |
-
"normalized": false,
|
| 15 |
-
"rstrip": false,
|
| 16 |
-
"single_word": false,
|
| 17 |
-
"special": true
|
| 18 |
-
},
|
| 19 |
-
"2": {
|
| 20 |
-
"content": "[SEP]",
|
| 21 |
-
"lstrip": false,
|
| 22 |
-
"normalized": false,
|
| 23 |
-
"rstrip": false,
|
| 24 |
-
"single_word": false,
|
| 25 |
-
"special": true
|
| 26 |
-
},
|
| 27 |
-
"3": {
|
| 28 |
-
"content": "[UNK]",
|
| 29 |
-
"lstrip": false,
|
| 30 |
-
"normalized": true,
|
| 31 |
-
"rstrip": false,
|
| 32 |
-
"single_word": false,
|
| 33 |
-
"special": true
|
| 34 |
-
},
|
| 35 |
-
"128000": {
|
| 36 |
-
"content": "[MASK]",
|
| 37 |
-
"lstrip": false,
|
| 38 |
-
"normalized": false,
|
| 39 |
-
"rstrip": false,
|
| 40 |
-
"single_word": false,
|
| 41 |
-
"special": true
|
| 42 |
-
}
|
| 43 |
-
},
|
| 44 |
-
"bos_token": "[CLS]",
|
| 45 |
-
"clean_up_tokenization_spaces": false,
|
| 46 |
-
"cls_token": "[CLS]",
|
| 47 |
-
"do_lower_case": false,
|
| 48 |
-
"eos_token": "[SEP]",
|
| 49 |
-
"extra_special_tokens": {},
|
| 50 |
-
"mask_token": "[MASK]",
|
| 51 |
-
"model_max_length": 1000000000000000019884624838656,
|
| 52 |
-
"pad_token": "[PAD]",
|
| 53 |
-
"sep_token": "[SEP]",
|
| 54 |
-
"sp_model_kwargs": {},
|
| 55 |
-
"split_by_punct": false,
|
| 56 |
-
"tokenizer_class": "DebertaV2Tokenizer",
|
| 57 |
-
"unk_token": "[UNK]",
|
| 58 |
-
"vocab_type": "spm"
|
| 59 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
checkpoint-6948/trainer_state.json
DELETED
|
@@ -1,1000 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"best_global_step": null,
|
| 3 |
-
"best_metric": null,
|
| 4 |
-
"best_model_checkpoint": null,
|
| 5 |
-
"epoch": 0.9022204908453447,
|
| 6 |
-
"eval_steps": 500,
|
| 7 |
-
"global_step": 6948,
|
| 8 |
-
"is_hyper_param_search": false,
|
| 9 |
-
"is_local_process_zero": true,
|
| 10 |
-
"is_world_process_zero": true,
|
| 11 |
-
"log_history": [
|
| 12 |
-
{
|
| 13 |
-
"epoch": 0.006492663290481756,
|
| 14 |
-
"grad_norm": 23.6250057220459,
|
| 15 |
-
"learning_rate": 1.9600000000000003e-06,
|
| 16 |
-
"loss": 3.3492,
|
| 17 |
-
"step": 50
|
| 18 |
-
},
|
| 19 |
-
{
|
| 20 |
-
"epoch": 0.012985326580963512,
|
| 21 |
-
"grad_norm": 1.5745655298233032,
|
| 22 |
-
"learning_rate": 3.96e-06,
|
| 23 |
-
"loss": 1.3837,
|
| 24 |
-
"step": 100
|
| 25 |
-
},
|
| 26 |
-
{
|
| 27 |
-
"epoch": 0.019477989871445268,
|
| 28 |
-
"grad_norm": 1.3241935968399048,
|
| 29 |
-
"learning_rate": 5.9600000000000005e-06,
|
| 30 |
-
"loss": 0.6673,
|
| 31 |
-
"step": 150
|
| 32 |
-
},
|
| 33 |
-
{
|
| 34 |
-
"epoch": 0.025970653161927024,
|
| 35 |
-
"grad_norm": 1.068557620048523,
|
| 36 |
-
"learning_rate": 7.960000000000002e-06,
|
| 37 |
-
"loss": 0.4426,
|
| 38 |
-
"step": 200
|
| 39 |
-
},
|
| 40 |
-
{
|
| 41 |
-
"epoch": 0.032463316452408776,
|
| 42 |
-
"grad_norm": 1.2953921556472778,
|
| 43 |
-
"learning_rate": 9.960000000000001e-06,
|
| 44 |
-
"loss": 0.363,
|
| 45 |
-
"step": 250
|
| 46 |
-
},
|
| 47 |
-
{
|
| 48 |
-
"epoch": 0.038955979742890535,
|
| 49 |
-
"grad_norm": 1.4753903150558472,
|
| 50 |
-
"learning_rate": 1.196e-05,
|
| 51 |
-
"loss": 0.3397,
|
| 52 |
-
"step": 300
|
| 53 |
-
},
|
| 54 |
-
{
|
| 55 |
-
"epoch": 0.04544864303337229,
|
| 56 |
-
"grad_norm": 1.245597243309021,
|
| 57 |
-
"learning_rate": 1.396e-05,
|
| 58 |
-
"loss": 0.3125,
|
| 59 |
-
"step": 350
|
| 60 |
-
},
|
| 61 |
-
{
|
| 62 |
-
"epoch": 0.05194130632385405,
|
| 63 |
-
"grad_norm": 1.3421887159347534,
|
| 64 |
-
"learning_rate": 1.5960000000000003e-05,
|
| 65 |
-
"loss": 0.3086,
|
| 66 |
-
"step": 400
|
| 67 |
-
},
|
| 68 |
-
{
|
| 69 |
-
"epoch": 0.0584339696143358,
|
| 70 |
-
"grad_norm": 1.7852529287338257,
|
| 71 |
-
"learning_rate": 1.796e-05,
|
| 72 |
-
"loss": 0.2883,
|
| 73 |
-
"step": 450
|
| 74 |
-
},
|
| 75 |
-
{
|
| 76 |
-
"epoch": 0.06492663290481755,
|
| 77 |
-
"grad_norm": 1.086064100265503,
|
| 78 |
-
"learning_rate": 1.9960000000000002e-05,
|
| 79 |
-
"loss": 0.2947,
|
| 80 |
-
"step": 500
|
| 81 |
-
},
|
| 82 |
-
{
|
| 83 |
-
"epoch": 0.07141929619529931,
|
| 84 |
-
"grad_norm": 0.8753414750099182,
|
| 85 |
-
"learning_rate": 1.9974213919221156e-05,
|
| 86 |
-
"loss": 0.2955,
|
| 87 |
-
"step": 550
|
| 88 |
-
},
|
| 89 |
-
{
|
| 90 |
-
"epoch": 0.07791195948578107,
|
| 91 |
-
"grad_norm": 0.7875685095787048,
|
| 92 |
-
"learning_rate": 1.9947901591895804e-05,
|
| 93 |
-
"loss": 0.2773,
|
| 94 |
-
"step": 600
|
| 95 |
-
},
|
| 96 |
-
{
|
| 97 |
-
"epoch": 0.08440462277626282,
|
| 98 |
-
"grad_norm": 1.1609506607055664,
|
| 99 |
-
"learning_rate": 1.9921589264570452e-05,
|
| 100 |
-
"loss": 0.2696,
|
| 101 |
-
"step": 650
|
| 102 |
-
},
|
| 103 |
-
{
|
| 104 |
-
"epoch": 0.09089728606674458,
|
| 105 |
-
"grad_norm": 0.9162603616714478,
|
| 106 |
-
"learning_rate": 1.98952769372451e-05,
|
| 107 |
-
"loss": 0.2667,
|
| 108 |
-
"step": 700
|
| 109 |
-
},
|
| 110 |
-
{
|
| 111 |
-
"epoch": 0.09738994935722634,
|
| 112 |
-
"grad_norm": 1.0398370027542114,
|
| 113 |
-
"learning_rate": 1.986896460991975e-05,
|
| 114 |
-
"loss": 0.277,
|
| 115 |
-
"step": 750
|
| 116 |
-
},
|
| 117 |
-
{
|
| 118 |
-
"epoch": 0.1038826126477081,
|
| 119 |
-
"grad_norm": 0.9404798150062561,
|
| 120 |
-
"learning_rate": 1.9842652282594397e-05,
|
| 121 |
-
"loss": 0.2584,
|
| 122 |
-
"step": 800
|
| 123 |
-
},
|
| 124 |
-
{
|
| 125 |
-
"epoch": 0.11037527593818984,
|
| 126 |
-
"grad_norm": 1.0353909730911255,
|
| 127 |
-
"learning_rate": 1.9816339955269045e-05,
|
| 128 |
-
"loss": 0.2533,
|
| 129 |
-
"step": 850
|
| 130 |
-
},
|
| 131 |
-
{
|
| 132 |
-
"epoch": 0.1168679392286716,
|
| 133 |
-
"grad_norm": 1.485606074333191,
|
| 134 |
-
"learning_rate": 1.9790027627943694e-05,
|
| 135 |
-
"loss": 0.2506,
|
| 136 |
-
"step": 900
|
| 137 |
-
},
|
| 138 |
-
{
|
| 139 |
-
"epoch": 0.12336060251915336,
|
| 140 |
-
"grad_norm": 0.9874151945114136,
|
| 141 |
-
"learning_rate": 1.9763715300618342e-05,
|
| 142 |
-
"loss": 0.2515,
|
| 143 |
-
"step": 950
|
| 144 |
-
},
|
| 145 |
-
{
|
| 146 |
-
"epoch": 0.1298532658096351,
|
| 147 |
-
"grad_norm": 1.036007285118103,
|
| 148 |
-
"learning_rate": 1.973740297329299e-05,
|
| 149 |
-
"loss": 0.244,
|
| 150 |
-
"step": 1000
|
| 151 |
-
},
|
| 152 |
-
{
|
| 153 |
-
"epoch": 0.13634592910011686,
|
| 154 |
-
"grad_norm": 0.946956217288971,
|
| 155 |
-
"learning_rate": 1.9711090645967635e-05,
|
| 156 |
-
"loss": 0.2518,
|
| 157 |
-
"step": 1050
|
| 158 |
-
},
|
| 159 |
-
{
|
| 160 |
-
"epoch": 0.14283859239059862,
|
| 161 |
-
"grad_norm": 0.9096735119819641,
|
| 162 |
-
"learning_rate": 1.9684778318642287e-05,
|
| 163 |
-
"loss": 0.2543,
|
| 164 |
-
"step": 1100
|
| 165 |
-
},
|
| 166 |
-
{
|
| 167 |
-
"epoch": 0.14933125568108038,
|
| 168 |
-
"grad_norm": 0.9371875524520874,
|
| 169 |
-
"learning_rate": 1.965846599131693e-05,
|
| 170 |
-
"loss": 0.2605,
|
| 171 |
-
"step": 1150
|
| 172 |
-
},
|
| 173 |
-
{
|
| 174 |
-
"epoch": 0.15582391897156214,
|
| 175 |
-
"grad_norm": 0.8072580695152283,
|
| 176 |
-
"learning_rate": 1.9632153663991583e-05,
|
| 177 |
-
"loss": 0.2423,
|
| 178 |
-
"step": 1200
|
| 179 |
-
},
|
| 180 |
-
{
|
| 181 |
-
"epoch": 0.1623165822620439,
|
| 182 |
-
"grad_norm": 0.6907565593719482,
|
| 183 |
-
"learning_rate": 1.9605841336666228e-05,
|
| 184 |
-
"loss": 0.2315,
|
| 185 |
-
"step": 1250
|
| 186 |
-
},
|
| 187 |
-
{
|
| 188 |
-
"epoch": 0.16880924555252563,
|
| 189 |
-
"grad_norm": 0.6676374673843384,
|
| 190 |
-
"learning_rate": 1.957952900934088e-05,
|
| 191 |
-
"loss": 0.2363,
|
| 192 |
-
"step": 1300
|
| 193 |
-
},
|
| 194 |
-
{
|
| 195 |
-
"epoch": 0.1753019088430074,
|
| 196 |
-
"grad_norm": 0.7281391024589539,
|
| 197 |
-
"learning_rate": 1.9553216682015525e-05,
|
| 198 |
-
"loss": 0.2399,
|
| 199 |
-
"step": 1350
|
| 200 |
-
},
|
| 201 |
-
{
|
| 202 |
-
"epoch": 0.18179457213348915,
|
| 203 |
-
"grad_norm": 0.8105588555335999,
|
| 204 |
-
"learning_rate": 1.9526904354690177e-05,
|
| 205 |
-
"loss": 0.2448,
|
| 206 |
-
"step": 1400
|
| 207 |
-
},
|
| 208 |
-
{
|
| 209 |
-
"epoch": 0.1882872354239709,
|
| 210 |
-
"grad_norm": 0.7466333508491516,
|
| 211 |
-
"learning_rate": 1.950059202736482e-05,
|
| 212 |
-
"loss": 0.2388,
|
| 213 |
-
"step": 1450
|
| 214 |
-
},
|
| 215 |
-
{
|
| 216 |
-
"epoch": 0.19477989871445267,
|
| 217 |
-
"grad_norm": 0.5949985384941101,
|
| 218 |
-
"learning_rate": 1.947427970003947e-05,
|
| 219 |
-
"loss": 0.233,
|
| 220 |
-
"step": 1500
|
| 221 |
-
},
|
| 222 |
-
{
|
| 223 |
-
"epoch": 0.20127256200493443,
|
| 224 |
-
"grad_norm": 0.7439960241317749,
|
| 225 |
-
"learning_rate": 1.9447967372714118e-05,
|
| 226 |
-
"loss": 0.2448,
|
| 227 |
-
"step": 1550
|
| 228 |
-
},
|
| 229 |
-
{
|
| 230 |
-
"epoch": 0.2077652252954162,
|
| 231 |
-
"grad_norm": 0.9508784413337708,
|
| 232 |
-
"learning_rate": 1.9421655045388766e-05,
|
| 233 |
-
"loss": 0.2353,
|
| 234 |
-
"step": 1600
|
| 235 |
-
},
|
| 236 |
-
{
|
| 237 |
-
"epoch": 0.21425788858589795,
|
| 238 |
-
"grad_norm": 0.8785332441329956,
|
| 239 |
-
"learning_rate": 1.9395342718063415e-05,
|
| 240 |
-
"loss": 0.2249,
|
| 241 |
-
"step": 1650
|
| 242 |
-
},
|
| 243 |
-
{
|
| 244 |
-
"epoch": 0.22075055187637968,
|
| 245 |
-
"grad_norm": 0.7568134069442749,
|
| 246 |
-
"learning_rate": 1.9369030390738063e-05,
|
| 247 |
-
"loss": 0.2307,
|
| 248 |
-
"step": 1700
|
| 249 |
-
},
|
| 250 |
-
{
|
| 251 |
-
"epoch": 0.22724321516686144,
|
| 252 |
-
"grad_norm": 0.6832641363143921,
|
| 253 |
-
"learning_rate": 1.934271806341271e-05,
|
| 254 |
-
"loss": 0.2263,
|
| 255 |
-
"step": 1750
|
| 256 |
-
},
|
| 257 |
-
{
|
| 258 |
-
"epoch": 0.2337358784573432,
|
| 259 |
-
"grad_norm": 0.9721575379371643,
|
| 260 |
-
"learning_rate": 1.931640573608736e-05,
|
| 261 |
-
"loss": 0.2287,
|
| 262 |
-
"step": 1800
|
| 263 |
-
},
|
| 264 |
-
{
|
| 265 |
-
"epoch": 0.24022854174782496,
|
| 266 |
-
"grad_norm": 0.620299220085144,
|
| 267 |
-
"learning_rate": 1.9290093408762008e-05,
|
| 268 |
-
"loss": 0.226,
|
| 269 |
-
"step": 1850
|
| 270 |
-
},
|
| 271 |
-
{
|
| 272 |
-
"epoch": 0.24672120503830672,
|
| 273 |
-
"grad_norm": 0.6656680703163147,
|
| 274 |
-
"learning_rate": 1.9263781081436656e-05,
|
| 275 |
-
"loss": 0.2223,
|
| 276 |
-
"step": 1900
|
| 277 |
-
},
|
| 278 |
-
{
|
| 279 |
-
"epoch": 0.25321386832878845,
|
| 280 |
-
"grad_norm": 0.8536450266838074,
|
| 281 |
-
"learning_rate": 1.9237468754111304e-05,
|
| 282 |
-
"loss": 0.2332,
|
| 283 |
-
"step": 1950
|
| 284 |
-
},
|
| 285 |
-
{
|
| 286 |
-
"epoch": 0.2597065316192702,
|
| 287 |
-
"grad_norm": 0.9754176139831543,
|
| 288 |
-
"learning_rate": 1.921115642678595e-05,
|
| 289 |
-
"loss": 0.2211,
|
| 290 |
-
"step": 2000
|
| 291 |
-
},
|
| 292 |
-
{
|
| 293 |
-
"epoch": 0.26619919490975197,
|
| 294 |
-
"grad_norm": 0.6166796684265137,
|
| 295 |
-
"learning_rate": 1.9184844099460598e-05,
|
| 296 |
-
"loss": 0.2201,
|
| 297 |
-
"step": 2050
|
| 298 |
-
},
|
| 299 |
-
{
|
| 300 |
-
"epoch": 0.2726918582002337,
|
| 301 |
-
"grad_norm": 0.8182518482208252,
|
| 302 |
-
"learning_rate": 1.9158531772135246e-05,
|
| 303 |
-
"loss": 0.2213,
|
| 304 |
-
"step": 2100
|
| 305 |
-
},
|
| 306 |
-
{
|
| 307 |
-
"epoch": 0.2791845214907155,
|
| 308 |
-
"grad_norm": 0.8245829939842224,
|
| 309 |
-
"learning_rate": 1.9132219444809894e-05,
|
| 310 |
-
"loss": 0.2208,
|
| 311 |
-
"step": 2150
|
| 312 |
-
},
|
| 313 |
-
{
|
| 314 |
-
"epoch": 0.28567718478119725,
|
| 315 |
-
"grad_norm": 0.6423109173774719,
|
| 316 |
-
"learning_rate": 1.9105907117484542e-05,
|
| 317 |
-
"loss": 0.2308,
|
| 318 |
-
"step": 2200
|
| 319 |
-
},
|
| 320 |
-
{
|
| 321 |
-
"epoch": 0.292169848071679,
|
| 322 |
-
"grad_norm": 0.7164100408554077,
|
| 323 |
-
"learning_rate": 1.907959479015919e-05,
|
| 324 |
-
"loss": 0.2193,
|
| 325 |
-
"step": 2250
|
| 326 |
-
},
|
| 327 |
-
{
|
| 328 |
-
"epoch": 0.29866251136216077,
|
| 329 |
-
"grad_norm": 0.7910536527633667,
|
| 330 |
-
"learning_rate": 1.905328246283384e-05,
|
| 331 |
-
"loss": 0.2292,
|
| 332 |
-
"step": 2300
|
| 333 |
-
},
|
| 334 |
-
{
|
| 335 |
-
"epoch": 0.3051551746526425,
|
| 336 |
-
"grad_norm": 0.9699934720993042,
|
| 337 |
-
"learning_rate": 1.9026970135508487e-05,
|
| 338 |
-
"loss": 0.2175,
|
| 339 |
-
"step": 2350
|
| 340 |
-
},
|
| 341 |
-
{
|
| 342 |
-
"epoch": 0.3116478379431243,
|
| 343 |
-
"grad_norm": 1.0665416717529297,
|
| 344 |
-
"learning_rate": 1.9000657808183136e-05,
|
| 345 |
-
"loss": 0.2252,
|
| 346 |
-
"step": 2400
|
| 347 |
-
},
|
| 348 |
-
{
|
| 349 |
-
"epoch": 0.31814050123360604,
|
| 350 |
-
"grad_norm": 0.9915699362754822,
|
| 351 |
-
"learning_rate": 1.8974345480857784e-05,
|
| 352 |
-
"loss": 0.215,
|
| 353 |
-
"step": 2450
|
| 354 |
-
},
|
| 355 |
-
{
|
| 356 |
-
"epoch": 0.3246331645240878,
|
| 357 |
-
"grad_norm": 0.7021859288215637,
|
| 358 |
-
"learning_rate": 1.8948033153532432e-05,
|
| 359 |
-
"loss": 0.2123,
|
| 360 |
-
"step": 2500
|
| 361 |
-
},
|
| 362 |
-
{
|
| 363 |
-
"epoch": 0.33112582781456956,
|
| 364 |
-
"grad_norm": 0.6402600407600403,
|
| 365 |
-
"learning_rate": 1.892172082620708e-05,
|
| 366 |
-
"loss": 0.2176,
|
| 367 |
-
"step": 2550
|
| 368 |
-
},
|
| 369 |
-
{
|
| 370 |
-
"epoch": 0.33761849110505127,
|
| 371 |
-
"grad_norm": 0.9925078749656677,
|
| 372 |
-
"learning_rate": 1.8895408498881725e-05,
|
| 373 |
-
"loss": 0.223,
|
| 374 |
-
"step": 2600
|
| 375 |
-
},
|
| 376 |
-
{
|
| 377 |
-
"epoch": 0.344111154395533,
|
| 378 |
-
"grad_norm": 0.5188687443733215,
|
| 379 |
-
"learning_rate": 1.8869096171556377e-05,
|
| 380 |
-
"loss": 0.2174,
|
| 381 |
-
"step": 2650
|
| 382 |
-
},
|
| 383 |
-
{
|
| 384 |
-
"epoch": 0.3506038176860148,
|
| 385 |
-
"grad_norm": 0.7427687644958496,
|
| 386 |
-
"learning_rate": 1.8842783844231022e-05,
|
| 387 |
-
"loss": 0.2181,
|
| 388 |
-
"step": 2700
|
| 389 |
-
},
|
| 390 |
-
{
|
| 391 |
-
"epoch": 0.35709648097649654,
|
| 392 |
-
"grad_norm": 0.6055501103401184,
|
| 393 |
-
"learning_rate": 1.8816471516905674e-05,
|
| 394 |
-
"loss": 0.2097,
|
| 395 |
-
"step": 2750
|
| 396 |
-
},
|
| 397 |
-
{
|
| 398 |
-
"epoch": 0.3635891442669783,
|
| 399 |
-
"grad_norm": 0.6529932022094727,
|
| 400 |
-
"learning_rate": 1.879015918958032e-05,
|
| 401 |
-
"loss": 0.2133,
|
| 402 |
-
"step": 2800
|
| 403 |
-
},
|
| 404 |
-
{
|
| 405 |
-
"epoch": 0.37008180755746006,
|
| 406 |
-
"grad_norm": 0.9499914646148682,
|
| 407 |
-
"learning_rate": 1.8763846862254967e-05,
|
| 408 |
-
"loss": 0.2189,
|
| 409 |
-
"step": 2850
|
| 410 |
-
},
|
| 411 |
-
{
|
| 412 |
-
"epoch": 0.3765744708479418,
|
| 413 |
-
"grad_norm": 0.8217771053314209,
|
| 414 |
-
"learning_rate": 1.8737534534929615e-05,
|
| 415 |
-
"loss": 0.2172,
|
| 416 |
-
"step": 2900
|
| 417 |
-
},
|
| 418 |
-
{
|
| 419 |
-
"epoch": 0.3830671341384236,
|
| 420 |
-
"grad_norm": 1.031895399093628,
|
| 421 |
-
"learning_rate": 1.8711222207604263e-05,
|
| 422 |
-
"loss": 0.2171,
|
| 423 |
-
"step": 2950
|
| 424 |
-
},
|
| 425 |
-
{
|
| 426 |
-
"epoch": 0.38955979742890534,
|
| 427 |
-
"grad_norm": 0.7179540395736694,
|
| 428 |
-
"learning_rate": 1.8684909880278912e-05,
|
| 429 |
-
"loss": 0.2179,
|
| 430 |
-
"step": 3000
|
| 431 |
-
},
|
| 432 |
-
{
|
| 433 |
-
"epoch": 0.3960524607193871,
|
| 434 |
-
"grad_norm": 0.8115535974502563,
|
| 435 |
-
"learning_rate": 1.865859755295356e-05,
|
| 436 |
-
"loss": 0.2131,
|
| 437 |
-
"step": 3050
|
| 438 |
-
},
|
| 439 |
-
{
|
| 440 |
-
"epoch": 0.40254512400986886,
|
| 441 |
-
"grad_norm": 0.7659513354301453,
|
| 442 |
-
"learning_rate": 1.863228522562821e-05,
|
| 443 |
-
"loss": 0.2189,
|
| 444 |
-
"step": 3100
|
| 445 |
-
},
|
| 446 |
-
{
|
| 447 |
-
"epoch": 0.4090377873003506,
|
| 448 |
-
"grad_norm": 0.8645811080932617,
|
| 449 |
-
"learning_rate": 1.8605972898302857e-05,
|
| 450 |
-
"loss": 0.2453,
|
| 451 |
-
"step": 3150
|
| 452 |
-
},
|
| 453 |
-
{
|
| 454 |
-
"epoch": 0.4155304505908324,
|
| 455 |
-
"grad_norm": 0.8571991920471191,
|
| 456 |
-
"learning_rate": 1.8579660570977505e-05,
|
| 457 |
-
"loss": 0.2076,
|
| 458 |
-
"step": 3200
|
| 459 |
-
},
|
| 460 |
-
{
|
| 461 |
-
"epoch": 0.42202311388131414,
|
| 462 |
-
"grad_norm": 1.222774863243103,
|
| 463 |
-
"learning_rate": 1.8553348243652153e-05,
|
| 464 |
-
"loss": 0.2094,
|
| 465 |
-
"step": 3250
|
| 466 |
-
},
|
| 467 |
-
{
|
| 468 |
-
"epoch": 0.4285157771717959,
|
| 469 |
-
"grad_norm": 0.9631436467170715,
|
| 470 |
-
"learning_rate": 1.85270359163268e-05,
|
| 471 |
-
"loss": 0.2123,
|
| 472 |
-
"step": 3300
|
| 473 |
-
},
|
| 474 |
-
{
|
| 475 |
-
"epoch": 0.4350084404622776,
|
| 476 |
-
"grad_norm": 0.5590534806251526,
|
| 477 |
-
"learning_rate": 1.8500723589001446e-05,
|
| 478 |
-
"loss": 0.2028,
|
| 479 |
-
"step": 3350
|
| 480 |
-
},
|
| 481 |
-
{
|
| 482 |
-
"epoch": 0.44150110375275936,
|
| 483 |
-
"grad_norm": 0.7615678310394287,
|
| 484 |
-
"learning_rate": 1.8474411261676098e-05,
|
| 485 |
-
"loss": 0.2125,
|
| 486 |
-
"step": 3400
|
| 487 |
-
},
|
| 488 |
-
{
|
| 489 |
-
"epoch": 0.4479937670432411,
|
| 490 |
-
"grad_norm": 0.7124540209770203,
|
| 491 |
-
"learning_rate": 1.8448098934350743e-05,
|
| 492 |
-
"loss": 0.206,
|
| 493 |
-
"step": 3450
|
| 494 |
-
},
|
| 495 |
-
{
|
| 496 |
-
"epoch": 0.4544864303337229,
|
| 497 |
-
"grad_norm": 0.8492136001586914,
|
| 498 |
-
"learning_rate": 1.8421786607025395e-05,
|
| 499 |
-
"loss": 0.2054,
|
| 500 |
-
"step": 3500
|
| 501 |
-
},
|
| 502 |
-
{
|
| 503 |
-
"epoch": 0.46097909362420464,
|
| 504 |
-
"grad_norm": 1.0388613939285278,
|
| 505 |
-
"learning_rate": 1.839547427970004e-05,
|
| 506 |
-
"loss": 0.2117,
|
| 507 |
-
"step": 3550
|
| 508 |
-
},
|
| 509 |
-
{
|
| 510 |
-
"epoch": 0.4674717569146864,
|
| 511 |
-
"grad_norm": 0.8087108731269836,
|
| 512 |
-
"learning_rate": 1.836916195237469e-05,
|
| 513 |
-
"loss": 0.2192,
|
| 514 |
-
"step": 3600
|
| 515 |
-
},
|
| 516 |
-
{
|
| 517 |
-
"epoch": 0.47396442020516816,
|
| 518 |
-
"grad_norm": 0.7446919679641724,
|
| 519 |
-
"learning_rate": 1.8342849625049336e-05,
|
| 520 |
-
"loss": 0.2112,
|
| 521 |
-
"step": 3650
|
| 522 |
-
},
|
| 523 |
-
{
|
| 524 |
-
"epoch": 0.4804570834956499,
|
| 525 |
-
"grad_norm": 0.6669062972068787,
|
| 526 |
-
"learning_rate": 1.8316537297723984e-05,
|
| 527 |
-
"loss": 0.2167,
|
| 528 |
-
"step": 3700
|
| 529 |
-
},
|
| 530 |
-
{
|
| 531 |
-
"epoch": 0.4869497467861317,
|
| 532 |
-
"grad_norm": 0.519498884677887,
|
| 533 |
-
"learning_rate": 1.8290224970398633e-05,
|
| 534 |
-
"loss": 0.2069,
|
| 535 |
-
"step": 3750
|
| 536 |
-
},
|
| 537 |
-
{
|
| 538 |
-
"epoch": 0.49344241007661344,
|
| 539 |
-
"grad_norm": 0.5735522508621216,
|
| 540 |
-
"learning_rate": 1.826391264307328e-05,
|
| 541 |
-
"loss": 0.2102,
|
| 542 |
-
"step": 3800
|
| 543 |
-
},
|
| 544 |
-
{
|
| 545 |
-
"epoch": 0.4999350733670952,
|
| 546 |
-
"grad_norm": 0.6683741211891174,
|
| 547 |
-
"learning_rate": 1.823760031574793e-05,
|
| 548 |
-
"loss": 0.2141,
|
| 549 |
-
"step": 3850
|
| 550 |
-
},
|
| 551 |
-
{
|
| 552 |
-
"epoch": 0.5064277366575769,
|
| 553 |
-
"grad_norm": 0.6239739060401917,
|
| 554 |
-
"learning_rate": 1.8211287988422578e-05,
|
| 555 |
-
"loss": 0.1975,
|
| 556 |
-
"step": 3900
|
| 557 |
-
},
|
| 558 |
-
{
|
| 559 |
-
"epoch": 0.5129203999480587,
|
| 560 |
-
"grad_norm": 0.6327545046806335,
|
| 561 |
-
"learning_rate": 1.8184975661097226e-05,
|
| 562 |
-
"loss": 0.2165,
|
| 563 |
-
"step": 3950
|
| 564 |
-
},
|
| 565 |
-
{
|
| 566 |
-
"epoch": 0.5194130632385404,
|
| 567 |
-
"grad_norm": 0.6098693609237671,
|
| 568 |
-
"learning_rate": 1.8158663333771874e-05,
|
| 569 |
-
"loss": 0.2034,
|
| 570 |
-
"step": 4000
|
| 571 |
-
},
|
| 572 |
-
{
|
| 573 |
-
"epoch": 0.5259057265290222,
|
| 574 |
-
"grad_norm": 0.7356197237968445,
|
| 575 |
-
"learning_rate": 1.8132351006446522e-05,
|
| 576 |
-
"loss": 0.2041,
|
| 577 |
-
"step": 4050
|
| 578 |
-
},
|
| 579 |
-
{
|
| 580 |
-
"epoch": 0.5323983898195039,
|
| 581 |
-
"grad_norm": 0.7616349458694458,
|
| 582 |
-
"learning_rate": 1.810603867912117e-05,
|
| 583 |
-
"loss": 0.2121,
|
| 584 |
-
"step": 4100
|
| 585 |
-
},
|
| 586 |
-
{
|
| 587 |
-
"epoch": 0.5388910531099858,
|
| 588 |
-
"grad_norm": 0.6561347842216492,
|
| 589 |
-
"learning_rate": 1.807972635179582e-05,
|
| 590 |
-
"loss": 0.2002,
|
| 591 |
-
"step": 4150
|
| 592 |
-
},
|
| 593 |
-
{
|
| 594 |
-
"epoch": 0.5453837164004675,
|
| 595 |
-
"grad_norm": 0.6744963526725769,
|
| 596 |
-
"learning_rate": 1.8053414024470467e-05,
|
| 597 |
-
"loss": 0.2029,
|
| 598 |
-
"step": 4200
|
| 599 |
-
},
|
| 600 |
-
{
|
| 601 |
-
"epoch": 0.5518763796909493,
|
| 602 |
-
"grad_norm": 0.7836194038391113,
|
| 603 |
-
"learning_rate": 1.8027101697145112e-05,
|
| 604 |
-
"loss": 0.2055,
|
| 605 |
-
"step": 4250
|
| 606 |
-
},
|
| 607 |
-
{
|
| 608 |
-
"epoch": 0.558369042981431,
|
| 609 |
-
"grad_norm": 0.6983553767204285,
|
| 610 |
-
"learning_rate": 1.800078936981976e-05,
|
| 611 |
-
"loss": 0.2139,
|
| 612 |
-
"step": 4300
|
| 613 |
-
},
|
| 614 |
-
{
|
| 615 |
-
"epoch": 0.5648617062719128,
|
| 616 |
-
"grad_norm": 0.6315485239028931,
|
| 617 |
-
"learning_rate": 1.797447704249441e-05,
|
| 618 |
-
"loss": 0.1978,
|
| 619 |
-
"step": 4350
|
| 620 |
-
},
|
| 621 |
-
{
|
| 622 |
-
"epoch": 0.5713543695623945,
|
| 623 |
-
"grad_norm": 0.5928835868835449,
|
| 624 |
-
"learning_rate": 1.7948164715169057e-05,
|
| 625 |
-
"loss": 0.2027,
|
| 626 |
-
"step": 4400
|
| 627 |
-
},
|
| 628 |
-
{
|
| 629 |
-
"epoch": 0.5778470328528762,
|
| 630 |
-
"grad_norm": 0.8622831702232361,
|
| 631 |
-
"learning_rate": 1.7921852387843705e-05,
|
| 632 |
-
"loss": 0.2104,
|
| 633 |
-
"step": 4450
|
| 634 |
-
},
|
| 635 |
-
{
|
| 636 |
-
"epoch": 0.584339696143358,
|
| 637 |
-
"grad_norm": 0.7967308759689331,
|
| 638 |
-
"learning_rate": 1.7895540060518354e-05,
|
| 639 |
-
"loss": 0.2328,
|
| 640 |
-
"step": 4500
|
| 641 |
-
},
|
| 642 |
-
{
|
| 643 |
-
"epoch": 0.5908323594338397,
|
| 644 |
-
"grad_norm": 0.8289620280265808,
|
| 645 |
-
"learning_rate": 1.7869227733193002e-05,
|
| 646 |
-
"loss": 0.2099,
|
| 647 |
-
"step": 4550
|
| 648 |
-
},
|
| 649 |
-
{
|
| 650 |
-
"epoch": 0.5973250227243215,
|
| 651 |
-
"grad_norm": 0.7185404300689697,
|
| 652 |
-
"learning_rate": 1.784291540586765e-05,
|
| 653 |
-
"loss": 0.2045,
|
| 654 |
-
"step": 4600
|
| 655 |
-
},
|
| 656 |
-
{
|
| 657 |
-
"epoch": 0.6038176860148032,
|
| 658 |
-
"grad_norm": 0.6304630637168884,
|
| 659 |
-
"learning_rate": 1.78166030785423e-05,
|
| 660 |
-
"loss": 0.2055,
|
| 661 |
-
"step": 4650
|
| 662 |
-
},
|
| 663 |
-
{
|
| 664 |
-
"epoch": 0.610310349305285,
|
| 665 |
-
"grad_norm": 0.614983856678009,
|
| 666 |
-
"learning_rate": 1.7790290751216947e-05,
|
| 667 |
-
"loss": 0.1979,
|
| 668 |
-
"step": 4700
|
| 669 |
-
},
|
| 670 |
-
{
|
| 671 |
-
"epoch": 0.6168030125957668,
|
| 672 |
-
"grad_norm": 0.568352460861206,
|
| 673 |
-
"learning_rate": 1.7763978423891595e-05,
|
| 674 |
-
"loss": 0.197,
|
| 675 |
-
"step": 4750
|
| 676 |
-
},
|
| 677 |
-
{
|
| 678 |
-
"epoch": 0.6232956758862486,
|
| 679 |
-
"grad_norm": 0.8283701539039612,
|
| 680 |
-
"learning_rate": 1.773766609656624e-05,
|
| 681 |
-
"loss": 0.1986,
|
| 682 |
-
"step": 4800
|
| 683 |
-
},
|
| 684 |
-
{
|
| 685 |
-
"epoch": 0.6297883391767303,
|
| 686 |
-
"grad_norm": 0.888967752456665,
|
| 687 |
-
"learning_rate": 1.7711353769240892e-05,
|
| 688 |
-
"loss": 0.2271,
|
| 689 |
-
"step": 4850
|
| 690 |
-
},
|
| 691 |
-
{
|
| 692 |
-
"epoch": 0.6362810024672121,
|
| 693 |
-
"grad_norm": 0.5292450189590454,
|
| 694 |
-
"learning_rate": 1.7685041441915537e-05,
|
| 695 |
-
"loss": 0.1971,
|
| 696 |
-
"step": 4900
|
| 697 |
-
},
|
| 698 |
-
{
|
| 699 |
-
"epoch": 0.6427736657576938,
|
| 700 |
-
"grad_norm": 0.5274189114570618,
|
| 701 |
-
"learning_rate": 1.765872911459019e-05,
|
| 702 |
-
"loss": 0.1963,
|
| 703 |
-
"step": 4950
|
| 704 |
-
},
|
| 705 |
-
{
|
| 706 |
-
"epoch": 0.6492663290481756,
|
| 707 |
-
"grad_norm": 0.5699307322502136,
|
| 708 |
-
"learning_rate": 1.7632416787264833e-05,
|
| 709 |
-
"loss": 0.1984,
|
| 710 |
-
"step": 5000
|
| 711 |
-
},
|
| 712 |
-
{
|
| 713 |
-
"epoch": 0.6557589923386573,
|
| 714 |
-
"grad_norm": 0.8022367358207703,
|
| 715 |
-
"learning_rate": 1.7606104459939485e-05,
|
| 716 |
-
"loss": 0.2053,
|
| 717 |
-
"step": 5050
|
| 718 |
-
},
|
| 719 |
-
{
|
| 720 |
-
"epoch": 0.6622516556291391,
|
| 721 |
-
"grad_norm": 0.6432430148124695,
|
| 722 |
-
"learning_rate": 1.757979213261413e-05,
|
| 723 |
-
"loss": 0.1991,
|
| 724 |
-
"step": 5100
|
| 725 |
-
},
|
| 726 |
-
{
|
| 727 |
-
"epoch": 0.6687443189196208,
|
| 728 |
-
"grad_norm": 0.7671304941177368,
|
| 729 |
-
"learning_rate": 1.755347980528878e-05,
|
| 730 |
-
"loss": 0.2064,
|
| 731 |
-
"step": 5150
|
| 732 |
-
},
|
| 733 |
-
{
|
| 734 |
-
"epoch": 0.6752369822101025,
|
| 735 |
-
"grad_norm": 0.5107030272483826,
|
| 736 |
-
"learning_rate": 1.7527167477963426e-05,
|
| 737 |
-
"loss": 0.2049,
|
| 738 |
-
"step": 5200
|
| 739 |
-
},
|
| 740 |
-
{
|
| 741 |
-
"epoch": 0.6817296455005843,
|
| 742 |
-
"grad_norm": 0.7239235639572144,
|
| 743 |
-
"learning_rate": 1.7500855150638075e-05,
|
| 744 |
-
"loss": 0.2001,
|
| 745 |
-
"step": 5250
|
| 746 |
-
},
|
| 747 |
-
{
|
| 748 |
-
"epoch": 0.688222308791066,
|
| 749 |
-
"grad_norm": 0.6116129755973816,
|
| 750 |
-
"learning_rate": 1.7474542823312723e-05,
|
| 751 |
-
"loss": 0.2193,
|
| 752 |
-
"step": 5300
|
| 753 |
-
},
|
| 754 |
-
{
|
| 755 |
-
"epoch": 0.6947149720815479,
|
| 756 |
-
"grad_norm": 0.5425911545753479,
|
| 757 |
-
"learning_rate": 1.744823049598737e-05,
|
| 758 |
-
"loss": 0.2001,
|
| 759 |
-
"step": 5350
|
| 760 |
-
},
|
| 761 |
-
{
|
| 762 |
-
"epoch": 0.7012076353720296,
|
| 763 |
-
"grad_norm": 0.6464748382568359,
|
| 764 |
-
"learning_rate": 1.742191816866202e-05,
|
| 765 |
-
"loss": 0.1963,
|
| 766 |
-
"step": 5400
|
| 767 |
-
},
|
| 768 |
-
{
|
| 769 |
-
"epoch": 0.7077002986625114,
|
| 770 |
-
"grad_norm": 0.8812252879142761,
|
| 771 |
-
"learning_rate": 1.7395605841336668e-05,
|
| 772 |
-
"loss": 0.196,
|
| 773 |
-
"step": 5450
|
| 774 |
-
},
|
| 775 |
-
{
|
| 776 |
-
"epoch": 0.7141929619529931,
|
| 777 |
-
"grad_norm": 0.6928241848945618,
|
| 778 |
-
"learning_rate": 1.7369293514011316e-05,
|
| 779 |
-
"loss": 0.2011,
|
| 780 |
-
"step": 5500
|
| 781 |
-
},
|
| 782 |
-
{
|
| 783 |
-
"epoch": 0.7206856252434749,
|
| 784 |
-
"grad_norm": 0.6892450451850891,
|
| 785 |
-
"learning_rate": 1.7342981186685965e-05,
|
| 786 |
-
"loss": 0.1949,
|
| 787 |
-
"step": 5550
|
| 788 |
-
},
|
| 789 |
-
{
|
| 790 |
-
"epoch": 0.7271782885339566,
|
| 791 |
-
"grad_norm": 0.4782065749168396,
|
| 792 |
-
"learning_rate": 1.7316668859360613e-05,
|
| 793 |
-
"loss": 0.194,
|
| 794 |
-
"step": 5600
|
| 795 |
-
},
|
| 796 |
-
{
|
| 797 |
-
"epoch": 0.7336709518244384,
|
| 798 |
-
"grad_norm": 0.6438505053520203,
|
| 799 |
-
"learning_rate": 1.729035653203526e-05,
|
| 800 |
-
"loss": 0.1967,
|
| 801 |
-
"step": 5650
|
| 802 |
-
},
|
| 803 |
-
{
|
| 804 |
-
"epoch": 0.7401636151149201,
|
| 805 |
-
"grad_norm": 0.5797818899154663,
|
| 806 |
-
"learning_rate": 1.726404420470991e-05,
|
| 807 |
-
"loss": 0.2185,
|
| 808 |
-
"step": 5700
|
| 809 |
-
},
|
| 810 |
-
{
|
| 811 |
-
"epoch": 0.7466562784054019,
|
| 812 |
-
"grad_norm": 0.6884586811065674,
|
| 813 |
-
"learning_rate": 1.7237731877384554e-05,
|
| 814 |
-
"loss": 0.1977,
|
| 815 |
-
"step": 5750
|
| 816 |
-
},
|
| 817 |
-
{
|
| 818 |
-
"epoch": 0.7531489416958836,
|
| 819 |
-
"grad_norm": 0.648883581161499,
|
| 820 |
-
"learning_rate": 1.7211419550059206e-05,
|
| 821 |
-
"loss": 0.1964,
|
| 822 |
-
"step": 5800
|
| 823 |
-
},
|
| 824 |
-
{
|
| 825 |
-
"epoch": 0.7596416049863655,
|
| 826 |
-
"grad_norm": 0.6440086960792542,
|
| 827 |
-
"learning_rate": 1.718510722273385e-05,
|
| 828 |
-
"loss": 0.2014,
|
| 829 |
-
"step": 5850
|
| 830 |
-
},
|
| 831 |
-
{
|
| 832 |
-
"epoch": 0.7661342682768472,
|
| 833 |
-
"grad_norm": 0.5619300007820129,
|
| 834 |
-
"learning_rate": 1.71587948954085e-05,
|
| 835 |
-
"loss": 0.1909,
|
| 836 |
-
"step": 5900
|
| 837 |
-
},
|
| 838 |
-
{
|
| 839 |
-
"epoch": 0.7726269315673289,
|
| 840 |
-
"grad_norm": 0.6859204769134521,
|
| 841 |
-
"learning_rate": 1.7132482568083147e-05,
|
| 842 |
-
"loss": 0.2049,
|
| 843 |
-
"step": 5950
|
| 844 |
-
},
|
| 845 |
-
{
|
| 846 |
-
"epoch": 0.7791195948578107,
|
| 847 |
-
"grad_norm": 0.6132592558860779,
|
| 848 |
-
"learning_rate": 1.7106170240757796e-05,
|
| 849 |
-
"loss": 0.2,
|
| 850 |
-
"step": 6000
|
| 851 |
-
},
|
| 852 |
-
{
|
| 853 |
-
"epoch": 0.7856122581482924,
|
| 854 |
-
"grad_norm": 0.7050901055335999,
|
| 855 |
-
"learning_rate": 1.7079857913432444e-05,
|
| 856 |
-
"loss": 0.189,
|
| 857 |
-
"step": 6050
|
| 858 |
-
},
|
| 859 |
-
{
|
| 860 |
-
"epoch": 0.7921049214387742,
|
| 861 |
-
"grad_norm": 0.6752614974975586,
|
| 862 |
-
"learning_rate": 1.7053545586107092e-05,
|
| 863 |
-
"loss": 0.2248,
|
| 864 |
-
"step": 6100
|
| 865 |
-
},
|
| 866 |
-
{
|
| 867 |
-
"epoch": 0.7985975847292559,
|
| 868 |
-
"grad_norm": 0.7186923623085022,
|
| 869 |
-
"learning_rate": 1.702723325878174e-05,
|
| 870 |
-
"loss": 0.1903,
|
| 871 |
-
"step": 6150
|
| 872 |
-
},
|
| 873 |
-
{
|
| 874 |
-
"epoch": 0.8050902480197377,
|
| 875 |
-
"grad_norm": 0.5991400480270386,
|
| 876 |
-
"learning_rate": 1.700092093145639e-05,
|
| 877 |
-
"loss": 0.197,
|
| 878 |
-
"step": 6200
|
| 879 |
-
},
|
| 880 |
-
{
|
| 881 |
-
"epoch": 0.8115829113102194,
|
| 882 |
-
"grad_norm": 0.9522245526313782,
|
| 883 |
-
"learning_rate": 1.6974608604131037e-05,
|
| 884 |
-
"loss": 0.1962,
|
| 885 |
-
"step": 6250
|
| 886 |
-
},
|
| 887 |
-
{
|
| 888 |
-
"epoch": 0.8180755746007012,
|
| 889 |
-
"grad_norm": 0.8645381927490234,
|
| 890 |
-
"learning_rate": 1.6948296276805686e-05,
|
| 891 |
-
"loss": 0.1901,
|
| 892 |
-
"step": 6300
|
| 893 |
-
},
|
| 894 |
-
{
|
| 895 |
-
"epoch": 0.8245682378911829,
|
| 896 |
-
"grad_norm": 0.5243034958839417,
|
| 897 |
-
"learning_rate": 1.6921983949480334e-05,
|
| 898 |
-
"loss": 0.194,
|
| 899 |
-
"step": 6350
|
| 900 |
-
},
|
| 901 |
-
{
|
| 902 |
-
"epoch": 0.8310609011816648,
|
| 903 |
-
"grad_norm": 0.5842151641845703,
|
| 904 |
-
"learning_rate": 1.6895671622154982e-05,
|
| 905 |
-
"loss": 0.1917,
|
| 906 |
-
"step": 6400
|
| 907 |
-
},
|
| 908 |
-
{
|
| 909 |
-
"epoch": 0.8375535644721465,
|
| 910 |
-
"grad_norm": 0.6111485362052917,
|
| 911 |
-
"learning_rate": 1.6869359294829627e-05,
|
| 912 |
-
"loss": 0.2042,
|
| 913 |
-
"step": 6450
|
| 914 |
-
},
|
| 915 |
-
{
|
| 916 |
-
"epoch": 0.8440462277626283,
|
| 917 |
-
"grad_norm": 0.6515288949012756,
|
| 918 |
-
"learning_rate": 1.684304696750428e-05,
|
| 919 |
-
"loss": 0.2032,
|
| 920 |
-
"step": 6500
|
| 921 |
-
},
|
| 922 |
-
{
|
| 923 |
-
"epoch": 0.85053889105311,
|
| 924 |
-
"grad_norm": 0.7596396207809448,
|
| 925 |
-
"learning_rate": 1.6816734640178924e-05,
|
| 926 |
-
"loss": 0.193,
|
| 927 |
-
"step": 6550
|
| 928 |
-
},
|
| 929 |
-
{
|
| 930 |
-
"epoch": 0.8570315543435918,
|
| 931 |
-
"grad_norm": 0.6944254636764526,
|
| 932 |
-
"learning_rate": 1.6790422312853575e-05,
|
| 933 |
-
"loss": 0.1899,
|
| 934 |
-
"step": 6600
|
| 935 |
-
},
|
| 936 |
-
{
|
| 937 |
-
"epoch": 0.8635242176340735,
|
| 938 |
-
"grad_norm": 0.6190508604049683,
|
| 939 |
-
"learning_rate": 1.676410998552822e-05,
|
| 940 |
-
"loss": 0.1987,
|
| 941 |
-
"step": 6650
|
| 942 |
-
},
|
| 943 |
-
{
|
| 944 |
-
"epoch": 0.8700168809245552,
|
| 945 |
-
"grad_norm": 1.1515477895736694,
|
| 946 |
-
"learning_rate": 1.673779765820287e-05,
|
| 947 |
-
"loss": 0.196,
|
| 948 |
-
"step": 6700
|
| 949 |
-
},
|
| 950 |
-
{
|
| 951 |
-
"epoch": 0.876509544215037,
|
| 952 |
-
"grad_norm": 0.5803254842758179,
|
| 953 |
-
"learning_rate": 1.6711485330877517e-05,
|
| 954 |
-
"loss": 0.1923,
|
| 955 |
-
"step": 6750
|
| 956 |
-
},
|
| 957 |
-
{
|
| 958 |
-
"epoch": 0.8830022075055187,
|
| 959 |
-
"grad_norm": 0.8052871227264404,
|
| 960 |
-
"learning_rate": 1.6685173003552165e-05,
|
| 961 |
-
"loss": 0.1894,
|
| 962 |
-
"step": 6800
|
| 963 |
-
},
|
| 964 |
-
{
|
| 965 |
-
"epoch": 0.8894948707960005,
|
| 966 |
-
"grad_norm": 0.9313941597938538,
|
| 967 |
-
"learning_rate": 1.6658860676226813e-05,
|
| 968 |
-
"loss": 0.1889,
|
| 969 |
-
"step": 6850
|
| 970 |
-
},
|
| 971 |
-
{
|
| 972 |
-
"epoch": 0.8959875340864822,
|
| 973 |
-
"grad_norm": 0.5186671614646912,
|
| 974 |
-
"learning_rate": 1.663254834890146e-05,
|
| 975 |
-
"loss": 0.1895,
|
| 976 |
-
"step": 6900
|
| 977 |
-
}
|
| 978 |
-
],
|
| 979 |
-
"logging_steps": 50,
|
| 980 |
-
"max_steps": 38505,
|
| 981 |
-
"num_input_tokens_seen": 0,
|
| 982 |
-
"num_train_epochs": 5,
|
| 983 |
-
"save_steps": 386,
|
| 984 |
-
"stateful_callbacks": {
|
| 985 |
-
"TrainerControl": {
|
| 986 |
-
"args": {
|
| 987 |
-
"should_epoch_stop": false,
|
| 988 |
-
"should_evaluate": false,
|
| 989 |
-
"should_log": false,
|
| 990 |
-
"should_save": true,
|
| 991 |
-
"should_training_stop": false
|
| 992 |
-
},
|
| 993 |
-
"attributes": {}
|
| 994 |
-
}
|
| 995 |
-
},
|
| 996 |
-
"total_flos": 5.810355262497751e+17,
|
| 997 |
-
"train_batch_size": 40,
|
| 998 |
-
"trial_name": null,
|
| 999 |
-
"trial_params": null
|
| 1000 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
checkpoint-7334/added_tokens.json
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"[MASK]": 128000
|
| 3 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
checkpoint-7334/config.json
DELETED
|
@@ -1,69 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"architectures": [
|
| 3 |
-
"DebertaV2ForTokenClassification"
|
| 4 |
-
],
|
| 5 |
-
"attention_probs_dropout_prob": 0.1,
|
| 6 |
-
"hidden_act": "gelu",
|
| 7 |
-
"hidden_dropout_prob": 0.1,
|
| 8 |
-
"hidden_size": 768,
|
| 9 |
-
"id2label": {
|
| 10 |
-
"0": "O",
|
| 11 |
-
"1": "B-CHA",
|
| 12 |
-
"2": "I-CHA",
|
| 13 |
-
"3": "B-LOC",
|
| 14 |
-
"4": "I-LOC",
|
| 15 |
-
"5": "B-FAC",
|
| 16 |
-
"6": "I-FAC",
|
| 17 |
-
"7": "B-OBJ",
|
| 18 |
-
"8": "I-OBJ",
|
| 19 |
-
"9": "B-EVT",
|
| 20 |
-
"10": "I-EVT",
|
| 21 |
-
"11": "B-ORG",
|
| 22 |
-
"12": "I-ORG",
|
| 23 |
-
"13": "B-MISC",
|
| 24 |
-
"14": "I-MISC"
|
| 25 |
-
},
|
| 26 |
-
"initializer_range": 0.02,
|
| 27 |
-
"intermediate_size": 3072,
|
| 28 |
-
"label2id": {
|
| 29 |
-
"B-CHA": 1,
|
| 30 |
-
"B-EVT": 9,
|
| 31 |
-
"B-FAC": 5,
|
| 32 |
-
"B-LOC": 3,
|
| 33 |
-
"B-MISC": 13,
|
| 34 |
-
"B-OBJ": 7,
|
| 35 |
-
"B-ORG": 11,
|
| 36 |
-
"I-CHA": 2,
|
| 37 |
-
"I-EVT": 10,
|
| 38 |
-
"I-FAC": 6,
|
| 39 |
-
"I-LOC": 4,
|
| 40 |
-
"I-MISC": 14,
|
| 41 |
-
"I-OBJ": 8,
|
| 42 |
-
"I-ORG": 12,
|
| 43 |
-
"O": 0
|
| 44 |
-
},
|
| 45 |
-
"layer_norm_eps": 1e-07,
|
| 46 |
-
"legacy": true,
|
| 47 |
-
"max_position_embeddings": 512,
|
| 48 |
-
"max_relative_positions": -1,
|
| 49 |
-
"model_type": "deberta-v2",
|
| 50 |
-
"norm_rel_ebd": "layer_norm",
|
| 51 |
-
"num_attention_heads": 12,
|
| 52 |
-
"num_hidden_layers": 12,
|
| 53 |
-
"pad_token_id": 0,
|
| 54 |
-
"pooler_dropout": 0,
|
| 55 |
-
"pooler_hidden_act": "gelu",
|
| 56 |
-
"pooler_hidden_size": 768,
|
| 57 |
-
"pos_att_type": [
|
| 58 |
-
"p2c",
|
| 59 |
-
"c2p"
|
| 60 |
-
],
|
| 61 |
-
"position_biased_input": false,
|
| 62 |
-
"position_buckets": 256,
|
| 63 |
-
"relative_attention": true,
|
| 64 |
-
"share_att_key": true,
|
| 65 |
-
"torch_dtype": "float32",
|
| 66 |
-
"transformers_version": "4.55.4",
|
| 67 |
-
"type_vocab_size": 0,
|
| 68 |
-
"vocab_size": 128100
|
| 69 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
checkpoint-7334/special_tokens_map.json
DELETED
|
@@ -1,15 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"bos_token": "[CLS]",
|
| 3 |
-
"cls_token": "[CLS]",
|
| 4 |
-
"eos_token": "[SEP]",
|
| 5 |
-
"mask_token": "[MASK]",
|
| 6 |
-
"pad_token": "[PAD]",
|
| 7 |
-
"sep_token": "[SEP]",
|
| 8 |
-
"unk_token": {
|
| 9 |
-
"content": "[UNK]",
|
| 10 |
-
"lstrip": false,
|
| 11 |
-
"normalized": true,
|
| 12 |
-
"rstrip": false,
|
| 13 |
-
"single_word": false
|
| 14 |
-
}
|
| 15 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
checkpoint-7334/tokenizer.json
DELETED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-7334/tokenizer_config.json
DELETED
|
@@ -1,59 +0,0 @@
|
|
| 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": "[CLS]",
|
| 13 |
-
"lstrip": false,
|
| 14 |
-
"normalized": false,
|
| 15 |
-
"rstrip": false,
|
| 16 |
-
"single_word": false,
|
| 17 |
-
"special": true
|
| 18 |
-
},
|
| 19 |
-
"2": {
|
| 20 |
-
"content": "[SEP]",
|
| 21 |
-
"lstrip": false,
|
| 22 |
-
"normalized": false,
|
| 23 |
-
"rstrip": false,
|
| 24 |
-
"single_word": false,
|
| 25 |
-
"special": true
|
| 26 |
-
},
|
| 27 |
-
"3": {
|
| 28 |
-
"content": "[UNK]",
|
| 29 |
-
"lstrip": false,
|
| 30 |
-
"normalized": true,
|
| 31 |
-
"rstrip": false,
|
| 32 |
-
"single_word": false,
|
| 33 |
-
"special": true
|
| 34 |
-
},
|
| 35 |
-
"128000": {
|
| 36 |
-
"content": "[MASK]",
|
| 37 |
-
"lstrip": false,
|
| 38 |
-
"normalized": false,
|
| 39 |
-
"rstrip": false,
|
| 40 |
-
"single_word": false,
|
| 41 |
-
"special": true
|
| 42 |
-
}
|
| 43 |
-
},
|
| 44 |
-
"bos_token": "[CLS]",
|
| 45 |
-
"clean_up_tokenization_spaces": false,
|
| 46 |
-
"cls_token": "[CLS]",
|
| 47 |
-
"do_lower_case": false,
|
| 48 |
-
"eos_token": "[SEP]",
|
| 49 |
-
"extra_special_tokens": {},
|
| 50 |
-
"mask_token": "[MASK]",
|
| 51 |
-
"model_max_length": 1000000000000000019884624838656,
|
| 52 |
-
"pad_token": "[PAD]",
|
| 53 |
-
"sep_token": "[SEP]",
|
| 54 |
-
"sp_model_kwargs": {},
|
| 55 |
-
"split_by_punct": false,
|
| 56 |
-
"tokenizer_class": "DebertaV2Tokenizer",
|
| 57 |
-
"unk_token": "[UNK]",
|
| 58 |
-
"vocab_type": "spm"
|
| 59 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
checkpoint-7334/trainer_state.json
DELETED
|
@@ -1,1056 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"best_global_step": null,
|
| 3 |
-
"best_metric": null,
|
| 4 |
-
"best_model_checkpoint": null,
|
| 5 |
-
"epoch": 0.9523438514478639,
|
| 6 |
-
"eval_steps": 500,
|
| 7 |
-
"global_step": 7334,
|
| 8 |
-
"is_hyper_param_search": false,
|
| 9 |
-
"is_local_process_zero": true,
|
| 10 |
-
"is_world_process_zero": true,
|
| 11 |
-
"log_history": [
|
| 12 |
-
{
|
| 13 |
-
"epoch": 0.006492663290481756,
|
| 14 |
-
"grad_norm": 23.6250057220459,
|
| 15 |
-
"learning_rate": 1.9600000000000003e-06,
|
| 16 |
-
"loss": 3.3492,
|
| 17 |
-
"step": 50
|
| 18 |
-
},
|
| 19 |
-
{
|
| 20 |
-
"epoch": 0.012985326580963512,
|
| 21 |
-
"grad_norm": 1.5745655298233032,
|
| 22 |
-
"learning_rate": 3.96e-06,
|
| 23 |
-
"loss": 1.3837,
|
| 24 |
-
"step": 100
|
| 25 |
-
},
|
| 26 |
-
{
|
| 27 |
-
"epoch": 0.019477989871445268,
|
| 28 |
-
"grad_norm": 1.3241935968399048,
|
| 29 |
-
"learning_rate": 5.9600000000000005e-06,
|
| 30 |
-
"loss": 0.6673,
|
| 31 |
-
"step": 150
|
| 32 |
-
},
|
| 33 |
-
{
|
| 34 |
-
"epoch": 0.025970653161927024,
|
| 35 |
-
"grad_norm": 1.068557620048523,
|
| 36 |
-
"learning_rate": 7.960000000000002e-06,
|
| 37 |
-
"loss": 0.4426,
|
| 38 |
-
"step": 200
|
| 39 |
-
},
|
| 40 |
-
{
|
| 41 |
-
"epoch": 0.032463316452408776,
|
| 42 |
-
"grad_norm": 1.2953921556472778,
|
| 43 |
-
"learning_rate": 9.960000000000001e-06,
|
| 44 |
-
"loss": 0.363,
|
| 45 |
-
"step": 250
|
| 46 |
-
},
|
| 47 |
-
{
|
| 48 |
-
"epoch": 0.038955979742890535,
|
| 49 |
-
"grad_norm": 1.4753903150558472,
|
| 50 |
-
"learning_rate": 1.196e-05,
|
| 51 |
-
"loss": 0.3397,
|
| 52 |
-
"step": 300
|
| 53 |
-
},
|
| 54 |
-
{
|
| 55 |
-
"epoch": 0.04544864303337229,
|
| 56 |
-
"grad_norm": 1.245597243309021,
|
| 57 |
-
"learning_rate": 1.396e-05,
|
| 58 |
-
"loss": 0.3125,
|
| 59 |
-
"step": 350
|
| 60 |
-
},
|
| 61 |
-
{
|
| 62 |
-
"epoch": 0.05194130632385405,
|
| 63 |
-
"grad_norm": 1.3421887159347534,
|
| 64 |
-
"learning_rate": 1.5960000000000003e-05,
|
| 65 |
-
"loss": 0.3086,
|
| 66 |
-
"step": 400
|
| 67 |
-
},
|
| 68 |
-
{
|
| 69 |
-
"epoch": 0.0584339696143358,
|
| 70 |
-
"grad_norm": 1.7852529287338257,
|
| 71 |
-
"learning_rate": 1.796e-05,
|
| 72 |
-
"loss": 0.2883,
|
| 73 |
-
"step": 450
|
| 74 |
-
},
|
| 75 |
-
{
|
| 76 |
-
"epoch": 0.06492663290481755,
|
| 77 |
-
"grad_norm": 1.086064100265503,
|
| 78 |
-
"learning_rate": 1.9960000000000002e-05,
|
| 79 |
-
"loss": 0.2947,
|
| 80 |
-
"step": 500
|
| 81 |
-
},
|
| 82 |
-
{
|
| 83 |
-
"epoch": 0.07141929619529931,
|
| 84 |
-
"grad_norm": 0.8753414750099182,
|
| 85 |
-
"learning_rate": 1.9974213919221156e-05,
|
| 86 |
-
"loss": 0.2955,
|
| 87 |
-
"step": 550
|
| 88 |
-
},
|
| 89 |
-
{
|
| 90 |
-
"epoch": 0.07791195948578107,
|
| 91 |
-
"grad_norm": 0.7875685095787048,
|
| 92 |
-
"learning_rate": 1.9947901591895804e-05,
|
| 93 |
-
"loss": 0.2773,
|
| 94 |
-
"step": 600
|
| 95 |
-
},
|
| 96 |
-
{
|
| 97 |
-
"epoch": 0.08440462277626282,
|
| 98 |
-
"grad_norm": 1.1609506607055664,
|
| 99 |
-
"learning_rate": 1.9921589264570452e-05,
|
| 100 |
-
"loss": 0.2696,
|
| 101 |
-
"step": 650
|
| 102 |
-
},
|
| 103 |
-
{
|
| 104 |
-
"epoch": 0.09089728606674458,
|
| 105 |
-
"grad_norm": 0.9162603616714478,
|
| 106 |
-
"learning_rate": 1.98952769372451e-05,
|
| 107 |
-
"loss": 0.2667,
|
| 108 |
-
"step": 700
|
| 109 |
-
},
|
| 110 |
-
{
|
| 111 |
-
"epoch": 0.09738994935722634,
|
| 112 |
-
"grad_norm": 1.0398370027542114,
|
| 113 |
-
"learning_rate": 1.986896460991975e-05,
|
| 114 |
-
"loss": 0.277,
|
| 115 |
-
"step": 750
|
| 116 |
-
},
|
| 117 |
-
{
|
| 118 |
-
"epoch": 0.1038826126477081,
|
| 119 |
-
"grad_norm": 0.9404798150062561,
|
| 120 |
-
"learning_rate": 1.9842652282594397e-05,
|
| 121 |
-
"loss": 0.2584,
|
| 122 |
-
"step": 800
|
| 123 |
-
},
|
| 124 |
-
{
|
| 125 |
-
"epoch": 0.11037527593818984,
|
| 126 |
-
"grad_norm": 1.0353909730911255,
|
| 127 |
-
"learning_rate": 1.9816339955269045e-05,
|
| 128 |
-
"loss": 0.2533,
|
| 129 |
-
"step": 850
|
| 130 |
-
},
|
| 131 |
-
{
|
| 132 |
-
"epoch": 0.1168679392286716,
|
| 133 |
-
"grad_norm": 1.485606074333191,
|
| 134 |
-
"learning_rate": 1.9790027627943694e-05,
|
| 135 |
-
"loss": 0.2506,
|
| 136 |
-
"step": 900
|
| 137 |
-
},
|
| 138 |
-
{
|
| 139 |
-
"epoch": 0.12336060251915336,
|
| 140 |
-
"grad_norm": 0.9874151945114136,
|
| 141 |
-
"learning_rate": 1.9763715300618342e-05,
|
| 142 |
-
"loss": 0.2515,
|
| 143 |
-
"step": 950
|
| 144 |
-
},
|
| 145 |
-
{
|
| 146 |
-
"epoch": 0.1298532658096351,
|
| 147 |
-
"grad_norm": 1.036007285118103,
|
| 148 |
-
"learning_rate": 1.973740297329299e-05,
|
| 149 |
-
"loss": 0.244,
|
| 150 |
-
"step": 1000
|
| 151 |
-
},
|
| 152 |
-
{
|
| 153 |
-
"epoch": 0.13634592910011686,
|
| 154 |
-
"grad_norm": 0.946956217288971,
|
| 155 |
-
"learning_rate": 1.9711090645967635e-05,
|
| 156 |
-
"loss": 0.2518,
|
| 157 |
-
"step": 1050
|
| 158 |
-
},
|
| 159 |
-
{
|
| 160 |
-
"epoch": 0.14283859239059862,
|
| 161 |
-
"grad_norm": 0.9096735119819641,
|
| 162 |
-
"learning_rate": 1.9684778318642287e-05,
|
| 163 |
-
"loss": 0.2543,
|
| 164 |
-
"step": 1100
|
| 165 |
-
},
|
| 166 |
-
{
|
| 167 |
-
"epoch": 0.14933125568108038,
|
| 168 |
-
"grad_norm": 0.9371875524520874,
|
| 169 |
-
"learning_rate": 1.965846599131693e-05,
|
| 170 |
-
"loss": 0.2605,
|
| 171 |
-
"step": 1150
|
| 172 |
-
},
|
| 173 |
-
{
|
| 174 |
-
"epoch": 0.15582391897156214,
|
| 175 |
-
"grad_norm": 0.8072580695152283,
|
| 176 |
-
"learning_rate": 1.9632153663991583e-05,
|
| 177 |
-
"loss": 0.2423,
|
| 178 |
-
"step": 1200
|
| 179 |
-
},
|
| 180 |
-
{
|
| 181 |
-
"epoch": 0.1623165822620439,
|
| 182 |
-
"grad_norm": 0.6907565593719482,
|
| 183 |
-
"learning_rate": 1.9605841336666228e-05,
|
| 184 |
-
"loss": 0.2315,
|
| 185 |
-
"step": 1250
|
| 186 |
-
},
|
| 187 |
-
{
|
| 188 |
-
"epoch": 0.16880924555252563,
|
| 189 |
-
"grad_norm": 0.6676374673843384,
|
| 190 |
-
"learning_rate": 1.957952900934088e-05,
|
| 191 |
-
"loss": 0.2363,
|
| 192 |
-
"step": 1300
|
| 193 |
-
},
|
| 194 |
-
{
|
| 195 |
-
"epoch": 0.1753019088430074,
|
| 196 |
-
"grad_norm": 0.7281391024589539,
|
| 197 |
-
"learning_rate": 1.9553216682015525e-05,
|
| 198 |
-
"loss": 0.2399,
|
| 199 |
-
"step": 1350
|
| 200 |
-
},
|
| 201 |
-
{
|
| 202 |
-
"epoch": 0.18179457213348915,
|
| 203 |
-
"grad_norm": 0.8105588555335999,
|
| 204 |
-
"learning_rate": 1.9526904354690177e-05,
|
| 205 |
-
"loss": 0.2448,
|
| 206 |
-
"step": 1400
|
| 207 |
-
},
|
| 208 |
-
{
|
| 209 |
-
"epoch": 0.1882872354239709,
|
| 210 |
-
"grad_norm": 0.7466333508491516,
|
| 211 |
-
"learning_rate": 1.950059202736482e-05,
|
| 212 |
-
"loss": 0.2388,
|
| 213 |
-
"step": 1450
|
| 214 |
-
},
|
| 215 |
-
{
|
| 216 |
-
"epoch": 0.19477989871445267,
|
| 217 |
-
"grad_norm": 0.5949985384941101,
|
| 218 |
-
"learning_rate": 1.947427970003947e-05,
|
| 219 |
-
"loss": 0.233,
|
| 220 |
-
"step": 1500
|
| 221 |
-
},
|
| 222 |
-
{
|
| 223 |
-
"epoch": 0.20127256200493443,
|
| 224 |
-
"grad_norm": 0.7439960241317749,
|
| 225 |
-
"learning_rate": 1.9447967372714118e-05,
|
| 226 |
-
"loss": 0.2448,
|
| 227 |
-
"step": 1550
|
| 228 |
-
},
|
| 229 |
-
{
|
| 230 |
-
"epoch": 0.2077652252954162,
|
| 231 |
-
"grad_norm": 0.9508784413337708,
|
| 232 |
-
"learning_rate": 1.9421655045388766e-05,
|
| 233 |
-
"loss": 0.2353,
|
| 234 |
-
"step": 1600
|
| 235 |
-
},
|
| 236 |
-
{
|
| 237 |
-
"epoch": 0.21425788858589795,
|
| 238 |
-
"grad_norm": 0.8785332441329956,
|
| 239 |
-
"learning_rate": 1.9395342718063415e-05,
|
| 240 |
-
"loss": 0.2249,
|
| 241 |
-
"step": 1650
|
| 242 |
-
},
|
| 243 |
-
{
|
| 244 |
-
"epoch": 0.22075055187637968,
|
| 245 |
-
"grad_norm": 0.7568134069442749,
|
| 246 |
-
"learning_rate": 1.9369030390738063e-05,
|
| 247 |
-
"loss": 0.2307,
|
| 248 |
-
"step": 1700
|
| 249 |
-
},
|
| 250 |
-
{
|
| 251 |
-
"epoch": 0.22724321516686144,
|
| 252 |
-
"grad_norm": 0.6832641363143921,
|
| 253 |
-
"learning_rate": 1.934271806341271e-05,
|
| 254 |
-
"loss": 0.2263,
|
| 255 |
-
"step": 1750
|
| 256 |
-
},
|
| 257 |
-
{
|
| 258 |
-
"epoch": 0.2337358784573432,
|
| 259 |
-
"grad_norm": 0.9721575379371643,
|
| 260 |
-
"learning_rate": 1.931640573608736e-05,
|
| 261 |
-
"loss": 0.2287,
|
| 262 |
-
"step": 1800
|
| 263 |
-
},
|
| 264 |
-
{
|
| 265 |
-
"epoch": 0.24022854174782496,
|
| 266 |
-
"grad_norm": 0.620299220085144,
|
| 267 |
-
"learning_rate": 1.9290093408762008e-05,
|
| 268 |
-
"loss": 0.226,
|
| 269 |
-
"step": 1850
|
| 270 |
-
},
|
| 271 |
-
{
|
| 272 |
-
"epoch": 0.24672120503830672,
|
| 273 |
-
"grad_norm": 0.6656680703163147,
|
| 274 |
-
"learning_rate": 1.9263781081436656e-05,
|
| 275 |
-
"loss": 0.2223,
|
| 276 |
-
"step": 1900
|
| 277 |
-
},
|
| 278 |
-
{
|
| 279 |
-
"epoch": 0.25321386832878845,
|
| 280 |
-
"grad_norm": 0.8536450266838074,
|
| 281 |
-
"learning_rate": 1.9237468754111304e-05,
|
| 282 |
-
"loss": 0.2332,
|
| 283 |
-
"step": 1950
|
| 284 |
-
},
|
| 285 |
-
{
|
| 286 |
-
"epoch": 0.2597065316192702,
|
| 287 |
-
"grad_norm": 0.9754176139831543,
|
| 288 |
-
"learning_rate": 1.921115642678595e-05,
|
| 289 |
-
"loss": 0.2211,
|
| 290 |
-
"step": 2000
|
| 291 |
-
},
|
| 292 |
-
{
|
| 293 |
-
"epoch": 0.26619919490975197,
|
| 294 |
-
"grad_norm": 0.6166796684265137,
|
| 295 |
-
"learning_rate": 1.9184844099460598e-05,
|
| 296 |
-
"loss": 0.2201,
|
| 297 |
-
"step": 2050
|
| 298 |
-
},
|
| 299 |
-
{
|
| 300 |
-
"epoch": 0.2726918582002337,
|
| 301 |
-
"grad_norm": 0.8182518482208252,
|
| 302 |
-
"learning_rate": 1.9158531772135246e-05,
|
| 303 |
-
"loss": 0.2213,
|
| 304 |
-
"step": 2100
|
| 305 |
-
},
|
| 306 |
-
{
|
| 307 |
-
"epoch": 0.2791845214907155,
|
| 308 |
-
"grad_norm": 0.8245829939842224,
|
| 309 |
-
"learning_rate": 1.9132219444809894e-05,
|
| 310 |
-
"loss": 0.2208,
|
| 311 |
-
"step": 2150
|
| 312 |
-
},
|
| 313 |
-
{
|
| 314 |
-
"epoch": 0.28567718478119725,
|
| 315 |
-
"grad_norm": 0.6423109173774719,
|
| 316 |
-
"learning_rate": 1.9105907117484542e-05,
|
| 317 |
-
"loss": 0.2308,
|
| 318 |
-
"step": 2200
|
| 319 |
-
},
|
| 320 |
-
{
|
| 321 |
-
"epoch": 0.292169848071679,
|
| 322 |
-
"grad_norm": 0.7164100408554077,
|
| 323 |
-
"learning_rate": 1.907959479015919e-05,
|
| 324 |
-
"loss": 0.2193,
|
| 325 |
-
"step": 2250
|
| 326 |
-
},
|
| 327 |
-
{
|
| 328 |
-
"epoch": 0.29866251136216077,
|
| 329 |
-
"grad_norm": 0.7910536527633667,
|
| 330 |
-
"learning_rate": 1.905328246283384e-05,
|
| 331 |
-
"loss": 0.2292,
|
| 332 |
-
"step": 2300
|
| 333 |
-
},
|
| 334 |
-
{
|
| 335 |
-
"epoch": 0.3051551746526425,
|
| 336 |
-
"grad_norm": 0.9699934720993042,
|
| 337 |
-
"learning_rate": 1.9026970135508487e-05,
|
| 338 |
-
"loss": 0.2175,
|
| 339 |
-
"step": 2350
|
| 340 |
-
},
|
| 341 |
-
{
|
| 342 |
-
"epoch": 0.3116478379431243,
|
| 343 |
-
"grad_norm": 1.0665416717529297,
|
| 344 |
-
"learning_rate": 1.9000657808183136e-05,
|
| 345 |
-
"loss": 0.2252,
|
| 346 |
-
"step": 2400
|
| 347 |
-
},
|
| 348 |
-
{
|
| 349 |
-
"epoch": 0.31814050123360604,
|
| 350 |
-
"grad_norm": 0.9915699362754822,
|
| 351 |
-
"learning_rate": 1.8974345480857784e-05,
|
| 352 |
-
"loss": 0.215,
|
| 353 |
-
"step": 2450
|
| 354 |
-
},
|
| 355 |
-
{
|
| 356 |
-
"epoch": 0.3246331645240878,
|
| 357 |
-
"grad_norm": 0.7021859288215637,
|
| 358 |
-
"learning_rate": 1.8948033153532432e-05,
|
| 359 |
-
"loss": 0.2123,
|
| 360 |
-
"step": 2500
|
| 361 |
-
},
|
| 362 |
-
{
|
| 363 |
-
"epoch": 0.33112582781456956,
|
| 364 |
-
"grad_norm": 0.6402600407600403,
|
| 365 |
-
"learning_rate": 1.892172082620708e-05,
|
| 366 |
-
"loss": 0.2176,
|
| 367 |
-
"step": 2550
|
| 368 |
-
},
|
| 369 |
-
{
|
| 370 |
-
"epoch": 0.33761849110505127,
|
| 371 |
-
"grad_norm": 0.9925078749656677,
|
| 372 |
-
"learning_rate": 1.8895408498881725e-05,
|
| 373 |
-
"loss": 0.223,
|
| 374 |
-
"step": 2600
|
| 375 |
-
},
|
| 376 |
-
{
|
| 377 |
-
"epoch": 0.344111154395533,
|
| 378 |
-
"grad_norm": 0.5188687443733215,
|
| 379 |
-
"learning_rate": 1.8869096171556377e-05,
|
| 380 |
-
"loss": 0.2174,
|
| 381 |
-
"step": 2650
|
| 382 |
-
},
|
| 383 |
-
{
|
| 384 |
-
"epoch": 0.3506038176860148,
|
| 385 |
-
"grad_norm": 0.7427687644958496,
|
| 386 |
-
"learning_rate": 1.8842783844231022e-05,
|
| 387 |
-
"loss": 0.2181,
|
| 388 |
-
"step": 2700
|
| 389 |
-
},
|
| 390 |
-
{
|
| 391 |
-
"epoch": 0.35709648097649654,
|
| 392 |
-
"grad_norm": 0.6055501103401184,
|
| 393 |
-
"learning_rate": 1.8816471516905674e-05,
|
| 394 |
-
"loss": 0.2097,
|
| 395 |
-
"step": 2750
|
| 396 |
-
},
|
| 397 |
-
{
|
| 398 |
-
"epoch": 0.3635891442669783,
|
| 399 |
-
"grad_norm": 0.6529932022094727,
|
| 400 |
-
"learning_rate": 1.879015918958032e-05,
|
| 401 |
-
"loss": 0.2133,
|
| 402 |
-
"step": 2800
|
| 403 |
-
},
|
| 404 |
-
{
|
| 405 |
-
"epoch": 0.37008180755746006,
|
| 406 |
-
"grad_norm": 0.9499914646148682,
|
| 407 |
-
"learning_rate": 1.8763846862254967e-05,
|
| 408 |
-
"loss": 0.2189,
|
| 409 |
-
"step": 2850
|
| 410 |
-
},
|
| 411 |
-
{
|
| 412 |
-
"epoch": 0.3765744708479418,
|
| 413 |
-
"grad_norm": 0.8217771053314209,
|
| 414 |
-
"learning_rate": 1.8737534534929615e-05,
|
| 415 |
-
"loss": 0.2172,
|
| 416 |
-
"step": 2900
|
| 417 |
-
},
|
| 418 |
-
{
|
| 419 |
-
"epoch": 0.3830671341384236,
|
| 420 |
-
"grad_norm": 1.031895399093628,
|
| 421 |
-
"learning_rate": 1.8711222207604263e-05,
|
| 422 |
-
"loss": 0.2171,
|
| 423 |
-
"step": 2950
|
| 424 |
-
},
|
| 425 |
-
{
|
| 426 |
-
"epoch": 0.38955979742890534,
|
| 427 |
-
"grad_norm": 0.7179540395736694,
|
| 428 |
-
"learning_rate": 1.8684909880278912e-05,
|
| 429 |
-
"loss": 0.2179,
|
| 430 |
-
"step": 3000
|
| 431 |
-
},
|
| 432 |
-
{
|
| 433 |
-
"epoch": 0.3960524607193871,
|
| 434 |
-
"grad_norm": 0.8115535974502563,
|
| 435 |
-
"learning_rate": 1.865859755295356e-05,
|
| 436 |
-
"loss": 0.2131,
|
| 437 |
-
"step": 3050
|
| 438 |
-
},
|
| 439 |
-
{
|
| 440 |
-
"epoch": 0.40254512400986886,
|
| 441 |
-
"grad_norm": 0.7659513354301453,
|
| 442 |
-
"learning_rate": 1.863228522562821e-05,
|
| 443 |
-
"loss": 0.2189,
|
| 444 |
-
"step": 3100
|
| 445 |
-
},
|
| 446 |
-
{
|
| 447 |
-
"epoch": 0.4090377873003506,
|
| 448 |
-
"grad_norm": 0.8645811080932617,
|
| 449 |
-
"learning_rate": 1.8605972898302857e-05,
|
| 450 |
-
"loss": 0.2453,
|
| 451 |
-
"step": 3150
|
| 452 |
-
},
|
| 453 |
-
{
|
| 454 |
-
"epoch": 0.4155304505908324,
|
| 455 |
-
"grad_norm": 0.8571991920471191,
|
| 456 |
-
"learning_rate": 1.8579660570977505e-05,
|
| 457 |
-
"loss": 0.2076,
|
| 458 |
-
"step": 3200
|
| 459 |
-
},
|
| 460 |
-
{
|
| 461 |
-
"epoch": 0.42202311388131414,
|
| 462 |
-
"grad_norm": 1.222774863243103,
|
| 463 |
-
"learning_rate": 1.8553348243652153e-05,
|
| 464 |
-
"loss": 0.2094,
|
| 465 |
-
"step": 3250
|
| 466 |
-
},
|
| 467 |
-
{
|
| 468 |
-
"epoch": 0.4285157771717959,
|
| 469 |
-
"grad_norm": 0.9631436467170715,
|
| 470 |
-
"learning_rate": 1.85270359163268e-05,
|
| 471 |
-
"loss": 0.2123,
|
| 472 |
-
"step": 3300
|
| 473 |
-
},
|
| 474 |
-
{
|
| 475 |
-
"epoch": 0.4350084404622776,
|
| 476 |
-
"grad_norm": 0.5590534806251526,
|
| 477 |
-
"learning_rate": 1.8500723589001446e-05,
|
| 478 |
-
"loss": 0.2028,
|
| 479 |
-
"step": 3350
|
| 480 |
-
},
|
| 481 |
-
{
|
| 482 |
-
"epoch": 0.44150110375275936,
|
| 483 |
-
"grad_norm": 0.7615678310394287,
|
| 484 |
-
"learning_rate": 1.8474411261676098e-05,
|
| 485 |
-
"loss": 0.2125,
|
| 486 |
-
"step": 3400
|
| 487 |
-
},
|
| 488 |
-
{
|
| 489 |
-
"epoch": 0.4479937670432411,
|
| 490 |
-
"grad_norm": 0.7124540209770203,
|
| 491 |
-
"learning_rate": 1.8448098934350743e-05,
|
| 492 |
-
"loss": 0.206,
|
| 493 |
-
"step": 3450
|
| 494 |
-
},
|
| 495 |
-
{
|
| 496 |
-
"epoch": 0.4544864303337229,
|
| 497 |
-
"grad_norm": 0.8492136001586914,
|
| 498 |
-
"learning_rate": 1.8421786607025395e-05,
|
| 499 |
-
"loss": 0.2054,
|
| 500 |
-
"step": 3500
|
| 501 |
-
},
|
| 502 |
-
{
|
| 503 |
-
"epoch": 0.46097909362420464,
|
| 504 |
-
"grad_norm": 1.0388613939285278,
|
| 505 |
-
"learning_rate": 1.839547427970004e-05,
|
| 506 |
-
"loss": 0.2117,
|
| 507 |
-
"step": 3550
|
| 508 |
-
},
|
| 509 |
-
{
|
| 510 |
-
"epoch": 0.4674717569146864,
|
| 511 |
-
"grad_norm": 0.8087108731269836,
|
| 512 |
-
"learning_rate": 1.836916195237469e-05,
|
| 513 |
-
"loss": 0.2192,
|
| 514 |
-
"step": 3600
|
| 515 |
-
},
|
| 516 |
-
{
|
| 517 |
-
"epoch": 0.47396442020516816,
|
| 518 |
-
"grad_norm": 0.7446919679641724,
|
| 519 |
-
"learning_rate": 1.8342849625049336e-05,
|
| 520 |
-
"loss": 0.2112,
|
| 521 |
-
"step": 3650
|
| 522 |
-
},
|
| 523 |
-
{
|
| 524 |
-
"epoch": 0.4804570834956499,
|
| 525 |
-
"grad_norm": 0.6669062972068787,
|
| 526 |
-
"learning_rate": 1.8316537297723984e-05,
|
| 527 |
-
"loss": 0.2167,
|
| 528 |
-
"step": 3700
|
| 529 |
-
},
|
| 530 |
-
{
|
| 531 |
-
"epoch": 0.4869497467861317,
|
| 532 |
-
"grad_norm": 0.519498884677887,
|
| 533 |
-
"learning_rate": 1.8290224970398633e-05,
|
| 534 |
-
"loss": 0.2069,
|
| 535 |
-
"step": 3750
|
| 536 |
-
},
|
| 537 |
-
{
|
| 538 |
-
"epoch": 0.49344241007661344,
|
| 539 |
-
"grad_norm": 0.5735522508621216,
|
| 540 |
-
"learning_rate": 1.826391264307328e-05,
|
| 541 |
-
"loss": 0.2102,
|
| 542 |
-
"step": 3800
|
| 543 |
-
},
|
| 544 |
-
{
|
| 545 |
-
"epoch": 0.4999350733670952,
|
| 546 |
-
"grad_norm": 0.6683741211891174,
|
| 547 |
-
"learning_rate": 1.823760031574793e-05,
|
| 548 |
-
"loss": 0.2141,
|
| 549 |
-
"step": 3850
|
| 550 |
-
},
|
| 551 |
-
{
|
| 552 |
-
"epoch": 0.5064277366575769,
|
| 553 |
-
"grad_norm": 0.6239739060401917,
|
| 554 |
-
"learning_rate": 1.8211287988422578e-05,
|
| 555 |
-
"loss": 0.1975,
|
| 556 |
-
"step": 3900
|
| 557 |
-
},
|
| 558 |
-
{
|
| 559 |
-
"epoch": 0.5129203999480587,
|
| 560 |
-
"grad_norm": 0.6327545046806335,
|
| 561 |
-
"learning_rate": 1.8184975661097226e-05,
|
| 562 |
-
"loss": 0.2165,
|
| 563 |
-
"step": 3950
|
| 564 |
-
},
|
| 565 |
-
{
|
| 566 |
-
"epoch": 0.5194130632385404,
|
| 567 |
-
"grad_norm": 0.6098693609237671,
|
| 568 |
-
"learning_rate": 1.8158663333771874e-05,
|
| 569 |
-
"loss": 0.2034,
|
| 570 |
-
"step": 4000
|
| 571 |
-
},
|
| 572 |
-
{
|
| 573 |
-
"epoch": 0.5259057265290222,
|
| 574 |
-
"grad_norm": 0.7356197237968445,
|
| 575 |
-
"learning_rate": 1.8132351006446522e-05,
|
| 576 |
-
"loss": 0.2041,
|
| 577 |
-
"step": 4050
|
| 578 |
-
},
|
| 579 |
-
{
|
| 580 |
-
"epoch": 0.5323983898195039,
|
| 581 |
-
"grad_norm": 0.7616349458694458,
|
| 582 |
-
"learning_rate": 1.810603867912117e-05,
|
| 583 |
-
"loss": 0.2121,
|
| 584 |
-
"step": 4100
|
| 585 |
-
},
|
| 586 |
-
{
|
| 587 |
-
"epoch": 0.5388910531099858,
|
| 588 |
-
"grad_norm": 0.6561347842216492,
|
| 589 |
-
"learning_rate": 1.807972635179582e-05,
|
| 590 |
-
"loss": 0.2002,
|
| 591 |
-
"step": 4150
|
| 592 |
-
},
|
| 593 |
-
{
|
| 594 |
-
"epoch": 0.5453837164004675,
|
| 595 |
-
"grad_norm": 0.6744963526725769,
|
| 596 |
-
"learning_rate": 1.8053414024470467e-05,
|
| 597 |
-
"loss": 0.2029,
|
| 598 |
-
"step": 4200
|
| 599 |
-
},
|
| 600 |
-
{
|
| 601 |
-
"epoch": 0.5518763796909493,
|
| 602 |
-
"grad_norm": 0.7836194038391113,
|
| 603 |
-
"learning_rate": 1.8027101697145112e-05,
|
| 604 |
-
"loss": 0.2055,
|
| 605 |
-
"step": 4250
|
| 606 |
-
},
|
| 607 |
-
{
|
| 608 |
-
"epoch": 0.558369042981431,
|
| 609 |
-
"grad_norm": 0.6983553767204285,
|
| 610 |
-
"learning_rate": 1.800078936981976e-05,
|
| 611 |
-
"loss": 0.2139,
|
| 612 |
-
"step": 4300
|
| 613 |
-
},
|
| 614 |
-
{
|
| 615 |
-
"epoch": 0.5648617062719128,
|
| 616 |
-
"grad_norm": 0.6315485239028931,
|
| 617 |
-
"learning_rate": 1.797447704249441e-05,
|
| 618 |
-
"loss": 0.1978,
|
| 619 |
-
"step": 4350
|
| 620 |
-
},
|
| 621 |
-
{
|
| 622 |
-
"epoch": 0.5713543695623945,
|
| 623 |
-
"grad_norm": 0.5928835868835449,
|
| 624 |
-
"learning_rate": 1.7948164715169057e-05,
|
| 625 |
-
"loss": 0.2027,
|
| 626 |
-
"step": 4400
|
| 627 |
-
},
|
| 628 |
-
{
|
| 629 |
-
"epoch": 0.5778470328528762,
|
| 630 |
-
"grad_norm": 0.8622831702232361,
|
| 631 |
-
"learning_rate": 1.7921852387843705e-05,
|
| 632 |
-
"loss": 0.2104,
|
| 633 |
-
"step": 4450
|
| 634 |
-
},
|
| 635 |
-
{
|
| 636 |
-
"epoch": 0.584339696143358,
|
| 637 |
-
"grad_norm": 0.7967308759689331,
|
| 638 |
-
"learning_rate": 1.7895540060518354e-05,
|
| 639 |
-
"loss": 0.2328,
|
| 640 |
-
"step": 4500
|
| 641 |
-
},
|
| 642 |
-
{
|
| 643 |
-
"epoch": 0.5908323594338397,
|
| 644 |
-
"grad_norm": 0.8289620280265808,
|
| 645 |
-
"learning_rate": 1.7869227733193002e-05,
|
| 646 |
-
"loss": 0.2099,
|
| 647 |
-
"step": 4550
|
| 648 |
-
},
|
| 649 |
-
{
|
| 650 |
-
"epoch": 0.5973250227243215,
|
| 651 |
-
"grad_norm": 0.7185404300689697,
|
| 652 |
-
"learning_rate": 1.784291540586765e-05,
|
| 653 |
-
"loss": 0.2045,
|
| 654 |
-
"step": 4600
|
| 655 |
-
},
|
| 656 |
-
{
|
| 657 |
-
"epoch": 0.6038176860148032,
|
| 658 |
-
"grad_norm": 0.6304630637168884,
|
| 659 |
-
"learning_rate": 1.78166030785423e-05,
|
| 660 |
-
"loss": 0.2055,
|
| 661 |
-
"step": 4650
|
| 662 |
-
},
|
| 663 |
-
{
|
| 664 |
-
"epoch": 0.610310349305285,
|
| 665 |
-
"grad_norm": 0.614983856678009,
|
| 666 |
-
"learning_rate": 1.7790290751216947e-05,
|
| 667 |
-
"loss": 0.1979,
|
| 668 |
-
"step": 4700
|
| 669 |
-
},
|
| 670 |
-
{
|
| 671 |
-
"epoch": 0.6168030125957668,
|
| 672 |
-
"grad_norm": 0.568352460861206,
|
| 673 |
-
"learning_rate": 1.7763978423891595e-05,
|
| 674 |
-
"loss": 0.197,
|
| 675 |
-
"step": 4750
|
| 676 |
-
},
|
| 677 |
-
{
|
| 678 |
-
"epoch": 0.6232956758862486,
|
| 679 |
-
"grad_norm": 0.8283701539039612,
|
| 680 |
-
"learning_rate": 1.773766609656624e-05,
|
| 681 |
-
"loss": 0.1986,
|
| 682 |
-
"step": 4800
|
| 683 |
-
},
|
| 684 |
-
{
|
| 685 |
-
"epoch": 0.6297883391767303,
|
| 686 |
-
"grad_norm": 0.888967752456665,
|
| 687 |
-
"learning_rate": 1.7711353769240892e-05,
|
| 688 |
-
"loss": 0.2271,
|
| 689 |
-
"step": 4850
|
| 690 |
-
},
|
| 691 |
-
{
|
| 692 |
-
"epoch": 0.6362810024672121,
|
| 693 |
-
"grad_norm": 0.5292450189590454,
|
| 694 |
-
"learning_rate": 1.7685041441915537e-05,
|
| 695 |
-
"loss": 0.1971,
|
| 696 |
-
"step": 4900
|
| 697 |
-
},
|
| 698 |
-
{
|
| 699 |
-
"epoch": 0.6427736657576938,
|
| 700 |
-
"grad_norm": 0.5274189114570618,
|
| 701 |
-
"learning_rate": 1.765872911459019e-05,
|
| 702 |
-
"loss": 0.1963,
|
| 703 |
-
"step": 4950
|
| 704 |
-
},
|
| 705 |
-
{
|
| 706 |
-
"epoch": 0.6492663290481756,
|
| 707 |
-
"grad_norm": 0.5699307322502136,
|
| 708 |
-
"learning_rate": 1.7632416787264833e-05,
|
| 709 |
-
"loss": 0.1984,
|
| 710 |
-
"step": 5000
|
| 711 |
-
},
|
| 712 |
-
{
|
| 713 |
-
"epoch": 0.6557589923386573,
|
| 714 |
-
"grad_norm": 0.8022367358207703,
|
| 715 |
-
"learning_rate": 1.7606104459939485e-05,
|
| 716 |
-
"loss": 0.2053,
|
| 717 |
-
"step": 5050
|
| 718 |
-
},
|
| 719 |
-
{
|
| 720 |
-
"epoch": 0.6622516556291391,
|
| 721 |
-
"grad_norm": 0.6432430148124695,
|
| 722 |
-
"learning_rate": 1.757979213261413e-05,
|
| 723 |
-
"loss": 0.1991,
|
| 724 |
-
"step": 5100
|
| 725 |
-
},
|
| 726 |
-
{
|
| 727 |
-
"epoch": 0.6687443189196208,
|
| 728 |
-
"grad_norm": 0.7671304941177368,
|
| 729 |
-
"learning_rate": 1.755347980528878e-05,
|
| 730 |
-
"loss": 0.2064,
|
| 731 |
-
"step": 5150
|
| 732 |
-
},
|
| 733 |
-
{
|
| 734 |
-
"epoch": 0.6752369822101025,
|
| 735 |
-
"grad_norm": 0.5107030272483826,
|
| 736 |
-
"learning_rate": 1.7527167477963426e-05,
|
| 737 |
-
"loss": 0.2049,
|
| 738 |
-
"step": 5200
|
| 739 |
-
},
|
| 740 |
-
{
|
| 741 |
-
"epoch": 0.6817296455005843,
|
| 742 |
-
"grad_norm": 0.7239235639572144,
|
| 743 |
-
"learning_rate": 1.7500855150638075e-05,
|
| 744 |
-
"loss": 0.2001,
|
| 745 |
-
"step": 5250
|
| 746 |
-
},
|
| 747 |
-
{
|
| 748 |
-
"epoch": 0.688222308791066,
|
| 749 |
-
"grad_norm": 0.6116129755973816,
|
| 750 |
-
"learning_rate": 1.7474542823312723e-05,
|
| 751 |
-
"loss": 0.2193,
|
| 752 |
-
"step": 5300
|
| 753 |
-
},
|
| 754 |
-
{
|
| 755 |
-
"epoch": 0.6947149720815479,
|
| 756 |
-
"grad_norm": 0.5425911545753479,
|
| 757 |
-
"learning_rate": 1.744823049598737e-05,
|
| 758 |
-
"loss": 0.2001,
|
| 759 |
-
"step": 5350
|
| 760 |
-
},
|
| 761 |
-
{
|
| 762 |
-
"epoch": 0.7012076353720296,
|
| 763 |
-
"grad_norm": 0.6464748382568359,
|
| 764 |
-
"learning_rate": 1.742191816866202e-05,
|
| 765 |
-
"loss": 0.1963,
|
| 766 |
-
"step": 5400
|
| 767 |
-
},
|
| 768 |
-
{
|
| 769 |
-
"epoch": 0.7077002986625114,
|
| 770 |
-
"grad_norm": 0.8812252879142761,
|
| 771 |
-
"learning_rate": 1.7395605841336668e-05,
|
| 772 |
-
"loss": 0.196,
|
| 773 |
-
"step": 5450
|
| 774 |
-
},
|
| 775 |
-
{
|
| 776 |
-
"epoch": 0.7141929619529931,
|
| 777 |
-
"grad_norm": 0.6928241848945618,
|
| 778 |
-
"learning_rate": 1.7369293514011316e-05,
|
| 779 |
-
"loss": 0.2011,
|
| 780 |
-
"step": 5500
|
| 781 |
-
},
|
| 782 |
-
{
|
| 783 |
-
"epoch": 0.7206856252434749,
|
| 784 |
-
"grad_norm": 0.6892450451850891,
|
| 785 |
-
"learning_rate": 1.7342981186685965e-05,
|
| 786 |
-
"loss": 0.1949,
|
| 787 |
-
"step": 5550
|
| 788 |
-
},
|
| 789 |
-
{
|
| 790 |
-
"epoch": 0.7271782885339566,
|
| 791 |
-
"grad_norm": 0.4782065749168396,
|
| 792 |
-
"learning_rate": 1.7316668859360613e-05,
|
| 793 |
-
"loss": 0.194,
|
| 794 |
-
"step": 5600
|
| 795 |
-
},
|
| 796 |
-
{
|
| 797 |
-
"epoch": 0.7336709518244384,
|
| 798 |
-
"grad_norm": 0.6438505053520203,
|
| 799 |
-
"learning_rate": 1.729035653203526e-05,
|
| 800 |
-
"loss": 0.1967,
|
| 801 |
-
"step": 5650
|
| 802 |
-
},
|
| 803 |
-
{
|
| 804 |
-
"epoch": 0.7401636151149201,
|
| 805 |
-
"grad_norm": 0.5797818899154663,
|
| 806 |
-
"learning_rate": 1.726404420470991e-05,
|
| 807 |
-
"loss": 0.2185,
|
| 808 |
-
"step": 5700
|
| 809 |
-
},
|
| 810 |
-
{
|
| 811 |
-
"epoch": 0.7466562784054019,
|
| 812 |
-
"grad_norm": 0.6884586811065674,
|
| 813 |
-
"learning_rate": 1.7237731877384554e-05,
|
| 814 |
-
"loss": 0.1977,
|
| 815 |
-
"step": 5750
|
| 816 |
-
},
|
| 817 |
-
{
|
| 818 |
-
"epoch": 0.7531489416958836,
|
| 819 |
-
"grad_norm": 0.648883581161499,
|
| 820 |
-
"learning_rate": 1.7211419550059206e-05,
|
| 821 |
-
"loss": 0.1964,
|
| 822 |
-
"step": 5800
|
| 823 |
-
},
|
| 824 |
-
{
|
| 825 |
-
"epoch": 0.7596416049863655,
|
| 826 |
-
"grad_norm": 0.6440086960792542,
|
| 827 |
-
"learning_rate": 1.718510722273385e-05,
|
| 828 |
-
"loss": 0.2014,
|
| 829 |
-
"step": 5850
|
| 830 |
-
},
|
| 831 |
-
{
|
| 832 |
-
"epoch": 0.7661342682768472,
|
| 833 |
-
"grad_norm": 0.5619300007820129,
|
| 834 |
-
"learning_rate": 1.71587948954085e-05,
|
| 835 |
-
"loss": 0.1909,
|
| 836 |
-
"step": 5900
|
| 837 |
-
},
|
| 838 |
-
{
|
| 839 |
-
"epoch": 0.7726269315673289,
|
| 840 |
-
"grad_norm": 0.6859204769134521,
|
| 841 |
-
"learning_rate": 1.7132482568083147e-05,
|
| 842 |
-
"loss": 0.2049,
|
| 843 |
-
"step": 5950
|
| 844 |
-
},
|
| 845 |
-
{
|
| 846 |
-
"epoch": 0.7791195948578107,
|
| 847 |
-
"grad_norm": 0.6132592558860779,
|
| 848 |
-
"learning_rate": 1.7106170240757796e-05,
|
| 849 |
-
"loss": 0.2,
|
| 850 |
-
"step": 6000
|
| 851 |
-
},
|
| 852 |
-
{
|
| 853 |
-
"epoch": 0.7856122581482924,
|
| 854 |
-
"grad_norm": 0.7050901055335999,
|
| 855 |
-
"learning_rate": 1.7079857913432444e-05,
|
| 856 |
-
"loss": 0.189,
|
| 857 |
-
"step": 6050
|
| 858 |
-
},
|
| 859 |
-
{
|
| 860 |
-
"epoch": 0.7921049214387742,
|
| 861 |
-
"grad_norm": 0.6752614974975586,
|
| 862 |
-
"learning_rate": 1.7053545586107092e-05,
|
| 863 |
-
"loss": 0.2248,
|
| 864 |
-
"step": 6100
|
| 865 |
-
},
|
| 866 |
-
{
|
| 867 |
-
"epoch": 0.7985975847292559,
|
| 868 |
-
"grad_norm": 0.7186923623085022,
|
| 869 |
-
"learning_rate": 1.702723325878174e-05,
|
| 870 |
-
"loss": 0.1903,
|
| 871 |
-
"step": 6150
|
| 872 |
-
},
|
| 873 |
-
{
|
| 874 |
-
"epoch": 0.8050902480197377,
|
| 875 |
-
"grad_norm": 0.5991400480270386,
|
| 876 |
-
"learning_rate": 1.700092093145639e-05,
|
| 877 |
-
"loss": 0.197,
|
| 878 |
-
"step": 6200
|
| 879 |
-
},
|
| 880 |
-
{
|
| 881 |
-
"epoch": 0.8115829113102194,
|
| 882 |
-
"grad_norm": 0.9522245526313782,
|
| 883 |
-
"learning_rate": 1.6974608604131037e-05,
|
| 884 |
-
"loss": 0.1962,
|
| 885 |
-
"step": 6250
|
| 886 |
-
},
|
| 887 |
-
{
|
| 888 |
-
"epoch": 0.8180755746007012,
|
| 889 |
-
"grad_norm": 0.8645381927490234,
|
| 890 |
-
"learning_rate": 1.6948296276805686e-05,
|
| 891 |
-
"loss": 0.1901,
|
| 892 |
-
"step": 6300
|
| 893 |
-
},
|
| 894 |
-
{
|
| 895 |
-
"epoch": 0.8245682378911829,
|
| 896 |
-
"grad_norm": 0.5243034958839417,
|
| 897 |
-
"learning_rate": 1.6921983949480334e-05,
|
| 898 |
-
"loss": 0.194,
|
| 899 |
-
"step": 6350
|
| 900 |
-
},
|
| 901 |
-
{
|
| 902 |
-
"epoch": 0.8310609011816648,
|
| 903 |
-
"grad_norm": 0.5842151641845703,
|
| 904 |
-
"learning_rate": 1.6895671622154982e-05,
|
| 905 |
-
"loss": 0.1917,
|
| 906 |
-
"step": 6400
|
| 907 |
-
},
|
| 908 |
-
{
|
| 909 |
-
"epoch": 0.8375535644721465,
|
| 910 |
-
"grad_norm": 0.6111485362052917,
|
| 911 |
-
"learning_rate": 1.6869359294829627e-05,
|
| 912 |
-
"loss": 0.2042,
|
| 913 |
-
"step": 6450
|
| 914 |
-
},
|
| 915 |
-
{
|
| 916 |
-
"epoch": 0.8440462277626283,
|
| 917 |
-
"grad_norm": 0.6515288949012756,
|
| 918 |
-
"learning_rate": 1.684304696750428e-05,
|
| 919 |
-
"loss": 0.2032,
|
| 920 |
-
"step": 6500
|
| 921 |
-
},
|
| 922 |
-
{
|
| 923 |
-
"epoch": 0.85053889105311,
|
| 924 |
-
"grad_norm": 0.7596396207809448,
|
| 925 |
-
"learning_rate": 1.6816734640178924e-05,
|
| 926 |
-
"loss": 0.193,
|
| 927 |
-
"step": 6550
|
| 928 |
-
},
|
| 929 |
-
{
|
| 930 |
-
"epoch": 0.8570315543435918,
|
| 931 |
-
"grad_norm": 0.6944254636764526,
|
| 932 |
-
"learning_rate": 1.6790422312853575e-05,
|
| 933 |
-
"loss": 0.1899,
|
| 934 |
-
"step": 6600
|
| 935 |
-
},
|
| 936 |
-
{
|
| 937 |
-
"epoch": 0.8635242176340735,
|
| 938 |
-
"grad_norm": 0.6190508604049683,
|
| 939 |
-
"learning_rate": 1.676410998552822e-05,
|
| 940 |
-
"loss": 0.1987,
|
| 941 |
-
"step": 6650
|
| 942 |
-
},
|
| 943 |
-
{
|
| 944 |
-
"epoch": 0.8700168809245552,
|
| 945 |
-
"grad_norm": 1.1515477895736694,
|
| 946 |
-
"learning_rate": 1.673779765820287e-05,
|
| 947 |
-
"loss": 0.196,
|
| 948 |
-
"step": 6700
|
| 949 |
-
},
|
| 950 |
-
{
|
| 951 |
-
"epoch": 0.876509544215037,
|
| 952 |
-
"grad_norm": 0.5803254842758179,
|
| 953 |
-
"learning_rate": 1.6711485330877517e-05,
|
| 954 |
-
"loss": 0.1923,
|
| 955 |
-
"step": 6750
|
| 956 |
-
},
|
| 957 |
-
{
|
| 958 |
-
"epoch": 0.8830022075055187,
|
| 959 |
-
"grad_norm": 0.8052871227264404,
|
| 960 |
-
"learning_rate": 1.6685173003552165e-05,
|
| 961 |
-
"loss": 0.1894,
|
| 962 |
-
"step": 6800
|
| 963 |
-
},
|
| 964 |
-
{
|
| 965 |
-
"epoch": 0.8894948707960005,
|
| 966 |
-
"grad_norm": 0.9313941597938538,
|
| 967 |
-
"learning_rate": 1.6658860676226813e-05,
|
| 968 |
-
"loss": 0.1889,
|
| 969 |
-
"step": 6850
|
| 970 |
-
},
|
| 971 |
-
{
|
| 972 |
-
"epoch": 0.8959875340864822,
|
| 973 |
-
"grad_norm": 0.5186671614646912,
|
| 974 |
-
"learning_rate": 1.663254834890146e-05,
|
| 975 |
-
"loss": 0.1895,
|
| 976 |
-
"step": 6900
|
| 977 |
-
},
|
| 978 |
-
{
|
| 979 |
-
"epoch": 0.902480197376964,
|
| 980 |
-
"grad_norm": 0.7533177137374878,
|
| 981 |
-
"learning_rate": 1.660623602157611e-05,
|
| 982 |
-
"loss": 0.1956,
|
| 983 |
-
"step": 6950
|
| 984 |
-
},
|
| 985 |
-
{
|
| 986 |
-
"epoch": 0.9089728606674458,
|
| 987 |
-
"grad_norm": 0.7142027020454407,
|
| 988 |
-
"learning_rate": 1.6579923694250758e-05,
|
| 989 |
-
"loss": 0.1921,
|
| 990 |
-
"step": 7000
|
| 991 |
-
},
|
| 992 |
-
{
|
| 993 |
-
"epoch": 0.9154655239579276,
|
| 994 |
-
"grad_norm": 1.0748203992843628,
|
| 995 |
-
"learning_rate": 1.6553611366925407e-05,
|
| 996 |
-
"loss": 0.1847,
|
| 997 |
-
"step": 7050
|
| 998 |
-
},
|
| 999 |
-
{
|
| 1000 |
-
"epoch": 0.9219581872484093,
|
| 1001 |
-
"grad_norm": 0.5605922341346741,
|
| 1002 |
-
"learning_rate": 1.6527299039600055e-05,
|
| 1003 |
-
"loss": 0.1887,
|
| 1004 |
-
"step": 7100
|
| 1005 |
-
},
|
| 1006 |
-
{
|
| 1007 |
-
"epoch": 0.9284508505388911,
|
| 1008 |
-
"grad_norm": 0.5474116802215576,
|
| 1009 |
-
"learning_rate": 1.6500986712274703e-05,
|
| 1010 |
-
"loss": 0.1948,
|
| 1011 |
-
"step": 7150
|
| 1012 |
-
},
|
| 1013 |
-
{
|
| 1014 |
-
"epoch": 0.9349435138293728,
|
| 1015 |
-
"grad_norm": 0.9507768154144287,
|
| 1016 |
-
"learning_rate": 1.6474674384949348e-05,
|
| 1017 |
-
"loss": 0.2126,
|
| 1018 |
-
"step": 7200
|
| 1019 |
-
},
|
| 1020 |
-
{
|
| 1021 |
-
"epoch": 0.9414361771198546,
|
| 1022 |
-
"grad_norm": 0.694097101688385,
|
| 1023 |
-
"learning_rate": 1.6448362057624e-05,
|
| 1024 |
-
"loss": 0.1904,
|
| 1025 |
-
"step": 7250
|
| 1026 |
-
},
|
| 1027 |
-
{
|
| 1028 |
-
"epoch": 0.9479288404103363,
|
| 1029 |
-
"grad_norm": 0.6153343915939331,
|
| 1030 |
-
"learning_rate": 1.6422049730298645e-05,
|
| 1031 |
-
"loss": 0.1921,
|
| 1032 |
-
"step": 7300
|
| 1033 |
-
}
|
| 1034 |
-
],
|
| 1035 |
-
"logging_steps": 50,
|
| 1036 |
-
"max_steps": 38505,
|
| 1037 |
-
"num_input_tokens_seen": 0,
|
| 1038 |
-
"num_train_epochs": 5,
|
| 1039 |
-
"save_steps": 386,
|
| 1040 |
-
"stateful_callbacks": {
|
| 1041 |
-
"TrainerControl": {
|
| 1042 |
-
"args": {
|
| 1043 |
-
"should_epoch_stop": false,
|
| 1044 |
-
"should_evaluate": false,
|
| 1045 |
-
"should_log": false,
|
| 1046 |
-
"should_save": true,
|
| 1047 |
-
"should_training_stop": false
|
| 1048 |
-
},
|
| 1049 |
-
"attributes": {}
|
| 1050 |
-
}
|
| 1051 |
-
},
|
| 1052 |
-
"total_flos": 6.133152775887913e+17,
|
| 1053 |
-
"train_batch_size": 40,
|
| 1054 |
-
"trial_name": null,
|
| 1055 |
-
"trial_params": null
|
| 1056 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
config.json
DELETED
|
@@ -1,69 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"architectures": [
|
| 3 |
-
"DebertaV2ForTokenClassification"
|
| 4 |
-
],
|
| 5 |
-
"attention_probs_dropout_prob": 0.1,
|
| 6 |
-
"hidden_act": "gelu",
|
| 7 |
-
"hidden_dropout_prob": 0.1,
|
| 8 |
-
"hidden_size": 768,
|
| 9 |
-
"id2label": {
|
| 10 |
-
"0": "O",
|
| 11 |
-
"1": "B-CHA",
|
| 12 |
-
"2": "I-CHA",
|
| 13 |
-
"3": "B-LOC",
|
| 14 |
-
"4": "I-LOC",
|
| 15 |
-
"5": "B-FAC",
|
| 16 |
-
"6": "I-FAC",
|
| 17 |
-
"7": "B-OBJ",
|
| 18 |
-
"8": "I-OBJ",
|
| 19 |
-
"9": "B-EVT",
|
| 20 |
-
"10": "I-EVT",
|
| 21 |
-
"11": "B-ORG",
|
| 22 |
-
"12": "I-ORG",
|
| 23 |
-
"13": "B-MISC",
|
| 24 |
-
"14": "I-MISC"
|
| 25 |
-
},
|
| 26 |
-
"initializer_range": 0.02,
|
| 27 |
-
"intermediate_size": 3072,
|
| 28 |
-
"label2id": {
|
| 29 |
-
"B-CHA": 1,
|
| 30 |
-
"B-EVT": 9,
|
| 31 |
-
"B-FAC": 5,
|
| 32 |
-
"B-LOC": 3,
|
| 33 |
-
"B-MISC": 13,
|
| 34 |
-
"B-OBJ": 7,
|
| 35 |
-
"B-ORG": 11,
|
| 36 |
-
"I-CHA": 2,
|
| 37 |
-
"I-EVT": 10,
|
| 38 |
-
"I-FAC": 6,
|
| 39 |
-
"I-LOC": 4,
|
| 40 |
-
"I-MISC": 14,
|
| 41 |
-
"I-OBJ": 8,
|
| 42 |
-
"I-ORG": 12,
|
| 43 |
-
"O": 0
|
| 44 |
-
},
|
| 45 |
-
"layer_norm_eps": 1e-07,
|
| 46 |
-
"legacy": true,
|
| 47 |
-
"max_position_embeddings": 512,
|
| 48 |
-
"max_relative_positions": -1,
|
| 49 |
-
"model_type": "deberta-v2",
|
| 50 |
-
"norm_rel_ebd": "layer_norm",
|
| 51 |
-
"num_attention_heads": 12,
|
| 52 |
-
"num_hidden_layers": 12,
|
| 53 |
-
"pad_token_id": 0,
|
| 54 |
-
"pooler_dropout": 0,
|
| 55 |
-
"pooler_hidden_act": "gelu",
|
| 56 |
-
"pooler_hidden_size": 768,
|
| 57 |
-
"pos_att_type": [
|
| 58 |
-
"p2c",
|
| 59 |
-
"c2p"
|
| 60 |
-
],
|
| 61 |
-
"position_biased_input": false,
|
| 62 |
-
"position_buckets": 256,
|
| 63 |
-
"relative_attention": true,
|
| 64 |
-
"share_att_key": true,
|
| 65 |
-
"torch_dtype": "float32",
|
| 66 |
-
"transformers_version": "4.55.4",
|
| 67 |
-
"type_vocab_size": 0,
|
| 68 |
-
"vocab_size": 128100
|
| 69 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
special_tokens_map.json
DELETED
|
@@ -1,15 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"bos_token": "[CLS]",
|
| 3 |
-
"cls_token": "[CLS]",
|
| 4 |
-
"eos_token": "[SEP]",
|
| 5 |
-
"mask_token": "[MASK]",
|
| 6 |
-
"pad_token": "[PAD]",
|
| 7 |
-
"sep_token": "[SEP]",
|
| 8 |
-
"unk_token": {
|
| 9 |
-
"content": "[UNK]",
|
| 10 |
-
"lstrip": false,
|
| 11 |
-
"normalized": true,
|
| 12 |
-
"rstrip": false,
|
| 13 |
-
"single_word": false
|
| 14 |
-
}
|
| 15 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
tokenizer.json
DELETED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
DELETED
|
@@ -1,59 +0,0 @@
|
|
| 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": "[CLS]",
|
| 13 |
-
"lstrip": false,
|
| 14 |
-
"normalized": false,
|
| 15 |
-
"rstrip": false,
|
| 16 |
-
"single_word": false,
|
| 17 |
-
"special": true
|
| 18 |
-
},
|
| 19 |
-
"2": {
|
| 20 |
-
"content": "[SEP]",
|
| 21 |
-
"lstrip": false,
|
| 22 |
-
"normalized": false,
|
| 23 |
-
"rstrip": false,
|
| 24 |
-
"single_word": false,
|
| 25 |
-
"special": true
|
| 26 |
-
},
|
| 27 |
-
"3": {
|
| 28 |
-
"content": "[UNK]",
|
| 29 |
-
"lstrip": false,
|
| 30 |
-
"normalized": true,
|
| 31 |
-
"rstrip": false,
|
| 32 |
-
"single_word": false,
|
| 33 |
-
"special": true
|
| 34 |
-
},
|
| 35 |
-
"128000": {
|
| 36 |
-
"content": "[MASK]",
|
| 37 |
-
"lstrip": false,
|
| 38 |
-
"normalized": false,
|
| 39 |
-
"rstrip": false,
|
| 40 |
-
"single_word": false,
|
| 41 |
-
"special": true
|
| 42 |
-
}
|
| 43 |
-
},
|
| 44 |
-
"bos_token": "[CLS]",
|
| 45 |
-
"clean_up_tokenization_spaces": false,
|
| 46 |
-
"cls_token": "[CLS]",
|
| 47 |
-
"do_lower_case": false,
|
| 48 |
-
"eos_token": "[SEP]",
|
| 49 |
-
"extra_special_tokens": {},
|
| 50 |
-
"mask_token": "[MASK]",
|
| 51 |
-
"model_max_length": 1000000000000000019884624838656,
|
| 52 |
-
"pad_token": "[PAD]",
|
| 53 |
-
"sep_token": "[SEP]",
|
| 54 |
-
"sp_model_kwargs": {},
|
| 55 |
-
"split_by_punct": false,
|
| 56 |
-
"tokenizer_class": "DebertaV2Tokenizer",
|
| 57 |
-
"unk_token": "[UNK]",
|
| 58 |
-
"vocab_type": "spm"
|
| 59 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|