Revert "update model"
Browse filesThis reverts commit b92314da2b840b4d670310b37317c0e8b808bbbe.
- .DS_Store +0 -0
- added_tokens.json +2 -1
- all_results.json +10 -10
- config.json +2 -2
- eval_results.json +5 -6
- latest +0 -1
- pytorch_model.bin +1 -1
- rng_state.pth +0 -3
- special_tokens_map.json +2 -2
- tokenizer_config.json +2 -2
- train_results.json +5 -5
- trainer_state.json +980 -260
- training_args.bin +1 -1
- vocab.json +0 -1
.DS_Store
DELETED
|
Binary file (6.15 kB)
|
|
|
added_tokens.json
CHANGED
|
@@ -17,6 +17,7 @@
|
|
| 17 |
"<|da|>": 50285,
|
| 18 |
"<|de|>": 50261,
|
| 19 |
"<|el|>": 50281,
|
|
|
|
| 20 |
"<|en|>": 50259,
|
| 21 |
"<|es|>": 50262,
|
| 22 |
"<|et|>": 50307,
|
|
@@ -29,7 +30,6 @@
|
|
| 29 |
"<|gu|>": 50333,
|
| 30 |
"<|haw|>": 50352,
|
| 31 |
"<|ha|>": 50354,
|
| 32 |
-
"<|he|>": 50279,
|
| 33 |
"<|hi|>": 50276,
|
| 34 |
"<|hr|>": 50291,
|
| 35 |
"<|ht|>": 50339,
|
|
@@ -38,6 +38,7 @@
|
|
| 38 |
"<|id|>": 50275,
|
| 39 |
"<|is|>": 50311,
|
| 40 |
"<|it|>": 50274,
|
|
|
|
| 41 |
"<|ja|>": 50266,
|
| 42 |
"<|jw|>": 50356,
|
| 43 |
"<|ka|>": 50329,
|
|
|
|
| 17 |
"<|da|>": 50285,
|
| 18 |
"<|de|>": 50261,
|
| 19 |
"<|el|>": 50281,
|
| 20 |
+
"<|endoftext|>": 50257,
|
| 21 |
"<|en|>": 50259,
|
| 22 |
"<|es|>": 50262,
|
| 23 |
"<|et|>": 50307,
|
|
|
|
| 30 |
"<|gu|>": 50333,
|
| 31 |
"<|haw|>": 50352,
|
| 32 |
"<|ha|>": 50354,
|
|
|
|
| 33 |
"<|hi|>": 50276,
|
| 34 |
"<|hr|>": 50291,
|
| 35 |
"<|ht|>": 50339,
|
|
|
|
| 38 |
"<|id|>": 50275,
|
| 39 |
"<|is|>": 50311,
|
| 40 |
"<|it|>": 50274,
|
| 41 |
+
"<|iw|>": 50279,
|
| 42 |
"<|ja|>": 50266,
|
| 43 |
"<|jw|>": 50356,
|
| 44 |
"<|ka|>": 50329,
|
all_results.json
CHANGED
|
@@ -1,12 +1,12 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"eval_loss": 0.
|
| 4 |
-
"eval_runtime":
|
| 5 |
-
"eval_samples_per_second":
|
| 6 |
-
"eval_steps_per_second": 0.
|
| 7 |
-
"eval_wer":
|
| 8 |
-
"train_loss": 0.
|
| 9 |
-
"train_runtime":
|
| 10 |
-
"train_samples_per_second":
|
| 11 |
-
"train_steps_per_second": 0.
|
| 12 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 41.67,
|
| 3 |
+
"eval_loss": 0.3455316126346588,
|
| 4 |
+
"eval_runtime": 597.4957,
|
| 5 |
+
"eval_samples_per_second": 3.398,
|
| 6 |
+
"eval_steps_per_second": 0.107,
|
| 7 |
+
"eval_wer": 17.93395978528797,
|
| 8 |
+
"train_loss": 0.02298767093420029,
|
| 9 |
+
"train_runtime": 31651.1636,
|
| 10 |
+
"train_samples_per_second": 5.055,
|
| 11 |
+
"train_steps_per_second": 0.158
|
| 12 |
}
|
config.json
CHANGED
|
@@ -1,5 +1,5 @@
|
|
| 1 |
{
|
| 2 |
-
"_name_or_path": "
|
| 3 |
"activation_dropout": 0.0,
|
| 4 |
"activation_function": "gelu",
|
| 5 |
"architectures": [
|
|
@@ -34,7 +34,7 @@
|
|
| 34 |
"num_mel_bins": 80,
|
| 35 |
"pad_token_id": 50257,
|
| 36 |
"scale_embedding": false,
|
| 37 |
-
"torch_dtype": "
|
| 38 |
"transformers_version": "4.27.0.dev0",
|
| 39 |
"use_cache": false,
|
| 40 |
"vocab_size": 51865
|
|
|
|
| 1 |
{
|
| 2 |
+
"_name_or_path": "./",
|
| 3 |
"activation_dropout": 0.0,
|
| 4 |
"activation_function": "gelu",
|
| 5 |
"architectures": [
|
|
|
|
| 34 |
"num_mel_bins": 80,
|
| 35 |
"pad_token_id": 50257,
|
| 36 |
"scale_embedding": false,
|
| 37 |
+
"torch_dtype": "float32",
|
| 38 |
"transformers_version": "4.27.0.dev0",
|
| 39 |
"use_cache": false,
|
| 40 |
"vocab_size": 51865
|
eval_results.json
CHANGED
|
@@ -1,8 +1,7 @@
|
|
| 1 |
{
|
| 2 |
-
"
|
| 3 |
-
"
|
| 4 |
-
"
|
| 5 |
-
"
|
| 6 |
-
"
|
| 7 |
-
"eval_wer": 16.930468837977504
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"eval_loss": 0.3455316126346588,
|
| 3 |
+
"eval_runtime": 597.4957,
|
| 4 |
+
"eval_samples_per_second": 3.398,
|
| 5 |
+
"eval_steps_per_second": 0.107,
|
| 6 |
+
"eval_wer": 17.93395978528797
|
|
|
|
| 7 |
}
|
latest
DELETED
|
@@ -1 +0,0 @@
|
|
| 1 |
-
global_step2000
|
|
|
|
|
|
pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1527847357
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7aafd9f00980283b3e88daf7c14ee8513283100b75871e2a69ce07b56eeab07c
|
| 3 |
size 1527847357
|
rng_state.pth
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:4d64e72aba0459326893d9ea79b331788cd9c1845d431a1d7f1a025afe964038
|
| 3 |
-
size 14511
|
|
|
|
|
|
|
|
|
|
|
|
special_tokens_map.json
CHANGED
|
@@ -22,7 +22,7 @@
|
|
| 22 |
"<|hi|>",
|
| 23 |
"<|fi|>",
|
| 24 |
"<|vi|>",
|
| 25 |
-
"<|
|
| 26 |
"<|uk|>",
|
| 27 |
"<|el|>",
|
| 28 |
"<|ms|>",
|
|
@@ -124,7 +124,7 @@
|
|
| 124 |
},
|
| 125 |
"pad_token": "<|endoftext|>",
|
| 126 |
"unk_token": {
|
| 127 |
-
"content": "
|
| 128 |
"lstrip": false,
|
| 129 |
"normalized": true,
|
| 130 |
"rstrip": false,
|
|
|
|
| 22 |
"<|hi|>",
|
| 23 |
"<|fi|>",
|
| 24 |
"<|vi|>",
|
| 25 |
+
"<|iw|>",
|
| 26 |
"<|uk|>",
|
| 27 |
"<|el|>",
|
| 28 |
"<|ms|>",
|
|
|
|
| 124 |
},
|
| 125 |
"pad_token": "<|endoftext|>",
|
| 126 |
"unk_token": {
|
| 127 |
+
"content": "",
|
| 128 |
"lstrip": false,
|
| 129 |
"normalized": true,
|
| 130 |
"rstrip": false,
|
tokenizer_config.json
CHANGED
|
@@ -19,7 +19,7 @@
|
|
| 19 |
},
|
| 20 |
"errors": "replace",
|
| 21 |
"model_max_length": 1024,
|
| 22 |
-
"name_or_path": "
|
| 23 |
"pad_token": null,
|
| 24 |
"processor_class": "WhisperProcessor",
|
| 25 |
"return_attention_mask": false,
|
|
@@ -27,7 +27,7 @@
|
|
| 27 |
"tokenizer_class": "WhisperTokenizer",
|
| 28 |
"unk_token": {
|
| 29 |
"__type": "AddedToken",
|
| 30 |
-
"content": "
|
| 31 |
"lstrip": false,
|
| 32 |
"normalized": true,
|
| 33 |
"rstrip": false,
|
|
|
|
| 19 |
},
|
| 20 |
"errors": "replace",
|
| 21 |
"model_max_length": 1024,
|
| 22 |
+
"name_or_path": "./",
|
| 23 |
"pad_token": null,
|
| 24 |
"processor_class": "WhisperProcessor",
|
| 25 |
"return_attention_mask": false,
|
|
|
|
| 27 |
"tokenizer_class": "WhisperTokenizer",
|
| 28 |
"unk_token": {
|
| 29 |
"__type": "AddedToken",
|
| 30 |
+
"content": "",
|
| 31 |
"lstrip": false,
|
| 32 |
"normalized": true,
|
| 33 |
"rstrip": false,
|
train_results.json
CHANGED
|
@@ -1,7 +1,7 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"train_loss": 0.
|
| 4 |
-
"train_runtime":
|
| 5 |
-
"train_samples_per_second":
|
| 6 |
-
"train_steps_per_second": 0.
|
| 7 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 41.67,
|
| 3 |
+
"train_loss": 0.02298767093420029,
|
| 4 |
+
"train_runtime": 31651.1636,
|
| 5 |
+
"train_samples_per_second": 5.055,
|
| 6 |
+
"train_steps_per_second": 0.158
|
| 7 |
}
|
trainer_state.json
CHANGED
|
@@ -1,514 +1,1234 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric":
|
| 3 |
-
"best_model_checkpoint": "./checkpoint-
|
| 4 |
-
"epoch":
|
| 5 |
-
"global_step":
|
| 6 |
"is_hyper_param_search": false,
|
| 7 |
"is_local_process_zero": true,
|
| 8 |
"is_world_process_zero": true,
|
| 9 |
"log_history": [
|
| 10 |
{
|
| 11 |
-
"epoch": 0.
|
| 12 |
-
"learning_rate":
|
| 13 |
-
"loss": 0.
|
| 14 |
"step": 25
|
| 15 |
},
|
| 16 |
{
|
| 17 |
-
"epoch": 0.
|
| 18 |
-
"learning_rate": 6.
|
| 19 |
-
"loss": 0.
|
| 20 |
"step": 50
|
| 21 |
},
|
| 22 |
{
|
| 23 |
-
"epoch": 0.
|
| 24 |
-
"learning_rate": 6.
|
| 25 |
-
"loss": 0.
|
| 26 |
"step": 75
|
| 27 |
},
|
| 28 |
{
|
| 29 |
-
"epoch": 0.
|
| 30 |
-
"learning_rate": 7.
|
| 31 |
-
"loss": 0.
|
| 32 |
"step": 100
|
| 33 |
},
|
| 34 |
{
|
| 35 |
-
"epoch": 1.
|
| 36 |
-
"learning_rate": 7.
|
| 37 |
-
"loss": 0.
|
| 38 |
"step": 125
|
| 39 |
},
|
| 40 |
{
|
| 41 |
-
"epoch": 1.
|
| 42 |
-
"learning_rate": 8.
|
| 43 |
-
"loss": 0.
|
| 44 |
"step": 150
|
| 45 |
},
|
| 46 |
{
|
| 47 |
-
"epoch": 1.
|
| 48 |
-
"learning_rate": 8.
|
| 49 |
-
"loss": 0.
|
| 50 |
"step": 175
|
| 51 |
},
|
| 52 |
{
|
| 53 |
-
"epoch": 1.
|
| 54 |
-
"learning_rate": 8.
|
| 55 |
-
"loss": 0.
|
| 56 |
"step": 200
|
| 57 |
},
|
| 58 |
{
|
| 59 |
-
"epoch":
|
| 60 |
-
"learning_rate": 8.
|
| 61 |
-
"loss": 0.
|
| 62 |
"step": 225
|
| 63 |
},
|
| 64 |
{
|
| 65 |
-
"epoch": 2.
|
| 66 |
-
"learning_rate": 8.
|
| 67 |
-
"loss": 0.
|
| 68 |
"step": 250
|
| 69 |
},
|
| 70 |
{
|
| 71 |
-
"epoch": 2.
|
| 72 |
-
"learning_rate": 9.
|
| 73 |
-
"loss": 0.
|
| 74 |
"step": 275
|
| 75 |
},
|
| 76 |
{
|
| 77 |
-
"epoch": 2.
|
| 78 |
-
"learning_rate": 9.
|
| 79 |
-
"loss": 0.
|
| 80 |
"step": 300
|
| 81 |
},
|
| 82 |
{
|
| 83 |
-
"epoch":
|
| 84 |
-
"learning_rate": 9.
|
| 85 |
-
"loss": 0.
|
| 86 |
"step": 325
|
| 87 |
},
|
| 88 |
{
|
| 89 |
-
"epoch":
|
| 90 |
-
"learning_rate": 9.
|
| 91 |
-
"loss": 0.
|
| 92 |
"step": 350
|
| 93 |
},
|
| 94 |
{
|
| 95 |
-
"epoch": 3.
|
| 96 |
-
"learning_rate": 9.
|
| 97 |
-
"loss": 0.
|
| 98 |
"step": 375
|
| 99 |
},
|
| 100 |
{
|
| 101 |
-
"epoch": 3.
|
| 102 |
-
"learning_rate": 9.
|
| 103 |
-
"loss": 0.
|
| 104 |
"step": 400
|
| 105 |
},
|
| 106 |
{
|
| 107 |
-
"epoch": 3.
|
| 108 |
-
"learning_rate": 9.
|
| 109 |
-
"loss": 0.
|
| 110 |
"step": 425
|
| 111 |
},
|
| 112 |
{
|
| 113 |
-
"epoch":
|
| 114 |
-
"learning_rate": 9.
|
| 115 |
-
"loss": 0.
|
| 116 |
"step": 450
|
| 117 |
},
|
| 118 |
{
|
| 119 |
-
"epoch":
|
| 120 |
-
"learning_rate": 9.
|
| 121 |
-
"loss": 0.
|
| 122 |
"step": 475
|
| 123 |
},
|
| 124 |
{
|
| 125 |
-
"epoch": 4.
|
| 126 |
-
"learning_rate": 9.
|
| 127 |
-
"loss": 0.
|
| 128 |
"step": 500
|
| 129 |
},
|
| 130 |
{
|
| 131 |
-
"epoch": 4.
|
| 132 |
-
"learning_rate": 9.
|
| 133 |
-
"loss": 0.
|
| 134 |
"step": 525
|
| 135 |
},
|
| 136 |
{
|
| 137 |
-
"epoch":
|
| 138 |
-
"learning_rate": 9.
|
| 139 |
-
"loss": 0.
|
| 140 |
"step": 550
|
| 141 |
},
|
| 142 |
{
|
| 143 |
-
"epoch":
|
| 144 |
-
"learning_rate": 9.
|
| 145 |
-
"loss": 0.
|
| 146 |
"step": 575
|
| 147 |
},
|
| 148 |
{
|
| 149 |
-
"epoch": 5.
|
| 150 |
-
"learning_rate": 9.
|
| 151 |
-
"loss": 0.
|
| 152 |
"step": 600
|
| 153 |
},
|
| 154 |
{
|
| 155 |
-
"epoch": 5.
|
| 156 |
-
"learning_rate": 9.
|
| 157 |
-
"loss": 0.
|
| 158 |
"step": 625
|
| 159 |
},
|
| 160 |
{
|
| 161 |
-
"epoch":
|
| 162 |
-
"learning_rate": 9.
|
| 163 |
-
"loss": 0.
|
| 164 |
"step": 650
|
| 165 |
},
|
| 166 |
{
|
| 167 |
-
"epoch":
|
| 168 |
-
"learning_rate":
|
| 169 |
-
"loss": 0.
|
| 170 |
"step": 675
|
| 171 |
},
|
| 172 |
{
|
| 173 |
-
"epoch":
|
| 174 |
-
"learning_rate":
|
| 175 |
-
"loss": 0.
|
| 176 |
"step": 700
|
| 177 |
},
|
| 178 |
{
|
| 179 |
-
"epoch": 6.
|
| 180 |
-
"learning_rate":
|
| 181 |
-
"loss": 0.
|
| 182 |
"step": 725
|
| 183 |
},
|
| 184 |
{
|
| 185 |
-
"epoch": 6.
|
| 186 |
-
"learning_rate":
|
| 187 |
-
"loss": 0.
|
| 188 |
"step": 750
|
| 189 |
},
|
| 190 |
{
|
| 191 |
-
"epoch":
|
| 192 |
-
"learning_rate":
|
| 193 |
-
"loss": 0.
|
| 194 |
"step": 775
|
| 195 |
},
|
| 196 |
{
|
| 197 |
-
"epoch":
|
| 198 |
-
"learning_rate":
|
| 199 |
-
"loss": 0.
|
| 200 |
"step": 800
|
| 201 |
},
|
| 202 |
{
|
| 203 |
-
"epoch":
|
| 204 |
-
"learning_rate":
|
| 205 |
-
"loss": 0.
|
| 206 |
"step": 825
|
| 207 |
},
|
| 208 |
{
|
| 209 |
-
"epoch": 7.
|
| 210 |
-
"learning_rate":
|
| 211 |
-
"loss": 0.
|
| 212 |
"step": 850
|
| 213 |
},
|
| 214 |
{
|
| 215 |
-
"epoch":
|
| 216 |
-
"learning_rate":
|
| 217 |
-
"loss": 0.
|
| 218 |
"step": 875
|
| 219 |
},
|
| 220 |
{
|
| 221 |
-
"epoch":
|
| 222 |
-
"learning_rate":
|
| 223 |
-
"loss": 0.
|
| 224 |
"step": 900
|
| 225 |
},
|
| 226 |
{
|
| 227 |
-
"epoch":
|
| 228 |
-
"learning_rate":
|
| 229 |
-
"loss": 0.
|
| 230 |
"step": 925
|
| 231 |
},
|
| 232 |
{
|
| 233 |
-
"epoch":
|
| 234 |
-
"learning_rate":
|
| 235 |
-
"loss": 0.
|
| 236 |
"step": 950
|
| 237 |
},
|
| 238 |
{
|
| 239 |
-
"epoch":
|
| 240 |
-
"learning_rate":
|
| 241 |
-
"loss": 0.
|
| 242 |
"step": 975
|
| 243 |
},
|
| 244 |
{
|
| 245 |
-
"epoch":
|
| 246 |
-
"learning_rate":
|
| 247 |
-
"loss": 0.
|
| 248 |
"step": 1000
|
| 249 |
},
|
| 250 |
{
|
| 251 |
-
"epoch":
|
| 252 |
-
"learning_rate":
|
| 253 |
-
"loss": 0.
|
| 254 |
"step": 1025
|
| 255 |
},
|
| 256 |
{
|
| 257 |
-
"epoch":
|
| 258 |
-
"learning_rate":
|
| 259 |
-
"loss": 0.
|
| 260 |
"step": 1050
|
| 261 |
},
|
| 262 |
{
|
| 263 |
-
"epoch":
|
| 264 |
-
"learning_rate":
|
| 265 |
-
"loss": 0.
|
| 266 |
"step": 1075
|
| 267 |
},
|
| 268 |
{
|
| 269 |
-
"epoch":
|
| 270 |
-
"learning_rate":
|
| 271 |
-
"loss": 0.
|
| 272 |
"step": 1100
|
| 273 |
},
|
| 274 |
{
|
| 275 |
-
"epoch":
|
| 276 |
-
"learning_rate":
|
| 277 |
-
"loss": 0.
|
| 278 |
"step": 1125
|
| 279 |
},
|
| 280 |
{
|
| 281 |
-
"epoch":
|
| 282 |
-
"learning_rate":
|
| 283 |
-
"loss": 0.
|
| 284 |
"step": 1150
|
| 285 |
},
|
| 286 |
{
|
| 287 |
-
"epoch":
|
| 288 |
-
"learning_rate":
|
| 289 |
-
"loss": 0.
|
| 290 |
"step": 1175
|
| 291 |
},
|
| 292 |
{
|
| 293 |
-
"epoch":
|
| 294 |
-
"learning_rate":
|
| 295 |
-
"loss": 0.
|
| 296 |
"step": 1200
|
| 297 |
},
|
| 298 |
{
|
| 299 |
-
"epoch":
|
| 300 |
-
"learning_rate":
|
| 301 |
-
"loss": 0.
|
| 302 |
"step": 1225
|
| 303 |
},
|
| 304 |
{
|
| 305 |
-
"epoch":
|
| 306 |
-
"learning_rate":
|
| 307 |
-
"loss": 0.
|
| 308 |
"step": 1250
|
| 309 |
},
|
| 310 |
{
|
| 311 |
-
"epoch":
|
| 312 |
-
"learning_rate":
|
| 313 |
-
"loss": 0.
|
| 314 |
"step": 1275
|
| 315 |
},
|
| 316 |
{
|
| 317 |
-
"epoch":
|
| 318 |
-
"learning_rate":
|
| 319 |
-
"loss": 0.
|
| 320 |
"step": 1300
|
| 321 |
},
|
| 322 |
{
|
| 323 |
-
"epoch":
|
| 324 |
-
"learning_rate":
|
| 325 |
-
"loss": 0.
|
| 326 |
"step": 1325
|
| 327 |
},
|
| 328 |
{
|
| 329 |
-
"epoch":
|
| 330 |
-
"learning_rate":
|
| 331 |
-
"loss": 0.
|
| 332 |
"step": 1350
|
| 333 |
},
|
| 334 |
{
|
| 335 |
-
"epoch":
|
| 336 |
-
"learning_rate":
|
| 337 |
-
"loss": 0.
|
| 338 |
"step": 1375
|
| 339 |
},
|
| 340 |
{
|
| 341 |
-
"epoch":
|
| 342 |
-
"learning_rate":
|
| 343 |
-
"loss": 0.
|
| 344 |
"step": 1400
|
| 345 |
},
|
| 346 |
{
|
| 347 |
-
"epoch":
|
| 348 |
-
"learning_rate":
|
| 349 |
-
"loss": 0.
|
| 350 |
"step": 1425
|
| 351 |
},
|
| 352 |
{
|
| 353 |
-
"epoch":
|
| 354 |
-
"learning_rate":
|
| 355 |
-
"loss": 0.
|
| 356 |
"step": 1450
|
| 357 |
},
|
| 358 |
{
|
| 359 |
-
"epoch":
|
| 360 |
-
"learning_rate":
|
| 361 |
-
"loss": 0.
|
| 362 |
"step": 1475
|
| 363 |
},
|
| 364 |
{
|
| 365 |
-
"epoch":
|
| 366 |
-
"learning_rate":
|
| 367 |
-
"loss": 0.
|
| 368 |
"step": 1500
|
| 369 |
},
|
| 370 |
{
|
| 371 |
-
"epoch":
|
| 372 |
-
"learning_rate":
|
| 373 |
-
"loss": 0.
|
| 374 |
"step": 1525
|
| 375 |
},
|
| 376 |
{
|
| 377 |
-
"epoch":
|
| 378 |
-
"learning_rate":
|
| 379 |
-
"loss": 0.
|
| 380 |
"step": 1550
|
| 381 |
},
|
| 382 |
{
|
| 383 |
-
"epoch":
|
| 384 |
-
"learning_rate":
|
| 385 |
-
"loss": 0.
|
| 386 |
"step": 1575
|
| 387 |
},
|
| 388 |
{
|
| 389 |
-
"epoch":
|
| 390 |
-
"learning_rate":
|
| 391 |
-
"loss": 0.
|
| 392 |
"step": 1600
|
| 393 |
},
|
| 394 |
{
|
| 395 |
-
"epoch":
|
| 396 |
-
"learning_rate":
|
| 397 |
-
"loss": 0.
|
| 398 |
"step": 1625
|
| 399 |
},
|
| 400 |
{
|
| 401 |
-
"epoch":
|
| 402 |
-
"learning_rate":
|
| 403 |
-
"loss": 0.
|
| 404 |
"step": 1650
|
| 405 |
},
|
| 406 |
{
|
| 407 |
-
"epoch":
|
| 408 |
-
"learning_rate":
|
| 409 |
-
"loss": 0.
|
| 410 |
"step": 1675
|
| 411 |
},
|
| 412 |
{
|
| 413 |
-
"epoch":
|
| 414 |
-
"learning_rate":
|
| 415 |
-
"loss": 0.
|
| 416 |
"step": 1700
|
| 417 |
},
|
| 418 |
{
|
| 419 |
-
"epoch":
|
| 420 |
-
"learning_rate":
|
| 421 |
-
"loss": 0.
|
| 422 |
"step": 1725
|
| 423 |
},
|
| 424 |
{
|
| 425 |
-
"epoch":
|
| 426 |
-
"learning_rate":
|
| 427 |
-
"loss": 0.
|
| 428 |
"step": 1750
|
| 429 |
},
|
| 430 |
{
|
| 431 |
-
"epoch":
|
| 432 |
-
"learning_rate":
|
| 433 |
-
"loss": 0.
|
| 434 |
"step": 1775
|
| 435 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 436 |
{
|
| 437 |
"epoch": 16.67,
|
| 438 |
-
"learning_rate":
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 439 |
"loss": 0.0001,
|
| 440 |
-
"step":
|
| 441 |
},
|
| 442 |
{
|
| 443 |
-
"epoch":
|
| 444 |
-
"learning_rate":
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 445 |
"loss": 0.0001,
|
| 446 |
-
"step":
|
| 447 |
},
|
| 448 |
{
|
| 449 |
-
"epoch":
|
| 450 |
-
"learning_rate":
|
| 451 |
"loss": 0.0001,
|
| 452 |
-
"step":
|
| 453 |
},
|
| 454 |
{
|
| 455 |
-
"epoch":
|
| 456 |
-
"learning_rate":
|
| 457 |
"loss": 0.0001,
|
| 458 |
-
"step":
|
| 459 |
},
|
| 460 |
{
|
| 461 |
-
"epoch":
|
| 462 |
-
"learning_rate":
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 463 |
"loss": 0.0001,
|
| 464 |
-
"step":
|
| 465 |
},
|
| 466 |
{
|
| 467 |
-
"epoch":
|
| 468 |
-
"learning_rate":
|
| 469 |
"loss": 0.0001,
|
| 470 |
-
"step":
|
| 471 |
},
|
| 472 |
{
|
| 473 |
-
"epoch":
|
| 474 |
-
"learning_rate": 3.
|
| 475 |
"loss": 0.0001,
|
| 476 |
-
"step":
|
| 477 |
},
|
| 478 |
{
|
| 479 |
-
"epoch":
|
| 480 |
-
"learning_rate":
|
| 481 |
"loss": 0.0001,
|
| 482 |
-
"step":
|
| 483 |
},
|
| 484 |
{
|
| 485 |
-
"epoch":
|
| 486 |
-
"learning_rate": 3.
|
| 487 |
"loss": 0.0001,
|
| 488 |
-
"step":
|
| 489 |
},
|
| 490 |
{
|
| 491 |
-
"epoch":
|
| 492 |
-
"
|
| 493 |
-
"
|
| 494 |
-
"
|
| 495 |
-
|
| 496 |
-
|
| 497 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 498 |
},
|
| 499 |
{
|
| 500 |
-
"epoch":
|
| 501 |
-
"step":
|
| 502 |
-
"total_flos":
|
| 503 |
-
"train_loss": 0.
|
| 504 |
-
"train_runtime":
|
| 505 |
-
"train_samples_per_second":
|
| 506 |
-
"train_steps_per_second": 0.
|
| 507 |
}
|
| 508 |
],
|
| 509 |
-
"max_steps":
|
| 510 |
-
"num_train_epochs":
|
| 511 |
-
"total_flos":
|
| 512 |
"trial_name": null,
|
| 513 |
"trial_params": null
|
| 514 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 17.947723376834624,
|
| 3 |
+
"best_model_checkpoint": "./checkpoint-5000",
|
| 4 |
+
"epoch": 41.666666666666664,
|
| 5 |
+
"global_step": 5000,
|
| 6 |
"is_hyper_param_search": false,
|
| 7 |
"is_local_process_zero": true,
|
| 8 |
"is_world_process_zero": true,
|
| 9 |
"log_history": [
|
| 10 |
{
|
| 11 |
+
"epoch": 0.21,
|
| 12 |
+
"learning_rate": 4.898977360288234e-06,
|
| 13 |
+
"loss": 0.546,
|
| 14 |
"step": 25
|
| 15 |
},
|
| 16 |
{
|
| 17 |
+
"epoch": 0.42,
|
| 18 |
+
"learning_rate": 6.160712527409633e-06,
|
| 19 |
+
"loss": 0.4065,
|
| 20 |
"step": 50
|
| 21 |
},
|
| 22 |
{
|
| 23 |
+
"epoch": 0.62,
|
| 24 |
+
"learning_rate": 6.85912902234906e-06,
|
| 25 |
+
"loss": 0.3709,
|
| 26 |
"step": 75
|
| 27 |
},
|
| 28 |
{
|
| 29 |
+
"epoch": 0.83,
|
| 30 |
+
"learning_rate": 7.344547104469332e-06,
|
| 31 |
+
"loss": 0.3362,
|
| 32 |
"step": 100
|
| 33 |
},
|
| 34 |
{
|
| 35 |
+
"epoch": 1.04,
|
| 36 |
+
"learning_rate": 7.716963756434345e-06,
|
| 37 |
+
"loss": 0.3056,
|
| 38 |
"step": 125
|
| 39 |
},
|
| 40 |
{
|
| 41 |
+
"epoch": 1.25,
|
| 42 |
+
"learning_rate": 8.019180844200955e-06,
|
| 43 |
+
"loss": 0.2381,
|
| 44 |
"step": 150
|
| 45 |
},
|
| 46 |
{
|
| 47 |
+
"epoch": 1.46,
|
| 48 |
+
"learning_rate": 8.27351214279797e-06,
|
| 49 |
+
"loss": 0.232,
|
| 50 |
"step": 175
|
| 51 |
},
|
| 52 |
{
|
| 53 |
+
"epoch": 1.67,
|
| 54 |
+
"learning_rate": 8.49307723936858e-06,
|
| 55 |
+
"loss": 0.2129,
|
| 56 |
"step": 200
|
| 57 |
},
|
| 58 |
{
|
| 59 |
+
"epoch": 1.88,
|
| 60 |
+
"learning_rate": 8.686247975778677e-06,
|
| 61 |
+
"loss": 0.214,
|
| 62 |
"step": 225
|
| 63 |
},
|
| 64 |
{
|
| 65 |
+
"epoch": 2.08,
|
| 66 |
+
"learning_rate": 8.858694625217149e-06,
|
| 67 |
+
"loss": 0.1779,
|
| 68 |
"step": 250
|
| 69 |
},
|
| 70 |
{
|
| 71 |
+
"epoch": 2.29,
|
| 72 |
+
"learning_rate": 9.014436199608479e-06,
|
| 73 |
+
"loss": 0.1383,
|
| 74 |
"step": 275
|
| 75 |
},
|
| 76 |
{
|
| 77 |
+
"epoch": 2.5,
|
| 78 |
+
"learning_rate": 9.156425255148058e-06,
|
| 79 |
+
"loss": 0.1322,
|
| 80 |
"step": 300
|
| 81 |
},
|
| 82 |
{
|
| 83 |
+
"epoch": 2.71,
|
| 84 |
+
"learning_rate": 9.28689473531776e-06,
|
| 85 |
+
"loss": 0.1354,
|
| 86 |
"step": 325
|
| 87 |
},
|
| 88 |
{
|
| 89 |
+
"epoch": 2.92,
|
| 90 |
+
"learning_rate": 9.407574351377137e-06,
|
| 91 |
+
"loss": 0.1304,
|
| 92 |
"step": 350
|
| 93 |
},
|
| 94 |
{
|
| 95 |
+
"epoch": 3.12,
|
| 96 |
+
"learning_rate": 9.519831289296397e-06,
|
| 97 |
+
"loss": 0.1044,
|
| 98 |
"step": 375
|
| 99 |
},
|
| 100 |
{
|
| 101 |
+
"epoch": 3.33,
|
| 102 |
+
"learning_rate": 9.624764935335318e-06,
|
| 103 |
+
"loss": 0.076,
|
| 104 |
"step": 400
|
| 105 |
},
|
| 106 |
{
|
| 107 |
+
"epoch": 3.54,
|
| 108 |
+
"learning_rate": 9.723272550712454e-06,
|
| 109 |
+
"loss": 0.079,
|
| 110 |
"step": 425
|
| 111 |
},
|
| 112 |
{
|
| 113 |
+
"epoch": 3.75,
|
| 114 |
+
"learning_rate": 9.816095971633122e-06,
|
| 115 |
+
"loss": 0.0791,
|
| 116 |
"step": 450
|
| 117 |
},
|
| 118 |
{
|
| 119 |
+
"epoch": 3.96,
|
| 120 |
+
"learning_rate": 9.90385555539545e-06,
|
| 121 |
+
"loss": 0.0794,
|
| 122 |
"step": 475
|
| 123 |
},
|
| 124 |
{
|
| 125 |
+
"epoch": 4.17,
|
| 126 |
+
"learning_rate": 9.987075336738768e-06,
|
| 127 |
+
"loss": 0.0534,
|
| 128 |
"step": 500
|
| 129 |
},
|
| 130 |
{
|
| 131 |
+
"epoch": 4.38,
|
| 132 |
+
"learning_rate": 9.955555555555556e-06,
|
| 133 |
+
"loss": 0.0446,
|
| 134 |
"step": 525
|
| 135 |
},
|
| 136 |
{
|
| 137 |
+
"epoch": 4.58,
|
| 138 |
+
"learning_rate": 9.9e-06,
|
| 139 |
+
"loss": 0.0456,
|
| 140 |
"step": 550
|
| 141 |
},
|
| 142 |
{
|
| 143 |
+
"epoch": 4.79,
|
| 144 |
+
"learning_rate": 9.844444444444446e-06,
|
| 145 |
+
"loss": 0.0457,
|
| 146 |
"step": 575
|
| 147 |
},
|
| 148 |
{
|
| 149 |
+
"epoch": 5.0,
|
| 150 |
+
"learning_rate": 9.78888888888889e-06,
|
| 151 |
+
"loss": 0.044,
|
| 152 |
"step": 600
|
| 153 |
},
|
| 154 |
{
|
| 155 |
+
"epoch": 5.21,
|
| 156 |
+
"learning_rate": 9.733333333333334e-06,
|
| 157 |
+
"loss": 0.0244,
|
| 158 |
"step": 625
|
| 159 |
},
|
| 160 |
{
|
| 161 |
+
"epoch": 5.42,
|
| 162 |
+
"learning_rate": 9.677777777777778e-06,
|
| 163 |
+
"loss": 0.0264,
|
| 164 |
"step": 650
|
| 165 |
},
|
| 166 |
{
|
| 167 |
+
"epoch": 5.62,
|
| 168 |
+
"learning_rate": 9.622222222222222e-06,
|
| 169 |
+
"loss": 0.0266,
|
| 170 |
"step": 675
|
| 171 |
},
|
| 172 |
{
|
| 173 |
+
"epoch": 5.83,
|
| 174 |
+
"learning_rate": 9.566666666666668e-06,
|
| 175 |
+
"loss": 0.0253,
|
| 176 |
"step": 700
|
| 177 |
},
|
| 178 |
{
|
| 179 |
+
"epoch": 6.04,
|
| 180 |
+
"learning_rate": 9.511111111111112e-06,
|
| 181 |
+
"loss": 0.024,
|
| 182 |
"step": 725
|
| 183 |
},
|
| 184 |
{
|
| 185 |
+
"epoch": 6.25,
|
| 186 |
+
"learning_rate": 9.455555555555557e-06,
|
| 187 |
+
"loss": 0.0136,
|
| 188 |
"step": 750
|
| 189 |
},
|
| 190 |
{
|
| 191 |
+
"epoch": 6.46,
|
| 192 |
+
"learning_rate": 9.4e-06,
|
| 193 |
+
"loss": 0.013,
|
| 194 |
"step": 775
|
| 195 |
},
|
| 196 |
{
|
| 197 |
+
"epoch": 6.67,
|
| 198 |
+
"learning_rate": 9.344444444444446e-06,
|
| 199 |
+
"loss": 0.0136,
|
| 200 |
"step": 800
|
| 201 |
},
|
| 202 |
{
|
| 203 |
+
"epoch": 6.88,
|
| 204 |
+
"learning_rate": 9.28888888888889e-06,
|
| 205 |
+
"loss": 0.0142,
|
| 206 |
"step": 825
|
| 207 |
},
|
| 208 |
{
|
| 209 |
+
"epoch": 7.08,
|
| 210 |
+
"learning_rate": 9.233333333333334e-06,
|
| 211 |
+
"loss": 0.0131,
|
| 212 |
"step": 850
|
| 213 |
},
|
| 214 |
{
|
| 215 |
+
"epoch": 7.29,
|
| 216 |
+
"learning_rate": 9.17777777777778e-06,
|
| 217 |
+
"loss": 0.0081,
|
| 218 |
"step": 875
|
| 219 |
},
|
| 220 |
{
|
| 221 |
+
"epoch": 7.5,
|
| 222 |
+
"learning_rate": 9.122222222222223e-06,
|
| 223 |
+
"loss": 0.0075,
|
| 224 |
"step": 900
|
| 225 |
},
|
| 226 |
{
|
| 227 |
+
"epoch": 7.71,
|
| 228 |
+
"learning_rate": 9.066666666666667e-06,
|
| 229 |
+
"loss": 0.0085,
|
| 230 |
"step": 925
|
| 231 |
},
|
| 232 |
{
|
| 233 |
+
"epoch": 7.92,
|
| 234 |
+
"learning_rate": 9.011111111111111e-06,
|
| 235 |
+
"loss": 0.0072,
|
| 236 |
"step": 950
|
| 237 |
},
|
| 238 |
{
|
| 239 |
+
"epoch": 8.12,
|
| 240 |
+
"learning_rate": 8.955555555555555e-06,
|
| 241 |
+
"loss": 0.0052,
|
| 242 |
"step": 975
|
| 243 |
},
|
| 244 |
{
|
| 245 |
+
"epoch": 8.33,
|
| 246 |
+
"learning_rate": 8.900000000000001e-06,
|
| 247 |
+
"loss": 0.004,
|
| 248 |
"step": 1000
|
| 249 |
},
|
| 250 |
{
|
| 251 |
+
"epoch": 8.54,
|
| 252 |
+
"learning_rate": 8.844444444444445e-06,
|
| 253 |
+
"loss": 0.0042,
|
| 254 |
"step": 1025
|
| 255 |
},
|
| 256 |
{
|
| 257 |
+
"epoch": 8.75,
|
| 258 |
+
"learning_rate": 8.788888888888891e-06,
|
| 259 |
+
"loss": 0.0045,
|
| 260 |
"step": 1050
|
| 261 |
},
|
| 262 |
{
|
| 263 |
+
"epoch": 8.96,
|
| 264 |
+
"learning_rate": 8.733333333333333e-06,
|
| 265 |
+
"loss": 0.0044,
|
| 266 |
"step": 1075
|
| 267 |
},
|
| 268 |
{
|
| 269 |
+
"epoch": 9.17,
|
| 270 |
+
"learning_rate": 8.677777777777779e-06,
|
| 271 |
+
"loss": 0.0035,
|
| 272 |
"step": 1100
|
| 273 |
},
|
| 274 |
{
|
| 275 |
+
"epoch": 9.38,
|
| 276 |
+
"learning_rate": 8.622222222222223e-06,
|
| 277 |
+
"loss": 0.0029,
|
| 278 |
"step": 1125
|
| 279 |
},
|
| 280 |
{
|
| 281 |
+
"epoch": 9.58,
|
| 282 |
+
"learning_rate": 8.566666666666667e-06,
|
| 283 |
+
"loss": 0.0024,
|
| 284 |
"step": 1150
|
| 285 |
},
|
| 286 |
{
|
| 287 |
+
"epoch": 9.79,
|
| 288 |
+
"learning_rate": 8.511111111111113e-06,
|
| 289 |
+
"loss": 0.0032,
|
| 290 |
"step": 1175
|
| 291 |
},
|
| 292 |
{
|
| 293 |
+
"epoch": 10.0,
|
| 294 |
+
"learning_rate": 8.455555555555555e-06,
|
| 295 |
+
"loss": 0.003,
|
| 296 |
"step": 1200
|
| 297 |
},
|
| 298 |
{
|
| 299 |
+
"epoch": 10.21,
|
| 300 |
+
"learning_rate": 8.400000000000001e-06,
|
| 301 |
+
"loss": 0.0021,
|
| 302 |
"step": 1225
|
| 303 |
},
|
| 304 |
{
|
| 305 |
+
"epoch": 10.42,
|
| 306 |
+
"learning_rate": 8.344444444444445e-06,
|
| 307 |
+
"loss": 0.0024,
|
| 308 |
"step": 1250
|
| 309 |
},
|
| 310 |
{
|
| 311 |
+
"epoch": 10.62,
|
| 312 |
+
"learning_rate": 8.288888888888889e-06,
|
| 313 |
+
"loss": 0.0023,
|
| 314 |
"step": 1275
|
| 315 |
},
|
| 316 |
{
|
| 317 |
+
"epoch": 10.83,
|
| 318 |
+
"learning_rate": 8.233333333333335e-06,
|
| 319 |
+
"loss": 0.0024,
|
| 320 |
"step": 1300
|
| 321 |
},
|
| 322 |
{
|
| 323 |
+
"epoch": 11.04,
|
| 324 |
+
"learning_rate": 8.177777777777779e-06,
|
| 325 |
+
"loss": 0.0019,
|
| 326 |
"step": 1325
|
| 327 |
},
|
| 328 |
{
|
| 329 |
+
"epoch": 11.25,
|
| 330 |
+
"learning_rate": 8.122222222222223e-06,
|
| 331 |
+
"loss": 0.0014,
|
| 332 |
"step": 1350
|
| 333 |
},
|
| 334 |
{
|
| 335 |
+
"epoch": 11.46,
|
| 336 |
+
"learning_rate": 8.066666666666667e-06,
|
| 337 |
+
"loss": 0.0015,
|
| 338 |
"step": 1375
|
| 339 |
},
|
| 340 |
{
|
| 341 |
+
"epoch": 11.67,
|
| 342 |
+
"learning_rate": 8.011111111111113e-06,
|
| 343 |
+
"loss": 0.0018,
|
| 344 |
"step": 1400
|
| 345 |
},
|
| 346 |
{
|
| 347 |
+
"epoch": 11.88,
|
| 348 |
+
"learning_rate": 7.955555555555557e-06,
|
| 349 |
+
"loss": 0.0015,
|
| 350 |
"step": 1425
|
| 351 |
},
|
| 352 |
{
|
| 353 |
+
"epoch": 12.08,
|
| 354 |
+
"learning_rate": 7.9e-06,
|
| 355 |
+
"loss": 0.0016,
|
| 356 |
"step": 1450
|
| 357 |
},
|
| 358 |
{
|
| 359 |
+
"epoch": 12.29,
|
| 360 |
+
"learning_rate": 7.844444444444446e-06,
|
| 361 |
+
"loss": 0.0015,
|
| 362 |
"step": 1475
|
| 363 |
},
|
| 364 |
{
|
| 365 |
+
"epoch": 12.5,
|
| 366 |
+
"learning_rate": 7.788888888888889e-06,
|
| 367 |
+
"loss": 0.0017,
|
| 368 |
"step": 1500
|
| 369 |
},
|
| 370 |
{
|
| 371 |
+
"epoch": 12.71,
|
| 372 |
+
"learning_rate": 7.733333333333334e-06,
|
| 373 |
+
"loss": 0.0016,
|
| 374 |
"step": 1525
|
| 375 |
},
|
| 376 |
{
|
| 377 |
+
"epoch": 12.92,
|
| 378 |
+
"learning_rate": 7.677777777777778e-06,
|
| 379 |
+
"loss": 0.002,
|
| 380 |
"step": 1550
|
| 381 |
},
|
| 382 |
{
|
| 383 |
+
"epoch": 13.12,
|
| 384 |
+
"learning_rate": 7.622222222222223e-06,
|
| 385 |
+
"loss": 0.0015,
|
| 386 |
"step": 1575
|
| 387 |
},
|
| 388 |
{
|
| 389 |
+
"epoch": 13.33,
|
| 390 |
+
"learning_rate": 7.566666666666667e-06,
|
| 391 |
+
"loss": 0.0015,
|
| 392 |
"step": 1600
|
| 393 |
},
|
| 394 |
{
|
| 395 |
+
"epoch": 13.54,
|
| 396 |
+
"learning_rate": 7.511111111111111e-06,
|
| 397 |
+
"loss": 0.0015,
|
| 398 |
"step": 1625
|
| 399 |
},
|
| 400 |
{
|
| 401 |
+
"epoch": 13.75,
|
| 402 |
+
"learning_rate": 7.455555555555556e-06,
|
| 403 |
+
"loss": 0.0014,
|
| 404 |
"step": 1650
|
| 405 |
},
|
| 406 |
{
|
| 407 |
+
"epoch": 13.96,
|
| 408 |
+
"learning_rate": 7.4e-06,
|
| 409 |
+
"loss": 0.0017,
|
| 410 |
"step": 1675
|
| 411 |
},
|
| 412 |
{
|
| 413 |
+
"epoch": 14.17,
|
| 414 |
+
"learning_rate": 7.344444444444445e-06,
|
| 415 |
+
"loss": 0.0014,
|
| 416 |
"step": 1700
|
| 417 |
},
|
| 418 |
{
|
| 419 |
+
"epoch": 14.38,
|
| 420 |
+
"learning_rate": 7.28888888888889e-06,
|
| 421 |
+
"loss": 0.0013,
|
| 422 |
"step": 1725
|
| 423 |
},
|
| 424 |
{
|
| 425 |
+
"epoch": 14.58,
|
| 426 |
+
"learning_rate": 7.233333333333334e-06,
|
| 427 |
+
"loss": 0.0013,
|
| 428 |
"step": 1750
|
| 429 |
},
|
| 430 |
{
|
| 431 |
+
"epoch": 14.79,
|
| 432 |
+
"learning_rate": 7.177777777777778e-06,
|
| 433 |
+
"loss": 0.0018,
|
| 434 |
"step": 1775
|
| 435 |
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 15.0,
|
| 438 |
+
"learning_rate": 7.122222222222222e-06,
|
| 439 |
+
"loss": 0.0015,
|
| 440 |
+
"step": 1800
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 15.21,
|
| 444 |
+
"learning_rate": 7.066666666666667e-06,
|
| 445 |
+
"loss": 0.0017,
|
| 446 |
+
"step": 1825
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 15.42,
|
| 450 |
+
"learning_rate": 7.011111111111112e-06,
|
| 451 |
+
"loss": 0.0014,
|
| 452 |
+
"step": 1850
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 15.62,
|
| 456 |
+
"learning_rate": 6.955555555555557e-06,
|
| 457 |
+
"loss": 0.0018,
|
| 458 |
+
"step": 1875
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 15.83,
|
| 462 |
+
"learning_rate": 6.9e-06,
|
| 463 |
+
"loss": 0.0018,
|
| 464 |
+
"step": 1900
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 16.04,
|
| 468 |
+
"learning_rate": 6.844444444444445e-06,
|
| 469 |
+
"loss": 0.0021,
|
| 470 |
+
"step": 1925
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 16.25,
|
| 474 |
+
"learning_rate": 6.788888888888889e-06,
|
| 475 |
+
"loss": 0.0024,
|
| 476 |
+
"step": 1950
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 16.46,
|
| 480 |
+
"learning_rate": 6.733333333333334e-06,
|
| 481 |
+
"loss": 0.0019,
|
| 482 |
+
"step": 1975
|
| 483 |
+
},
|
| 484 |
{
|
| 485 |
"epoch": 16.67,
|
| 486 |
+
"learning_rate": 6.677777777777779e-06,
|
| 487 |
+
"loss": 0.0018,
|
| 488 |
+
"step": 2000
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 16.88,
|
| 492 |
+
"learning_rate": 6.6222222222222236e-06,
|
| 493 |
+
"loss": 0.0021,
|
| 494 |
+
"step": 2025
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 17.08,
|
| 498 |
+
"learning_rate": 6.566666666666667e-06,
|
| 499 |
+
"loss": 0.0019,
|
| 500 |
+
"step": 2050
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 17.29,
|
| 504 |
+
"learning_rate": 6.511111111111112e-06,
|
| 505 |
+
"loss": 0.0017,
|
| 506 |
+
"step": 2075
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 17.5,
|
| 510 |
+
"learning_rate": 6.455555555555556e-06,
|
| 511 |
+
"loss": 0.0013,
|
| 512 |
+
"step": 2100
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 17.71,
|
| 516 |
+
"learning_rate": 6.4000000000000006e-06,
|
| 517 |
+
"loss": 0.0021,
|
| 518 |
+
"step": 2125
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 17.92,
|
| 522 |
+
"learning_rate": 6.3444444444444454e-06,
|
| 523 |
+
"loss": 0.0015,
|
| 524 |
+
"step": 2150
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 18.12,
|
| 528 |
+
"learning_rate": 6.28888888888889e-06,
|
| 529 |
+
"loss": 0.0015,
|
| 530 |
+
"step": 2175
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 18.33,
|
| 534 |
+
"learning_rate": 6.2333333333333335e-06,
|
| 535 |
+
"loss": 0.0016,
|
| 536 |
+
"step": 2200
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 18.54,
|
| 540 |
+
"learning_rate": 6.177777777777778e-06,
|
| 541 |
+
"loss": 0.0014,
|
| 542 |
+
"step": 2225
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 18.75,
|
| 546 |
+
"learning_rate": 6.1222222222222224e-06,
|
| 547 |
+
"loss": 0.0017,
|
| 548 |
+
"step": 2250
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 18.96,
|
| 552 |
+
"learning_rate": 6.066666666666667e-06,
|
| 553 |
+
"loss": 0.0015,
|
| 554 |
+
"step": 2275
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 19.17,
|
| 558 |
+
"learning_rate": 6.011111111111112e-06,
|
| 559 |
+
"loss": 0.0016,
|
| 560 |
+
"step": 2300
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 19.38,
|
| 564 |
+
"learning_rate": 5.955555555555555e-06,
|
| 565 |
+
"loss": 0.0014,
|
| 566 |
+
"step": 2325
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 19.58,
|
| 570 |
+
"learning_rate": 5.9e-06,
|
| 571 |
+
"loss": 0.0013,
|
| 572 |
+
"step": 2350
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 19.79,
|
| 576 |
+
"learning_rate": 5.844444444444445e-06,
|
| 577 |
+
"loss": 0.0013,
|
| 578 |
+
"step": 2375
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 20.0,
|
| 582 |
+
"learning_rate": 5.788888888888889e-06,
|
| 583 |
+
"loss": 0.0014,
|
| 584 |
+
"step": 2400
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 20.21,
|
| 588 |
+
"learning_rate": 5.733333333333334e-06,
|
| 589 |
+
"loss": 0.0009,
|
| 590 |
+
"step": 2425
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 20.42,
|
| 594 |
+
"learning_rate": 5.677777777777779e-06,
|
| 595 |
+
"loss": 0.0012,
|
| 596 |
+
"step": 2450
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 20.62,
|
| 600 |
+
"learning_rate": 5.622222222222222e-06,
|
| 601 |
+
"loss": 0.0014,
|
| 602 |
+
"step": 2475
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 20.83,
|
| 606 |
+
"learning_rate": 5.566666666666667e-06,
|
| 607 |
+
"loss": 0.0011,
|
| 608 |
+
"step": 2500
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 21.04,
|
| 612 |
+
"learning_rate": 5.511111111111112e-06,
|
| 613 |
+
"loss": 0.0011,
|
| 614 |
+
"step": 2525
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 21.25,
|
| 618 |
+
"learning_rate": 5.455555555555556e-06,
|
| 619 |
+
"loss": 0.001,
|
| 620 |
+
"step": 2550
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 21.46,
|
| 624 |
+
"learning_rate": 5.400000000000001e-06,
|
| 625 |
+
"loss": 0.0011,
|
| 626 |
+
"step": 2575
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 21.67,
|
| 630 |
+
"learning_rate": 5.344444444444446e-06,
|
| 631 |
+
"loss": 0.001,
|
| 632 |
+
"step": 2600
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 21.88,
|
| 636 |
+
"learning_rate": 5.288888888888889e-06,
|
| 637 |
+
"loss": 0.0012,
|
| 638 |
+
"step": 2625
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 22.08,
|
| 642 |
+
"learning_rate": 5.233333333333334e-06,
|
| 643 |
+
"loss": 0.0008,
|
| 644 |
+
"step": 2650
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 22.29,
|
| 648 |
+
"learning_rate": 5.177777777777779e-06,
|
| 649 |
+
"loss": 0.0007,
|
| 650 |
+
"step": 2675
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 22.5,
|
| 654 |
+
"learning_rate": 5.122222222222223e-06,
|
| 655 |
+
"loss": 0.0007,
|
| 656 |
+
"step": 2700
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 22.71,
|
| 660 |
+
"learning_rate": 5.0666666666666676e-06,
|
| 661 |
+
"loss": 0.0009,
|
| 662 |
+
"step": 2725
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 22.92,
|
| 666 |
+
"learning_rate": 5.011111111111111e-06,
|
| 667 |
+
"loss": 0.0006,
|
| 668 |
+
"step": 2750
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 23.12,
|
| 672 |
+
"learning_rate": 4.9555555555555565e-06,
|
| 673 |
+
"loss": 0.0006,
|
| 674 |
+
"step": 2775
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 23.33,
|
| 678 |
+
"learning_rate": 4.9000000000000005e-06,
|
| 679 |
+
"loss": 0.0003,
|
| 680 |
+
"step": 2800
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 23.54,
|
| 684 |
+
"learning_rate": 4.8444444444444446e-06,
|
| 685 |
+
"loss": 0.0007,
|
| 686 |
+
"step": 2825
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 23.75,
|
| 690 |
+
"learning_rate": 4.7888888888888894e-06,
|
| 691 |
+
"loss": 0.0004,
|
| 692 |
+
"step": 2850
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 23.96,
|
| 696 |
+
"learning_rate": 4.7333333333333335e-06,
|
| 697 |
+
"loss": 0.0004,
|
| 698 |
+
"step": 2875
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 24.17,
|
| 702 |
+
"learning_rate": 4.677777777777778e-06,
|
| 703 |
+
"loss": 0.0003,
|
| 704 |
+
"step": 2900
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 24.38,
|
| 708 |
+
"learning_rate": 4.622222222222222e-06,
|
| 709 |
+
"loss": 0.0003,
|
| 710 |
+
"step": 2925
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 24.58,
|
| 714 |
+
"learning_rate": 4.566666666666667e-06,
|
| 715 |
+
"loss": 0.0003,
|
| 716 |
+
"step": 2950
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 24.79,
|
| 720 |
+
"learning_rate": 4.511111111111111e-06,
|
| 721 |
+
"loss": 0.0003,
|
| 722 |
+
"step": 2975
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 25.0,
|
| 726 |
+
"learning_rate": 4.455555555555555e-06,
|
| 727 |
+
"loss": 0.0002,
|
| 728 |
+
"step": 3000
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 25.21,
|
| 732 |
+
"learning_rate": 4.404444444444445e-06,
|
| 733 |
+
"loss": 0.0002,
|
| 734 |
+
"step": 3025
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 25.42,
|
| 738 |
+
"learning_rate": 4.348888888888889e-06,
|
| 739 |
+
"loss": 0.0002,
|
| 740 |
+
"step": 3050
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 25.62,
|
| 744 |
+
"learning_rate": 4.2933333333333334e-06,
|
| 745 |
+
"loss": 0.0002,
|
| 746 |
+
"step": 3075
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 25.83,
|
| 750 |
+
"learning_rate": 4.2377777777777775e-06,
|
| 751 |
+
"loss": 0.0002,
|
| 752 |
+
"step": 3100
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 26.04,
|
| 756 |
+
"learning_rate": 4.182222222222222e-06,
|
| 757 |
+
"loss": 0.0002,
|
| 758 |
+
"step": 3125
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 26.25,
|
| 762 |
+
"learning_rate": 4.126666666666667e-06,
|
| 763 |
+
"loss": 0.0003,
|
| 764 |
+
"step": 3150
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 26.46,
|
| 768 |
+
"learning_rate": 4.071111111111111e-06,
|
| 769 |
+
"loss": 0.0002,
|
| 770 |
+
"step": 3175
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 26.67,
|
| 774 |
+
"learning_rate": 4.015555555555556e-06,
|
| 775 |
+
"loss": 0.0003,
|
| 776 |
+
"step": 3200
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 26.88,
|
| 780 |
+
"learning_rate": 3.96e-06,
|
| 781 |
+
"loss": 0.0002,
|
| 782 |
+
"step": 3225
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 27.08,
|
| 786 |
+
"learning_rate": 3.904444444444444e-06,
|
| 787 |
"loss": 0.0001,
|
| 788 |
+
"step": 3250
|
| 789 |
},
|
| 790 |
{
|
| 791 |
+
"epoch": 27.29,
|
| 792 |
+
"learning_rate": 3.848888888888889e-06,
|
| 793 |
+
"loss": 0.0002,
|
| 794 |
+
"step": 3275
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 27.5,
|
| 798 |
+
"learning_rate": 3.793333333333334e-06,
|
| 799 |
+
"loss": 0.0002,
|
| 800 |
+
"step": 3300
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 27.71,
|
| 804 |
+
"learning_rate": 3.737777777777778e-06,
|
| 805 |
+
"loss": 0.0002,
|
| 806 |
+
"step": 3325
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 27.92,
|
| 810 |
+
"learning_rate": 3.6822222222222225e-06,
|
| 811 |
"loss": 0.0001,
|
| 812 |
+
"step": 3350
|
| 813 |
},
|
| 814 |
{
|
| 815 |
+
"epoch": 28.12,
|
| 816 |
+
"learning_rate": 3.6266666666666674e-06,
|
| 817 |
"loss": 0.0001,
|
| 818 |
+
"step": 3375
|
| 819 |
},
|
| 820 |
{
|
| 821 |
+
"epoch": 28.33,
|
| 822 |
+
"learning_rate": 3.5711111111111114e-06,
|
| 823 |
"loss": 0.0001,
|
| 824 |
+
"step": 3400
|
| 825 |
},
|
| 826 |
{
|
| 827 |
+
"epoch": 28.54,
|
| 828 |
+
"learning_rate": 3.515555555555556e-06,
|
| 829 |
+
"loss": 0.0002,
|
| 830 |
+
"step": 3425
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 28.75,
|
| 834 |
+
"learning_rate": 3.46e-06,
|
| 835 |
"loss": 0.0001,
|
| 836 |
+
"step": 3450
|
| 837 |
},
|
| 838 |
{
|
| 839 |
+
"epoch": 28.96,
|
| 840 |
+
"learning_rate": 3.404444444444445e-06,
|
| 841 |
"loss": 0.0001,
|
| 842 |
+
"step": 3475
|
| 843 |
},
|
| 844 |
{
|
| 845 |
+
"epoch": 29.17,
|
| 846 |
+
"learning_rate": 3.3488888888888892e-06,
|
| 847 |
"loss": 0.0001,
|
| 848 |
+
"step": 3500
|
| 849 |
},
|
| 850 |
{
|
| 851 |
+
"epoch": 29.38,
|
| 852 |
+
"learning_rate": 3.2933333333333333e-06,
|
| 853 |
"loss": 0.0001,
|
| 854 |
+
"step": 3525
|
| 855 |
},
|
| 856 |
{
|
| 857 |
+
"epoch": 29.58,
|
| 858 |
+
"learning_rate": 3.237777777777778e-06,
|
| 859 |
"loss": 0.0001,
|
| 860 |
+
"step": 3550
|
| 861 |
},
|
| 862 |
{
|
| 863 |
+
"epoch": 29.79,
|
| 864 |
+
"learning_rate": 3.1822222222222226e-06,
|
| 865 |
+
"loss": 0.0001,
|
| 866 |
+
"step": 3575
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 30.0,
|
| 870 |
+
"learning_rate": 3.1266666666666667e-06,
|
| 871 |
+
"loss": 0.0001,
|
| 872 |
+
"step": 3600
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 30.21,
|
| 876 |
+
"learning_rate": 3.0711111111111115e-06,
|
| 877 |
+
"loss": 0.0001,
|
| 878 |
+
"step": 3625
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 30.42,
|
| 882 |
+
"learning_rate": 3.015555555555556e-06,
|
| 883 |
+
"loss": 0.0001,
|
| 884 |
+
"step": 3650
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 30.62,
|
| 888 |
+
"learning_rate": 2.96e-06,
|
| 889 |
+
"loss": 0.0001,
|
| 890 |
+
"step": 3675
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 30.83,
|
| 894 |
+
"learning_rate": 2.904444444444445e-06,
|
| 895 |
+
"loss": 0.0001,
|
| 896 |
+
"step": 3700
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 31.04,
|
| 900 |
+
"learning_rate": 2.8488888888888894e-06,
|
| 901 |
+
"loss": 0.0001,
|
| 902 |
+
"step": 3725
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 31.25,
|
| 906 |
+
"learning_rate": 2.7933333333333334e-06,
|
| 907 |
+
"loss": 0.0001,
|
| 908 |
+
"step": 3750
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 31.46,
|
| 912 |
+
"learning_rate": 2.7377777777777783e-06,
|
| 913 |
+
"loss": 0.0001,
|
| 914 |
+
"step": 3775
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 31.67,
|
| 918 |
+
"learning_rate": 2.6822222222222223e-06,
|
| 919 |
+
"loss": 0.0001,
|
| 920 |
+
"step": 3800
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 31.88,
|
| 924 |
+
"learning_rate": 2.6266666666666668e-06,
|
| 925 |
+
"loss": 0.0001,
|
| 926 |
+
"step": 3825
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 32.08,
|
| 930 |
+
"learning_rate": 2.5711111111111112e-06,
|
| 931 |
+
"loss": 0.0001,
|
| 932 |
+
"step": 3850
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 32.29,
|
| 936 |
+
"learning_rate": 2.5155555555555557e-06,
|
| 937 |
+
"loss": 0.0001,
|
| 938 |
+
"step": 3875
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 32.5,
|
| 942 |
+
"learning_rate": 2.46e-06,
|
| 943 |
+
"loss": 0.0001,
|
| 944 |
+
"step": 3900
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 32.71,
|
| 948 |
+
"learning_rate": 2.4044444444444446e-06,
|
| 949 |
+
"loss": 0.0001,
|
| 950 |
+
"step": 3925
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 32.92,
|
| 954 |
+
"learning_rate": 2.348888888888889e-06,
|
| 955 |
+
"loss": 0.0001,
|
| 956 |
+
"step": 3950
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 33.12,
|
| 960 |
+
"learning_rate": 2.2933333333333335e-06,
|
| 961 |
+
"loss": 0.0001,
|
| 962 |
+
"step": 3975
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 33.33,
|
| 966 |
+
"learning_rate": 2.237777777777778e-06,
|
| 967 |
+
"loss": 0.0001,
|
| 968 |
+
"step": 4000
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 33.54,
|
| 972 |
+
"learning_rate": 2.1866666666666668e-06,
|
| 973 |
+
"loss": 0.0001,
|
| 974 |
+
"step": 4025
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 33.75,
|
| 978 |
+
"learning_rate": 2.1311111111111112e-06,
|
| 979 |
+
"loss": 0.0001,
|
| 980 |
+
"step": 4050
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 33.96,
|
| 984 |
+
"learning_rate": 2.0755555555555557e-06,
|
| 985 |
+
"loss": 0.0001,
|
| 986 |
+
"step": 4075
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 34.17,
|
| 990 |
+
"learning_rate": 2.02e-06,
|
| 991 |
+
"loss": 0.0001,
|
| 992 |
+
"step": 4100
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 34.38,
|
| 996 |
+
"learning_rate": 1.9644444444444446e-06,
|
| 997 |
+
"loss": 0.0001,
|
| 998 |
+
"step": 4125
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 34.58,
|
| 1002 |
+
"learning_rate": 1.908888888888889e-06,
|
| 1003 |
+
"loss": 0.0001,
|
| 1004 |
+
"step": 4150
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 34.79,
|
| 1008 |
+
"learning_rate": 1.8533333333333333e-06,
|
| 1009 |
+
"loss": 0.0001,
|
| 1010 |
+
"step": 4175
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 35.0,
|
| 1014 |
+
"learning_rate": 1.797777777777778e-06,
|
| 1015 |
+
"loss": 0.0001,
|
| 1016 |
+
"step": 4200
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 35.21,
|
| 1020 |
+
"learning_rate": 1.7422222222222224e-06,
|
| 1021 |
+
"loss": 0.0001,
|
| 1022 |
+
"step": 4225
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 35.42,
|
| 1026 |
+
"learning_rate": 1.6866666666666667e-06,
|
| 1027 |
+
"loss": 0.0001,
|
| 1028 |
+
"step": 4250
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 35.62,
|
| 1032 |
+
"learning_rate": 1.6311111111111114e-06,
|
| 1033 |
+
"loss": 0.0001,
|
| 1034 |
+
"step": 4275
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 35.83,
|
| 1038 |
+
"learning_rate": 1.5755555555555558e-06,
|
| 1039 |
+
"loss": 0.0001,
|
| 1040 |
+
"step": 4300
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 36.04,
|
| 1044 |
+
"learning_rate": 1.52e-06,
|
| 1045 |
+
"loss": 0.0001,
|
| 1046 |
+
"step": 4325
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 36.25,
|
| 1050 |
+
"learning_rate": 1.4644444444444445e-06,
|
| 1051 |
+
"loss": 0.0001,
|
| 1052 |
+
"step": 4350
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 36.46,
|
| 1056 |
+
"learning_rate": 1.4088888888888892e-06,
|
| 1057 |
+
"loss": 0.0001,
|
| 1058 |
+
"step": 4375
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 36.67,
|
| 1062 |
+
"learning_rate": 1.3533333333333334e-06,
|
| 1063 |
+
"loss": 0.0001,
|
| 1064 |
+
"step": 4400
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 36.88,
|
| 1068 |
+
"learning_rate": 1.2977777777777779e-06,
|
| 1069 |
+
"loss": 0.0001,
|
| 1070 |
+
"step": 4425
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 37.08,
|
| 1074 |
+
"learning_rate": 1.2422222222222224e-06,
|
| 1075 |
+
"loss": 0.0001,
|
| 1076 |
+
"step": 4450
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 37.29,
|
| 1080 |
+
"learning_rate": 1.1866666666666668e-06,
|
| 1081 |
+
"loss": 0.0001,
|
| 1082 |
+
"step": 4475
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 37.5,
|
| 1086 |
+
"learning_rate": 1.131111111111111e-06,
|
| 1087 |
+
"loss": 0.0001,
|
| 1088 |
+
"step": 4500
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 37.71,
|
| 1092 |
+
"learning_rate": 1.0755555555555557e-06,
|
| 1093 |
+
"loss": 0.0001,
|
| 1094 |
+
"step": 4525
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 37.92,
|
| 1098 |
+
"learning_rate": 1.02e-06,
|
| 1099 |
+
"loss": 0.0001,
|
| 1100 |
+
"step": 4550
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 38.12,
|
| 1104 |
+
"learning_rate": 9.644444444444444e-07,
|
| 1105 |
+
"loss": 0.0001,
|
| 1106 |
+
"step": 4575
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 38.33,
|
| 1110 |
+
"learning_rate": 9.08888888888889e-07,
|
| 1111 |
+
"loss": 0.0001,
|
| 1112 |
+
"step": 4600
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 38.54,
|
| 1116 |
+
"learning_rate": 8.533333333333334e-07,
|
| 1117 |
+
"loss": 0.0001,
|
| 1118 |
+
"step": 4625
|
| 1119 |
+
},
|
| 1120 |
+
{
|
| 1121 |
+
"epoch": 38.75,
|
| 1122 |
+
"learning_rate": 7.977777777777779e-07,
|
| 1123 |
+
"loss": 0.0001,
|
| 1124 |
+
"step": 4650
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 38.96,
|
| 1128 |
+
"learning_rate": 7.422222222222223e-07,
|
| 1129 |
+
"loss": 0.0001,
|
| 1130 |
+
"step": 4675
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 39.17,
|
| 1134 |
+
"learning_rate": 6.866666666666667e-07,
|
| 1135 |
+
"loss": 0.0001,
|
| 1136 |
+
"step": 4700
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 39.38,
|
| 1140 |
+
"learning_rate": 6.311111111111112e-07,
|
| 1141 |
+
"loss": 0.0001,
|
| 1142 |
+
"step": 4725
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 39.58,
|
| 1146 |
+
"learning_rate": 5.755555555555555e-07,
|
| 1147 |
+
"loss": 0.0001,
|
| 1148 |
+
"step": 4750
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 39.79,
|
| 1152 |
+
"learning_rate": 5.2e-07,
|
| 1153 |
+
"loss": 0.0001,
|
| 1154 |
+
"step": 4775
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 40.0,
|
| 1158 |
+
"learning_rate": 4.6444444444444446e-07,
|
| 1159 |
+
"loss": 0.0001,
|
| 1160 |
+
"step": 4800
|
| 1161 |
+
},
|
| 1162 |
+
{
|
| 1163 |
+
"epoch": 40.21,
|
| 1164 |
+
"learning_rate": 4.0888888888888897e-07,
|
| 1165 |
+
"loss": 0.0001,
|
| 1166 |
+
"step": 4825
|
| 1167 |
+
},
|
| 1168 |
+
{
|
| 1169 |
+
"epoch": 40.42,
|
| 1170 |
+
"learning_rate": 3.533333333333334e-07,
|
| 1171 |
+
"loss": 0.0001,
|
| 1172 |
+
"step": 4850
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 40.62,
|
| 1176 |
+
"learning_rate": 2.977777777777778e-07,
|
| 1177 |
+
"loss": 0.0001,
|
| 1178 |
+
"step": 4875
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 40.83,
|
| 1182 |
+
"learning_rate": 2.4222222222222224e-07,
|
| 1183 |
+
"loss": 0.0001,
|
| 1184 |
+
"step": 4900
|
| 1185 |
+
},
|
| 1186 |
+
{
|
| 1187 |
+
"epoch": 41.04,
|
| 1188 |
+
"learning_rate": 1.866666666666667e-07,
|
| 1189 |
+
"loss": 0.0001,
|
| 1190 |
+
"step": 4925
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 41.25,
|
| 1194 |
+
"learning_rate": 1.3111111111111113e-07,
|
| 1195 |
+
"loss": 0.0001,
|
| 1196 |
+
"step": 4950
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 41.46,
|
| 1200 |
+
"learning_rate": 7.555555555555556e-08,
|
| 1201 |
+
"loss": 0.0001,
|
| 1202 |
+
"step": 4975
|
| 1203 |
+
},
|
| 1204 |
+
{
|
| 1205 |
+
"epoch": 41.67,
|
| 1206 |
+
"learning_rate": 2e-08,
|
| 1207 |
+
"loss": 0.0001,
|
| 1208 |
+
"step": 5000
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 41.67,
|
| 1212 |
+
"eval_loss": 0.345458984375,
|
| 1213 |
+
"eval_runtime": 344.0011,
|
| 1214 |
+
"eval_samples_per_second": 5.901,
|
| 1215 |
+
"eval_steps_per_second": 0.186,
|
| 1216 |
+
"eval_wer": 17.947723376834624,
|
| 1217 |
+
"step": 5000
|
| 1218 |
},
|
| 1219 |
{
|
| 1220 |
+
"epoch": 41.67,
|
| 1221 |
+
"step": 5000,
|
| 1222 |
+
"total_flos": 1.6225066659249494e+20,
|
| 1223 |
+
"train_loss": 0.02298767093420029,
|
| 1224 |
+
"train_runtime": 31651.1636,
|
| 1225 |
+
"train_samples_per_second": 5.055,
|
| 1226 |
+
"train_steps_per_second": 0.158
|
| 1227 |
}
|
| 1228 |
],
|
| 1229 |
+
"max_steps": 5000,
|
| 1230 |
+
"num_train_epochs": 42,
|
| 1231 |
+
"total_flos": 1.6225066659249494e+20,
|
| 1232 |
"trial_name": null,
|
| 1233 |
"trial_params": null
|
| 1234 |
}
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4731
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:faf5417f1671844821e708f0d99b9717f45d1adab1b8fbbfe170676b5af68ed8
|
| 3 |
size 4731
|
vocab.json
CHANGED
|
@@ -314,7 +314,6 @@
|
|
| 314 |
";;": 35746,
|
| 315 |
"<": 27,
|
| 316 |
"</": 3433,
|
| 317 |
-
"<|endoftext|>": 50257,
|
| 318 |
"=": 28,
|
| 319 |
"=\"": 13114,
|
| 320 |
"=\"#": 34106,
|
|
|
|
| 314 |
";;": 35746,
|
| 315 |
"<": 27,
|
| 316 |
"</": 3433,
|
|
|
|
| 317 |
"=": 28,
|
| 318 |
"=\"": 13114,
|
| 319 |
"=\"#": 34106,
|