Commit ·
329df2f
1
Parent(s): 1573f37
End of training
Browse files- README.md +14 -2
- all_results.json +14 -14
- eval_results.json +9 -9
- train_results.json +5 -5
- trainer_state.json +627 -873
README.md
CHANGED
|
@@ -2,11 +2,23 @@
|
|
| 2 |
base_model: reza-alipour/ft5
|
| 3 |
tags:
|
| 4 |
- generated_from_trainer
|
|
|
|
|
|
|
| 5 |
metrics:
|
| 6 |
- rouge
|
| 7 |
model-index:
|
| 8 |
- name: ft5
|
| 9 |
-
results:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 10 |
---
|
| 11 |
|
| 12 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
|
@@ -14,7 +26,7 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 14 |
|
| 15 |
# ft5
|
| 16 |
|
| 17 |
-
This model is a fine-tuned version of [reza-alipour/ft5](https://huggingface.co/reza-alipour/ft5) on
|
| 18 |
It achieves the following results on the evaluation set:
|
| 19 |
- Loss: 0.3032
|
| 20 |
- Rouge1: 86.5313
|
|
|
|
| 2 |
base_model: reza-alipour/ft5
|
| 3 |
tags:
|
| 4 |
- generated_from_trainer
|
| 5 |
+
datasets:
|
| 6 |
+
- reza-alipour/Text-Edit-Instruct-Preprocessed-4m
|
| 7 |
metrics:
|
| 8 |
- rouge
|
| 9 |
model-index:
|
| 10 |
- name: ft5
|
| 11 |
+
results:
|
| 12 |
+
- task:
|
| 13 |
+
name: Summarization
|
| 14 |
+
type: summarization
|
| 15 |
+
dataset:
|
| 16 |
+
name: reza-alipour/Text-Edit-Instruct-Preprocessed-4m
|
| 17 |
+
type: reza-alipour/Text-Edit-Instruct-Preprocessed-4m
|
| 18 |
+
metrics:
|
| 19 |
+
- name: Rouge1
|
| 20 |
+
type: rouge
|
| 21 |
+
value: 86.5313
|
| 22 |
---
|
| 23 |
|
| 24 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
|
|
|
| 26 |
|
| 27 |
# ft5
|
| 28 |
|
| 29 |
+
This model is a fine-tuned version of [reza-alipour/ft5](https://huggingface.co/reza-alipour/ft5) on the reza-alipour/Text-Edit-Instruct-Preprocessed-4m dataset.
|
| 30 |
It achieves the following results on the evaluation set:
|
| 31 |
- Loss: 0.3032
|
| 32 |
- Rouge1: 86.5313
|
all_results.json
CHANGED
|
@@ -1,18 +1,18 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 1.0,
|
| 3 |
-
"eval_gen_len": 24.
|
| 4 |
-
"eval_loss": 0.
|
| 5 |
-
"eval_rouge1": 86.
|
| 6 |
-
"eval_rouge2": 75.
|
| 7 |
-
"eval_rougeL": 85.
|
| 8 |
-
"eval_rougeLsum": 85.
|
| 9 |
-
"eval_runtime":
|
| 10 |
"eval_samples": 6697,
|
| 11 |
-
"eval_samples_per_second": 4.
|
| 12 |
-
"eval_steps_per_second": 1.
|
| 13 |
-
"train_loss": 0.
|
| 14 |
-
"train_runtime":
|
| 15 |
-
"train_samples":
|
| 16 |
-
"train_samples_per_second": 25.
|
| 17 |
-
"train_steps_per_second": 0.
|
| 18 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 1.0,
|
| 3 |
+
"eval_gen_len": 24.568762132298044,
|
| 4 |
+
"eval_loss": 0.3032413423061371,
|
| 5 |
+
"eval_rouge1": 86.5313,
|
| 6 |
+
"eval_rouge2": 75.3351,
|
| 7 |
+
"eval_rougeL": 85.9565,
|
| 8 |
+
"eval_rougeLsum": 85.9785,
|
| 9 |
+
"eval_runtime": 1490.9314,
|
| 10 |
"eval_samples": 6697,
|
| 11 |
+
"eval_samples_per_second": 4.492,
|
| 12 |
+
"eval_steps_per_second": 1.123,
|
| 13 |
+
"train_loss": 0.35858034201253897,
|
| 14 |
+
"train_runtime": 38927.7212,
|
| 15 |
+
"train_samples": 993343,
|
| 16 |
+
"train_samples_per_second": 25.518,
|
| 17 |
+
"train_steps_per_second": 0.797
|
| 18 |
}
|
eval_results.json
CHANGED
|
@@ -1,13 +1,13 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 1.0,
|
| 3 |
-
"eval_gen_len": 24.
|
| 4 |
-
"eval_loss": 0.
|
| 5 |
-
"eval_rouge1": 86.
|
| 6 |
-
"eval_rouge2": 75.
|
| 7 |
-
"eval_rougeL": 85.
|
| 8 |
-
"eval_rougeLsum": 85.
|
| 9 |
-
"eval_runtime":
|
| 10 |
"eval_samples": 6697,
|
| 11 |
-
"eval_samples_per_second": 4.
|
| 12 |
-
"eval_steps_per_second": 1.
|
| 13 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 1.0,
|
| 3 |
+
"eval_gen_len": 24.568762132298044,
|
| 4 |
+
"eval_loss": 0.3032413423061371,
|
| 5 |
+
"eval_rouge1": 86.5313,
|
| 6 |
+
"eval_rouge2": 75.3351,
|
| 7 |
+
"eval_rougeL": 85.9565,
|
| 8 |
+
"eval_rougeLsum": 85.9785,
|
| 9 |
+
"eval_runtime": 1490.9314,
|
| 10 |
"eval_samples": 6697,
|
| 11 |
+
"eval_samples_per_second": 4.492,
|
| 12 |
+
"eval_steps_per_second": 1.123
|
| 13 |
}
|
train_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 1.0,
|
| 3 |
-
"train_loss": 0.
|
| 4 |
-
"train_runtime":
|
| 5 |
-
"train_samples":
|
| 6 |
-
"train_samples_per_second": 25.
|
| 7 |
-
"train_steps_per_second": 0.
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 1.0,
|
| 3 |
+
"train_loss": 0.35858034201253897,
|
| 4 |
+
"train_runtime": 38927.7212,
|
| 5 |
+
"train_samples": 993343,
|
| 6 |
+
"train_samples_per_second": 25.518,
|
| 7 |
+
"train_steps_per_second": 0.797
|
| 8 |
}
|
trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch":
|
| 5 |
"eval_steps": 500,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -11,2138 +11,1892 @@
|
|
| 11 |
{
|
| 12 |
"epoch": 0.0,
|
| 13 |
"learning_rate": 7e-05,
|
| 14 |
-
"loss": 0.
|
| 15 |
"step": 100
|
| 16 |
},
|
| 17 |
{
|
| 18 |
"epoch": 0.01,
|
| 19 |
"learning_rate": 7e-05,
|
| 20 |
-
"loss": 0.
|
| 21 |
"step": 200
|
| 22 |
},
|
| 23 |
{
|
| 24 |
"epoch": 0.01,
|
| 25 |
"learning_rate": 7e-05,
|
| 26 |
-
"loss": 0.
|
| 27 |
"step": 300
|
| 28 |
},
|
| 29 |
{
|
| 30 |
"epoch": 0.01,
|
| 31 |
"learning_rate": 7e-05,
|
| 32 |
-
"loss": 0.
|
| 33 |
"step": 400
|
| 34 |
},
|
| 35 |
{
|
| 36 |
-
"epoch": 0.
|
| 37 |
"learning_rate": 7e-05,
|
| 38 |
-
"loss": 0.
|
| 39 |
"step": 500
|
| 40 |
},
|
| 41 |
{
|
| 42 |
"epoch": 0.02,
|
| 43 |
"learning_rate": 7e-05,
|
| 44 |
-
"loss": 0.
|
| 45 |
"step": 600
|
| 46 |
},
|
| 47 |
{
|
| 48 |
"epoch": 0.02,
|
| 49 |
"learning_rate": 7e-05,
|
| 50 |
-
"loss": 0.
|
| 51 |
"step": 700
|
| 52 |
},
|
| 53 |
{
|
| 54 |
-
"epoch": 0.
|
| 55 |
"learning_rate": 7e-05,
|
| 56 |
-
"loss": 0.
|
| 57 |
"step": 800
|
| 58 |
},
|
| 59 |
{
|
| 60 |
"epoch": 0.03,
|
| 61 |
"learning_rate": 7e-05,
|
| 62 |
-
"loss": 0.
|
| 63 |
"step": 900
|
| 64 |
},
|
| 65 |
{
|
| 66 |
"epoch": 0.03,
|
| 67 |
"learning_rate": 7e-05,
|
| 68 |
-
"loss": 0.
|
| 69 |
"step": 1000
|
| 70 |
},
|
| 71 |
{
|
| 72 |
-
"epoch": 0.
|
| 73 |
"learning_rate": 7e-05,
|
| 74 |
-
"loss": 0.
|
| 75 |
"step": 1100
|
| 76 |
},
|
| 77 |
{
|
| 78 |
-
"epoch": 0.
|
| 79 |
"learning_rate": 7e-05,
|
| 80 |
-
"loss": 0.
|
| 81 |
"step": 1200
|
| 82 |
},
|
| 83 |
{
|
| 84 |
"epoch": 0.04,
|
| 85 |
"learning_rate": 7e-05,
|
| 86 |
-
"loss": 0.
|
| 87 |
"step": 1300
|
| 88 |
},
|
| 89 |
{
|
| 90 |
-
"epoch": 0.
|
| 91 |
"learning_rate": 7e-05,
|
| 92 |
-
"loss": 0.
|
| 93 |
"step": 1400
|
| 94 |
},
|
| 95 |
{
|
| 96 |
-
"epoch": 0.
|
| 97 |
"learning_rate": 7e-05,
|
| 98 |
-
"loss": 0.
|
| 99 |
"step": 1500
|
| 100 |
},
|
| 101 |
{
|
| 102 |
"epoch": 0.05,
|
| 103 |
"learning_rate": 7e-05,
|
| 104 |
-
"loss": 0.
|
| 105 |
"step": 1600
|
| 106 |
},
|
| 107 |
{
|
| 108 |
"epoch": 0.05,
|
| 109 |
"learning_rate": 7e-05,
|
| 110 |
-
"loss": 0.
|
| 111 |
"step": 1700
|
| 112 |
},
|
| 113 |
{
|
| 114 |
-
"epoch": 0.
|
| 115 |
"learning_rate": 7e-05,
|
| 116 |
-
"loss": 0.
|
| 117 |
"step": 1800
|
| 118 |
},
|
| 119 |
{
|
| 120 |
-
"epoch": 0.
|
| 121 |
"learning_rate": 7e-05,
|
| 122 |
-
"loss": 0.
|
| 123 |
"step": 1900
|
| 124 |
},
|
| 125 |
{
|
| 126 |
"epoch": 0.06,
|
| 127 |
"learning_rate": 7e-05,
|
| 128 |
-
"loss": 0.
|
| 129 |
"step": 2000
|
| 130 |
},
|
| 131 |
{
|
| 132 |
-
"epoch": 0.
|
| 133 |
"learning_rate": 7e-05,
|
| 134 |
-
"loss": 0.
|
| 135 |
"step": 2100
|
| 136 |
},
|
| 137 |
{
|
| 138 |
-
"epoch": 0.
|
| 139 |
"learning_rate": 7e-05,
|
| 140 |
-
"loss": 0.
|
| 141 |
"step": 2200
|
| 142 |
},
|
| 143 |
{
|
| 144 |
"epoch": 0.07,
|
| 145 |
"learning_rate": 7e-05,
|
| 146 |
-
"loss": 0.
|
| 147 |
"step": 2300
|
| 148 |
},
|
| 149 |
{
|
| 150 |
-
"epoch": 0.
|
| 151 |
"learning_rate": 7e-05,
|
| 152 |
-
"loss": 0.
|
| 153 |
"step": 2400
|
| 154 |
},
|
| 155 |
{
|
| 156 |
-
"epoch": 0.
|
| 157 |
"learning_rate": 7e-05,
|
| 158 |
-
"loss": 0.
|
| 159 |
"step": 2500
|
| 160 |
},
|
| 161 |
{
|
| 162 |
-
"epoch": 0.
|
| 163 |
"learning_rate": 7e-05,
|
| 164 |
-
"loss": 0.
|
| 165 |
"step": 2600
|
| 166 |
},
|
| 167 |
{
|
| 168 |
-
"epoch": 0.
|
| 169 |
"learning_rate": 7e-05,
|
| 170 |
-
"loss": 0.
|
| 171 |
"step": 2700
|
| 172 |
},
|
| 173 |
{
|
| 174 |
-
"epoch": 0.
|
| 175 |
"learning_rate": 7e-05,
|
| 176 |
-
"loss": 0.
|
| 177 |
"step": 2800
|
| 178 |
},
|
| 179 |
{
|
| 180 |
-
"epoch": 0.
|
| 181 |
"learning_rate": 7e-05,
|
| 182 |
-
"loss": 0.
|
| 183 |
"step": 2900
|
| 184 |
},
|
| 185 |
{
|
| 186 |
-
"epoch": 0.
|
| 187 |
"learning_rate": 7e-05,
|
| 188 |
-
"loss": 0.
|
| 189 |
"step": 3000
|
| 190 |
},
|
| 191 |
{
|
| 192 |
-
"epoch": 0.
|
| 193 |
"learning_rate": 7e-05,
|
| 194 |
-
"loss": 0.
|
| 195 |
"step": 3100
|
| 196 |
},
|
| 197 |
{
|
| 198 |
-
"epoch": 0.
|
| 199 |
"learning_rate": 7e-05,
|
| 200 |
-
"loss": 0.
|
| 201 |
"step": 3200
|
| 202 |
},
|
| 203 |
{
|
| 204 |
-
"epoch": 0.
|
| 205 |
"learning_rate": 7e-05,
|
| 206 |
-
"loss": 0.
|
| 207 |
"step": 3300
|
| 208 |
},
|
| 209 |
{
|
| 210 |
-
"epoch": 0.
|
| 211 |
"learning_rate": 7e-05,
|
| 212 |
-
"loss": 0.
|
| 213 |
"step": 3400
|
| 214 |
},
|
| 215 |
{
|
| 216 |
-
"epoch": 0.
|
| 217 |
"learning_rate": 7e-05,
|
| 218 |
-
"loss": 0.
|
| 219 |
"step": 3500
|
| 220 |
},
|
| 221 |
{
|
| 222 |
-
"epoch": 0.
|
| 223 |
"learning_rate": 7e-05,
|
| 224 |
-
"loss": 0.
|
| 225 |
"step": 3600
|
| 226 |
},
|
| 227 |
{
|
| 228 |
-
"epoch": 0.
|
| 229 |
"learning_rate": 7e-05,
|
| 230 |
-
"loss": 0.
|
| 231 |
"step": 3700
|
| 232 |
},
|
| 233 |
{
|
| 234 |
-
"epoch": 0.
|
| 235 |
"learning_rate": 7e-05,
|
| 236 |
-
"loss": 0.
|
| 237 |
"step": 3800
|
| 238 |
},
|
| 239 |
{
|
| 240 |
-
"epoch": 0.
|
| 241 |
"learning_rate": 7e-05,
|
| 242 |
-
"loss": 0.
|
| 243 |
"step": 3900
|
| 244 |
},
|
| 245 |
{
|
| 246 |
-
"epoch": 0.
|
| 247 |
"learning_rate": 7e-05,
|
| 248 |
-
"loss": 0.
|
| 249 |
"step": 4000
|
| 250 |
},
|
| 251 |
{
|
| 252 |
-
"epoch": 0.
|
| 253 |
"learning_rate": 7e-05,
|
| 254 |
-
"loss": 0.
|
| 255 |
"step": 4100
|
| 256 |
},
|
| 257 |
{
|
| 258 |
-
"epoch": 0.
|
| 259 |
"learning_rate": 7e-05,
|
| 260 |
-
"loss": 0.
|
| 261 |
"step": 4200
|
| 262 |
},
|
| 263 |
{
|
| 264 |
-
"epoch": 0.
|
| 265 |
"learning_rate": 7e-05,
|
| 266 |
-
"loss": 0.
|
| 267 |
"step": 4300
|
| 268 |
},
|
| 269 |
{
|
| 270 |
-
"epoch": 0.
|
| 271 |
"learning_rate": 7e-05,
|
| 272 |
-
"loss": 0.
|
| 273 |
"step": 4400
|
| 274 |
},
|
| 275 |
{
|
| 276 |
-
"epoch": 0.
|
| 277 |
"learning_rate": 7e-05,
|
| 278 |
-
"loss": 0.
|
| 279 |
"step": 4500
|
| 280 |
},
|
| 281 |
{
|
| 282 |
-
"epoch": 0.
|
| 283 |
"learning_rate": 7e-05,
|
| 284 |
-
"loss": 0.
|
| 285 |
"step": 4600
|
| 286 |
},
|
| 287 |
{
|
| 288 |
-
"epoch": 0.
|
| 289 |
"learning_rate": 7e-05,
|
| 290 |
-
"loss": 0.
|
| 291 |
"step": 4700
|
| 292 |
},
|
| 293 |
{
|
| 294 |
-
"epoch": 0.
|
| 295 |
"learning_rate": 7e-05,
|
| 296 |
-
"loss": 0.
|
| 297 |
"step": 4800
|
| 298 |
},
|
| 299 |
{
|
| 300 |
-
"epoch": 0.
|
| 301 |
"learning_rate": 7e-05,
|
| 302 |
-
"loss": 0.
|
| 303 |
"step": 4900
|
| 304 |
},
|
| 305 |
{
|
| 306 |
-
"epoch": 0.
|
| 307 |
"learning_rate": 7e-05,
|
| 308 |
-
"loss": 0.
|
| 309 |
"step": 5000
|
| 310 |
},
|
| 311 |
{
|
| 312 |
-
"epoch": 0.
|
| 313 |
"learning_rate": 7e-05,
|
| 314 |
-
"loss": 0.
|
| 315 |
"step": 5100
|
| 316 |
},
|
| 317 |
{
|
| 318 |
-
"epoch": 0.
|
| 319 |
"learning_rate": 7e-05,
|
| 320 |
-
"loss": 0.
|
| 321 |
"step": 5200
|
| 322 |
},
|
| 323 |
{
|
| 324 |
-
"epoch": 0.
|
| 325 |
"learning_rate": 7e-05,
|
| 326 |
-
"loss": 0.
|
| 327 |
"step": 5300
|
| 328 |
},
|
| 329 |
{
|
| 330 |
-
"epoch": 0.
|
| 331 |
"learning_rate": 7e-05,
|
| 332 |
-
"loss": 0.
|
| 333 |
"step": 5400
|
| 334 |
},
|
| 335 |
{
|
| 336 |
-
"epoch": 0.
|
| 337 |
"learning_rate": 7e-05,
|
| 338 |
-
"loss": 0.
|
| 339 |
"step": 5500
|
| 340 |
},
|
| 341 |
{
|
| 342 |
-
"epoch": 0.
|
| 343 |
"learning_rate": 7e-05,
|
| 344 |
-
"loss": 0.
|
| 345 |
"step": 5600
|
| 346 |
},
|
| 347 |
{
|
| 348 |
-
"epoch": 0.
|
| 349 |
"learning_rate": 7e-05,
|
| 350 |
-
"loss": 0.
|
| 351 |
"step": 5700
|
| 352 |
},
|
| 353 |
{
|
| 354 |
-
"epoch": 0.
|
| 355 |
"learning_rate": 7e-05,
|
| 356 |
-
"loss": 0.
|
| 357 |
"step": 5800
|
| 358 |
},
|
| 359 |
{
|
| 360 |
-
"epoch": 0.
|
| 361 |
"learning_rate": 7e-05,
|
| 362 |
-
"loss": 0.
|
| 363 |
"step": 5900
|
| 364 |
},
|
| 365 |
{
|
| 366 |
-
"epoch": 0.
|
| 367 |
"learning_rate": 7e-05,
|
| 368 |
-
"loss": 0.
|
| 369 |
"step": 6000
|
| 370 |
},
|
| 371 |
{
|
| 372 |
-
"epoch": 0.
|
| 373 |
"learning_rate": 7e-05,
|
| 374 |
-
"loss": 0.
|
| 375 |
"step": 6100
|
| 376 |
},
|
| 377 |
{
|
| 378 |
-
"epoch": 0.
|
| 379 |
"learning_rate": 7e-05,
|
| 380 |
-
"loss": 0.
|
| 381 |
"step": 6200
|
| 382 |
},
|
| 383 |
{
|
| 384 |
-
"epoch": 0.
|
| 385 |
"learning_rate": 7e-05,
|
| 386 |
-
"loss": 0.
|
| 387 |
"step": 6300
|
| 388 |
},
|
| 389 |
{
|
| 390 |
-
"epoch": 0.
|
| 391 |
"learning_rate": 7e-05,
|
| 392 |
-
"loss": 0.
|
| 393 |
"step": 6400
|
| 394 |
},
|
| 395 |
{
|
| 396 |
-
"epoch": 0.
|
| 397 |
"learning_rate": 7e-05,
|
| 398 |
-
"loss": 0.
|
| 399 |
"step": 6500
|
| 400 |
},
|
| 401 |
{
|
| 402 |
-
"epoch": 0.
|
| 403 |
"learning_rate": 7e-05,
|
| 404 |
-
"loss": 0.
|
| 405 |
"step": 6600
|
| 406 |
},
|
| 407 |
{
|
| 408 |
-
"epoch": 0.
|
| 409 |
"learning_rate": 7e-05,
|
| 410 |
-
"loss": 0.
|
| 411 |
"step": 6700
|
| 412 |
},
|
| 413 |
{
|
| 414 |
-
"epoch": 0.
|
| 415 |
"learning_rate": 7e-05,
|
| 416 |
-
"loss": 0.
|
| 417 |
"step": 6800
|
| 418 |
},
|
| 419 |
{
|
| 420 |
-
"epoch": 0.
|
| 421 |
"learning_rate": 7e-05,
|
| 422 |
-
"loss": 0.
|
| 423 |
"step": 6900
|
| 424 |
},
|
| 425 |
{
|
| 426 |
-
"epoch": 0.
|
| 427 |
"learning_rate": 7e-05,
|
| 428 |
-
"loss": 0.
|
| 429 |
"step": 7000
|
| 430 |
},
|
| 431 |
{
|
| 432 |
-
"epoch": 0.
|
| 433 |
"learning_rate": 7e-05,
|
| 434 |
-
"loss": 0.
|
| 435 |
"step": 7100
|
| 436 |
},
|
| 437 |
{
|
| 438 |
-
"epoch": 0.
|
| 439 |
"learning_rate": 7e-05,
|
| 440 |
-
"loss": 0.
|
| 441 |
"step": 7200
|
| 442 |
},
|
| 443 |
{
|
| 444 |
-
"epoch": 0.
|
| 445 |
"learning_rate": 7e-05,
|
| 446 |
-
"loss": 0.
|
| 447 |
"step": 7300
|
| 448 |
},
|
| 449 |
{
|
| 450 |
-
"epoch": 0.
|
| 451 |
"learning_rate": 7e-05,
|
| 452 |
-
"loss": 0.
|
| 453 |
"step": 7400
|
| 454 |
},
|
| 455 |
{
|
| 456 |
-
"epoch": 0.
|
| 457 |
"learning_rate": 7e-05,
|
| 458 |
-
"loss": 0.
|
| 459 |
"step": 7500
|
| 460 |
},
|
| 461 |
{
|
| 462 |
-
"epoch": 0.
|
| 463 |
"learning_rate": 7e-05,
|
| 464 |
-
"loss": 0.
|
| 465 |
"step": 7600
|
| 466 |
},
|
| 467 |
{
|
| 468 |
-
"epoch": 0.
|
| 469 |
"learning_rate": 7e-05,
|
| 470 |
-
"loss": 0.
|
| 471 |
"step": 7700
|
| 472 |
},
|
| 473 |
{
|
| 474 |
-
"epoch": 0.
|
| 475 |
"learning_rate": 7e-05,
|
| 476 |
-
"loss": 0.
|
| 477 |
"step": 7800
|
| 478 |
},
|
| 479 |
{
|
| 480 |
-
"epoch": 0.
|
| 481 |
"learning_rate": 7e-05,
|
| 482 |
-
"loss": 0.
|
| 483 |
"step": 7900
|
| 484 |
},
|
| 485 |
{
|
| 486 |
-
"epoch": 0.
|
| 487 |
"learning_rate": 7e-05,
|
| 488 |
-
"loss": 0.
|
| 489 |
"step": 8000
|
| 490 |
},
|
| 491 |
{
|
| 492 |
-
"epoch": 0.
|
| 493 |
"learning_rate": 7e-05,
|
| 494 |
-
"loss": 0.
|
| 495 |
"step": 8100
|
| 496 |
},
|
| 497 |
{
|
| 498 |
-
"epoch": 0.
|
| 499 |
"learning_rate": 7e-05,
|
| 500 |
-
"loss": 0.
|
| 501 |
"step": 8200
|
| 502 |
},
|
| 503 |
{
|
| 504 |
-
"epoch": 0.
|
| 505 |
"learning_rate": 7e-05,
|
| 506 |
-
"loss": 0.
|
| 507 |
"step": 8300
|
| 508 |
},
|
| 509 |
{
|
| 510 |
-
"epoch": 0.
|
| 511 |
"learning_rate": 7e-05,
|
| 512 |
-
"loss": 0.
|
| 513 |
"step": 8400
|
| 514 |
},
|
| 515 |
{
|
| 516 |
-
"epoch": 0.
|
| 517 |
"learning_rate": 7e-05,
|
| 518 |
-
"loss": 0.
|
| 519 |
"step": 8500
|
| 520 |
},
|
| 521 |
{
|
| 522 |
-
"epoch": 0.
|
| 523 |
"learning_rate": 7e-05,
|
| 524 |
-
"loss": 0.
|
| 525 |
"step": 8600
|
| 526 |
},
|
| 527 |
{
|
| 528 |
-
"epoch": 0.
|
| 529 |
"learning_rate": 7e-05,
|
| 530 |
-
"loss": 0.
|
| 531 |
"step": 8700
|
| 532 |
},
|
| 533 |
{
|
| 534 |
-
"epoch": 0.
|
| 535 |
"learning_rate": 7e-05,
|
| 536 |
-
"loss": 0.
|
| 537 |
"step": 8800
|
| 538 |
},
|
| 539 |
{
|
| 540 |
-
"epoch": 0.
|
| 541 |
"learning_rate": 7e-05,
|
| 542 |
-
"loss": 0.
|
| 543 |
"step": 8900
|
| 544 |
},
|
| 545 |
{
|
| 546 |
-
"epoch": 0.
|
| 547 |
"learning_rate": 7e-05,
|
| 548 |
-
"loss": 0.
|
| 549 |
"step": 9000
|
| 550 |
},
|
| 551 |
{
|
| 552 |
-
"epoch": 0.
|
| 553 |
"learning_rate": 7e-05,
|
| 554 |
-
"loss": 0.
|
| 555 |
"step": 9100
|
| 556 |
},
|
| 557 |
{
|
| 558 |
-
"epoch": 0.
|
| 559 |
"learning_rate": 7e-05,
|
| 560 |
-
"loss": 0.
|
| 561 |
"step": 9200
|
| 562 |
},
|
| 563 |
{
|
| 564 |
-
"epoch": 0.
|
| 565 |
"learning_rate": 7e-05,
|
| 566 |
-
"loss": 0.
|
| 567 |
"step": 9300
|
| 568 |
},
|
| 569 |
{
|
| 570 |
-
"epoch": 0.
|
| 571 |
"learning_rate": 7e-05,
|
| 572 |
-
"loss": 0.
|
| 573 |
"step": 9400
|
| 574 |
},
|
| 575 |
{
|
| 576 |
-
"epoch": 0.
|
| 577 |
"learning_rate": 7e-05,
|
| 578 |
-
"loss": 0.
|
| 579 |
"step": 9500
|
| 580 |
},
|
| 581 |
{
|
| 582 |
-
"epoch": 0.
|
| 583 |
"learning_rate": 7e-05,
|
| 584 |
-
"loss": 0.
|
| 585 |
"step": 9600
|
| 586 |
},
|
| 587 |
{
|
| 588 |
-
"epoch": 0.
|
| 589 |
"learning_rate": 7e-05,
|
| 590 |
-
"loss": 0.
|
| 591 |
"step": 9700
|
| 592 |
},
|
| 593 |
{
|
| 594 |
-
"epoch": 0.
|
| 595 |
"learning_rate": 7e-05,
|
| 596 |
-
"loss": 0.
|
| 597 |
"step": 9800
|
| 598 |
},
|
| 599 |
{
|
| 600 |
-
"epoch": 0.
|
| 601 |
"learning_rate": 7e-05,
|
| 602 |
-
"loss": 0.
|
| 603 |
"step": 9900
|
| 604 |
},
|
| 605 |
{
|
| 606 |
-
"epoch": 0.
|
| 607 |
"learning_rate": 7e-05,
|
| 608 |
-
"loss": 0.
|
| 609 |
"step": 10000
|
| 610 |
},
|
| 611 |
{
|
| 612 |
-
"epoch": 0.
|
| 613 |
"learning_rate": 7e-05,
|
| 614 |
-
"loss": 0.
|
| 615 |
"step": 10100
|
| 616 |
},
|
| 617 |
{
|
| 618 |
-
"epoch": 0.
|
| 619 |
"learning_rate": 7e-05,
|
| 620 |
-
"loss": 0.
|
| 621 |
"step": 10200
|
| 622 |
},
|
| 623 |
{
|
| 624 |
-
"epoch": 0.
|
| 625 |
"learning_rate": 7e-05,
|
| 626 |
-
"loss": 0.
|
| 627 |
"step": 10300
|
| 628 |
},
|
| 629 |
{
|
| 630 |
-
"epoch": 0.
|
| 631 |
"learning_rate": 7e-05,
|
| 632 |
-
"loss": 0.
|
| 633 |
"step": 10400
|
| 634 |
},
|
| 635 |
{
|
| 636 |
-
"epoch": 0.
|
| 637 |
"learning_rate": 7e-05,
|
| 638 |
-
"loss": 0.
|
| 639 |
"step": 10500
|
| 640 |
},
|
| 641 |
{
|
| 642 |
-
"epoch": 0.
|
| 643 |
"learning_rate": 7e-05,
|
| 644 |
-
"loss": 0.
|
| 645 |
"step": 10600
|
| 646 |
},
|
| 647 |
{
|
| 648 |
-
"epoch": 0.
|
| 649 |
"learning_rate": 7e-05,
|
| 650 |
-
"loss": 0.
|
| 651 |
"step": 10700
|
| 652 |
},
|
| 653 |
{
|
| 654 |
-
"epoch": 0.
|
| 655 |
"learning_rate": 7e-05,
|
| 656 |
-
"loss": 0.
|
| 657 |
"step": 10800
|
| 658 |
},
|
| 659 |
{
|
| 660 |
-
"epoch": 0.
|
| 661 |
"learning_rate": 7e-05,
|
| 662 |
-
"loss": 0.
|
| 663 |
"step": 10900
|
| 664 |
},
|
| 665 |
{
|
| 666 |
-
"epoch": 0.
|
| 667 |
"learning_rate": 7e-05,
|
| 668 |
-
"loss": 0.
|
| 669 |
"step": 11000
|
| 670 |
},
|
| 671 |
{
|
| 672 |
-
"epoch": 0.
|
| 673 |
"learning_rate": 7e-05,
|
| 674 |
-
"loss": 0.
|
| 675 |
"step": 11100
|
| 676 |
},
|
| 677 |
{
|
| 678 |
-
"epoch": 0.
|
| 679 |
"learning_rate": 7e-05,
|
| 680 |
-
"loss": 0.
|
| 681 |
"step": 11200
|
| 682 |
},
|
| 683 |
{
|
| 684 |
-
"epoch": 0.
|
| 685 |
"learning_rate": 7e-05,
|
| 686 |
-
"loss": 0.
|
| 687 |
"step": 11300
|
| 688 |
},
|
| 689 |
{
|
| 690 |
-
"epoch": 0.
|
| 691 |
"learning_rate": 7e-05,
|
| 692 |
-
"loss": 0.
|
| 693 |
"step": 11400
|
| 694 |
},
|
| 695 |
{
|
| 696 |
-
"epoch": 0.
|
| 697 |
"learning_rate": 7e-05,
|
| 698 |
-
"loss": 0.
|
| 699 |
"step": 11500
|
| 700 |
},
|
| 701 |
{
|
| 702 |
-
"epoch": 0.
|
| 703 |
"learning_rate": 7e-05,
|
| 704 |
-
"loss": 0.
|
| 705 |
"step": 11600
|
| 706 |
},
|
| 707 |
{
|
| 708 |
-
"epoch": 0.
|
| 709 |
"learning_rate": 7e-05,
|
| 710 |
-
"loss": 0.
|
| 711 |
"step": 11700
|
| 712 |
},
|
| 713 |
{
|
| 714 |
-
"epoch": 0.
|
| 715 |
"learning_rate": 7e-05,
|
| 716 |
-
"loss": 0.
|
| 717 |
"step": 11800
|
| 718 |
},
|
| 719 |
{
|
| 720 |
-
"epoch": 0.
|
| 721 |
"learning_rate": 7e-05,
|
| 722 |
-
"loss": 0.
|
| 723 |
"step": 11900
|
| 724 |
},
|
| 725 |
{
|
| 726 |
-
"epoch": 0.
|
| 727 |
"learning_rate": 7e-05,
|
| 728 |
-
"loss": 0.
|
| 729 |
"step": 12000
|
| 730 |
},
|
| 731 |
{
|
| 732 |
-
"epoch": 0.
|
| 733 |
"learning_rate": 7e-05,
|
| 734 |
-
"loss": 0.
|
| 735 |
"step": 12100
|
| 736 |
},
|
| 737 |
{
|
| 738 |
-
"epoch": 0.
|
| 739 |
"learning_rate": 7e-05,
|
| 740 |
-
"loss": 0.
|
| 741 |
"step": 12200
|
| 742 |
},
|
| 743 |
{
|
| 744 |
-
"epoch": 0.
|
| 745 |
"learning_rate": 7e-05,
|
| 746 |
-
"loss": 0.
|
| 747 |
"step": 12300
|
| 748 |
},
|
| 749 |
{
|
| 750 |
-
"epoch": 0.
|
| 751 |
"learning_rate": 7e-05,
|
| 752 |
-
"loss": 0.
|
| 753 |
"step": 12400
|
| 754 |
},
|
| 755 |
{
|
| 756 |
-
"epoch": 0.
|
| 757 |
"learning_rate": 7e-05,
|
| 758 |
-
"loss": 0.
|
| 759 |
"step": 12500
|
| 760 |
},
|
| 761 |
{
|
| 762 |
-
"epoch": 0.
|
| 763 |
"learning_rate": 7e-05,
|
| 764 |
-
"loss": 0.
|
| 765 |
"step": 12600
|
| 766 |
},
|
| 767 |
{
|
| 768 |
-
"epoch": 0.
|
| 769 |
"learning_rate": 7e-05,
|
| 770 |
-
"loss": 0.
|
| 771 |
"step": 12700
|
| 772 |
},
|
| 773 |
{
|
| 774 |
-
"epoch": 0.
|
| 775 |
"learning_rate": 7e-05,
|
| 776 |
-
"loss": 0.
|
| 777 |
"step": 12800
|
| 778 |
},
|
| 779 |
{
|
| 780 |
-
"epoch": 0.
|
| 781 |
"learning_rate": 7e-05,
|
| 782 |
-
"loss": 0.
|
| 783 |
"step": 12900
|
| 784 |
},
|
| 785 |
{
|
| 786 |
-
"epoch": 0.
|
| 787 |
"learning_rate": 7e-05,
|
| 788 |
-
"loss": 0.
|
| 789 |
"step": 13000
|
| 790 |
},
|
| 791 |
{
|
| 792 |
-
"epoch": 0.
|
| 793 |
"learning_rate": 7e-05,
|
| 794 |
-
"loss": 0.
|
| 795 |
"step": 13100
|
| 796 |
},
|
| 797 |
{
|
| 798 |
-
"epoch": 0.
|
| 799 |
"learning_rate": 7e-05,
|
| 800 |
-
"loss": 0.
|
| 801 |
"step": 13200
|
| 802 |
},
|
| 803 |
{
|
| 804 |
-
"epoch": 0.
|
| 805 |
"learning_rate": 7e-05,
|
| 806 |
-
"loss": 0.
|
| 807 |
"step": 13300
|
| 808 |
},
|
| 809 |
{
|
| 810 |
-
"epoch": 0.
|
| 811 |
"learning_rate": 7e-05,
|
| 812 |
-
"loss": 0.
|
| 813 |
"step": 13400
|
| 814 |
},
|
| 815 |
{
|
| 816 |
-
"epoch": 0.
|
| 817 |
"learning_rate": 7e-05,
|
| 818 |
-
"loss": 0.
|
| 819 |
"step": 13500
|
| 820 |
},
|
| 821 |
{
|
| 822 |
-
"epoch": 0.
|
| 823 |
"learning_rate": 7e-05,
|
| 824 |
-
"loss": 0.
|
| 825 |
"step": 13600
|
| 826 |
},
|
| 827 |
{
|
| 828 |
-
"epoch": 0.
|
| 829 |
"learning_rate": 7e-05,
|
| 830 |
-
"loss": 0.
|
| 831 |
"step": 13700
|
| 832 |
},
|
| 833 |
{
|
| 834 |
-
"epoch": 0.
|
| 835 |
"learning_rate": 7e-05,
|
| 836 |
-
"loss": 0.
|
| 837 |
"step": 13800
|
| 838 |
},
|
| 839 |
{
|
| 840 |
-
"epoch": 0.
|
| 841 |
"learning_rate": 7e-05,
|
| 842 |
-
"loss": 0.
|
| 843 |
"step": 13900
|
| 844 |
},
|
| 845 |
{
|
| 846 |
-
"epoch": 0.
|
| 847 |
"learning_rate": 7e-05,
|
| 848 |
-
"loss": 0.
|
| 849 |
"step": 14000
|
| 850 |
},
|
| 851 |
{
|
| 852 |
-
"epoch": 0.
|
| 853 |
"learning_rate": 7e-05,
|
| 854 |
-
"loss": 0.
|
| 855 |
"step": 14100
|
| 856 |
},
|
| 857 |
{
|
| 858 |
-
"epoch": 0.
|
| 859 |
"learning_rate": 7e-05,
|
| 860 |
-
"loss": 0.
|
| 861 |
"step": 14200
|
| 862 |
},
|
| 863 |
{
|
| 864 |
-
"epoch": 0.
|
| 865 |
"learning_rate": 7e-05,
|
| 866 |
-
"loss": 0.
|
| 867 |
"step": 14300
|
| 868 |
},
|
| 869 |
{
|
| 870 |
-
"epoch": 0.
|
| 871 |
"learning_rate": 7e-05,
|
| 872 |
-
"loss": 0.
|
| 873 |
"step": 14400
|
| 874 |
},
|
| 875 |
{
|
| 876 |
-
"epoch": 0.
|
| 877 |
"learning_rate": 7e-05,
|
| 878 |
-
"loss": 0.
|
| 879 |
"step": 14500
|
| 880 |
},
|
| 881 |
{
|
| 882 |
-
"epoch": 0.
|
| 883 |
"learning_rate": 7e-05,
|
| 884 |
-
"loss": 0.
|
| 885 |
"step": 14600
|
| 886 |
},
|
| 887 |
{
|
| 888 |
-
"epoch": 0.
|
| 889 |
"learning_rate": 7e-05,
|
| 890 |
-
"loss": 0.
|
| 891 |
"step": 14700
|
| 892 |
},
|
| 893 |
{
|
| 894 |
-
"epoch": 0.
|
| 895 |
"learning_rate": 7e-05,
|
| 896 |
-
"loss": 0.
|
| 897 |
"step": 14800
|
| 898 |
},
|
| 899 |
{
|
| 900 |
-
"epoch": 0.
|
| 901 |
"learning_rate": 7e-05,
|
| 902 |
-
"loss": 0.
|
| 903 |
"step": 14900
|
| 904 |
},
|
| 905 |
{
|
| 906 |
-
"epoch": 0.
|
| 907 |
"learning_rate": 7e-05,
|
| 908 |
-
"loss": 0.
|
| 909 |
"step": 15000
|
| 910 |
},
|
| 911 |
{
|
| 912 |
-
"epoch": 0.
|
| 913 |
"learning_rate": 7e-05,
|
| 914 |
-
"loss": 0.
|
| 915 |
"step": 15100
|
| 916 |
},
|
| 917 |
{
|
| 918 |
-
"epoch": 0.
|
| 919 |
"learning_rate": 7e-05,
|
| 920 |
-
"loss": 0.
|
| 921 |
"step": 15200
|
| 922 |
},
|
| 923 |
{
|
| 924 |
-
"epoch": 0.
|
| 925 |
"learning_rate": 7e-05,
|
| 926 |
-
"loss": 0.
|
| 927 |
"step": 15300
|
| 928 |
},
|
| 929 |
{
|
| 930 |
-
"epoch": 0.
|
| 931 |
"learning_rate": 7e-05,
|
| 932 |
-
"loss": 0.
|
| 933 |
"step": 15400
|
| 934 |
},
|
| 935 |
{
|
| 936 |
-
"epoch": 0.
|
| 937 |
"learning_rate": 7e-05,
|
| 938 |
-
"loss": 0.
|
| 939 |
"step": 15500
|
| 940 |
},
|
| 941 |
{
|
| 942 |
-
"epoch": 0.
|
| 943 |
"learning_rate": 7e-05,
|
| 944 |
-
"loss": 0.
|
| 945 |
"step": 15600
|
| 946 |
},
|
| 947 |
{
|
| 948 |
-
"epoch": 0.
|
| 949 |
"learning_rate": 7e-05,
|
| 950 |
-
"loss": 0.
|
| 951 |
"step": 15700
|
| 952 |
},
|
| 953 |
{
|
| 954 |
-
"epoch": 0.
|
| 955 |
"learning_rate": 7e-05,
|
| 956 |
-
"loss": 0.
|
| 957 |
"step": 15800
|
| 958 |
},
|
| 959 |
{
|
| 960 |
-
"epoch": 0.
|
| 961 |
"learning_rate": 7e-05,
|
| 962 |
-
"loss": 0.
|
| 963 |
"step": 15900
|
| 964 |
},
|
| 965 |
{
|
| 966 |
-
"epoch": 0.
|
| 967 |
"learning_rate": 7e-05,
|
| 968 |
-
"loss": 0.
|
| 969 |
"step": 16000
|
| 970 |
},
|
| 971 |
{
|
| 972 |
-
"epoch": 0.
|
| 973 |
"learning_rate": 7e-05,
|
| 974 |
-
"loss": 0.
|
| 975 |
"step": 16100
|
| 976 |
},
|
| 977 |
{
|
| 978 |
-
"epoch": 0.
|
| 979 |
"learning_rate": 7e-05,
|
| 980 |
-
"loss": 0.
|
| 981 |
"step": 16200
|
| 982 |
},
|
| 983 |
{
|
| 984 |
-
"epoch": 0.
|
| 985 |
"learning_rate": 7e-05,
|
| 986 |
-
"loss": 0.
|
| 987 |
"step": 16300
|
| 988 |
},
|
| 989 |
{
|
| 990 |
-
"epoch": 0.
|
| 991 |
"learning_rate": 7e-05,
|
| 992 |
-
"loss": 0.
|
| 993 |
"step": 16400
|
| 994 |
},
|
| 995 |
{
|
| 996 |
-
"epoch": 0.
|
| 997 |
"learning_rate": 7e-05,
|
| 998 |
-
"loss": 0.
|
| 999 |
"step": 16500
|
| 1000 |
},
|
| 1001 |
{
|
| 1002 |
-
"epoch": 0.
|
| 1003 |
"learning_rate": 7e-05,
|
| 1004 |
-
"loss": 0.
|
| 1005 |
"step": 16600
|
| 1006 |
},
|
| 1007 |
{
|
| 1008 |
-
"epoch": 0.
|
| 1009 |
"learning_rate": 7e-05,
|
| 1010 |
-
"loss": 0.
|
| 1011 |
"step": 16700
|
| 1012 |
},
|
| 1013 |
{
|
| 1014 |
-
"epoch": 0.
|
| 1015 |
"learning_rate": 7e-05,
|
| 1016 |
-
"loss": 0.
|
| 1017 |
"step": 16800
|
| 1018 |
},
|
| 1019 |
{
|
| 1020 |
-
"epoch": 0.
|
| 1021 |
"learning_rate": 7e-05,
|
| 1022 |
-
"loss": 0.
|
| 1023 |
"step": 16900
|
| 1024 |
},
|
| 1025 |
{
|
| 1026 |
-
"epoch": 0.
|
| 1027 |
"learning_rate": 7e-05,
|
| 1028 |
-
"loss": 0.
|
| 1029 |
"step": 17000
|
| 1030 |
},
|
| 1031 |
{
|
| 1032 |
-
"epoch": 0.
|
| 1033 |
"learning_rate": 7e-05,
|
| 1034 |
-
"loss": 0.
|
| 1035 |
"step": 17100
|
| 1036 |
},
|
| 1037 |
{
|
| 1038 |
-
"epoch": 0.
|
| 1039 |
"learning_rate": 7e-05,
|
| 1040 |
-
"loss": 0.
|
| 1041 |
"step": 17200
|
| 1042 |
},
|
| 1043 |
{
|
| 1044 |
-
"epoch": 0.
|
| 1045 |
"learning_rate": 7e-05,
|
| 1046 |
-
"loss": 0.
|
| 1047 |
"step": 17300
|
| 1048 |
},
|
| 1049 |
{
|
| 1050 |
-
"epoch": 0.
|
| 1051 |
"learning_rate": 7e-05,
|
| 1052 |
-
"loss": 0.
|
| 1053 |
"step": 17400
|
| 1054 |
},
|
| 1055 |
{
|
| 1056 |
-
"epoch": 0.
|
| 1057 |
"learning_rate": 7e-05,
|
| 1058 |
-
"loss": 0.
|
| 1059 |
"step": 17500
|
| 1060 |
},
|
| 1061 |
{
|
| 1062 |
-
"epoch": 0.
|
| 1063 |
"learning_rate": 7e-05,
|
| 1064 |
-
"loss": 0.
|
| 1065 |
"step": 17600
|
| 1066 |
},
|
| 1067 |
{
|
| 1068 |
-
"epoch": 0.
|
| 1069 |
"learning_rate": 7e-05,
|
| 1070 |
-
"loss": 0.
|
| 1071 |
"step": 17700
|
| 1072 |
},
|
| 1073 |
{
|
| 1074 |
-
"epoch": 0.
|
| 1075 |
"learning_rate": 7e-05,
|
| 1076 |
-
"loss": 0.
|
| 1077 |
"step": 17800
|
| 1078 |
},
|
| 1079 |
{
|
| 1080 |
-
"epoch": 0.
|
| 1081 |
"learning_rate": 7e-05,
|
| 1082 |
-
"loss": 0.
|
| 1083 |
"step": 17900
|
| 1084 |
},
|
| 1085 |
{
|
| 1086 |
-
"epoch": 0.
|
| 1087 |
"learning_rate": 7e-05,
|
| 1088 |
-
"loss": 0.
|
| 1089 |
"step": 18000
|
| 1090 |
},
|
| 1091 |
{
|
| 1092 |
-
"epoch": 0.
|
| 1093 |
"learning_rate": 7e-05,
|
| 1094 |
-
"loss": 0.
|
| 1095 |
"step": 18100
|
| 1096 |
},
|
| 1097 |
{
|
| 1098 |
-
"epoch": 0.
|
| 1099 |
"learning_rate": 7e-05,
|
| 1100 |
-
"loss": 0.
|
| 1101 |
"step": 18200
|
| 1102 |
},
|
| 1103 |
{
|
| 1104 |
-
"epoch": 0.
|
| 1105 |
"learning_rate": 7e-05,
|
| 1106 |
-
"loss": 0.
|
| 1107 |
"step": 18300
|
| 1108 |
},
|
| 1109 |
{
|
| 1110 |
-
"epoch": 0.
|
| 1111 |
"learning_rate": 7e-05,
|
| 1112 |
-
"loss": 0.
|
| 1113 |
"step": 18400
|
| 1114 |
},
|
| 1115 |
{
|
| 1116 |
-
"epoch": 0.
|
| 1117 |
"learning_rate": 7e-05,
|
| 1118 |
-
"loss": 0.
|
| 1119 |
"step": 18500
|
| 1120 |
},
|
| 1121 |
{
|
| 1122 |
-
"epoch": 0.
|
| 1123 |
"learning_rate": 7e-05,
|
| 1124 |
-
"loss": 0.
|
| 1125 |
"step": 18600
|
| 1126 |
},
|
| 1127 |
{
|
| 1128 |
-
"epoch": 0.
|
| 1129 |
"learning_rate": 7e-05,
|
| 1130 |
-
"loss": 0.
|
| 1131 |
"step": 18700
|
| 1132 |
},
|
| 1133 |
{
|
| 1134 |
-
"epoch": 0.
|
| 1135 |
"learning_rate": 7e-05,
|
| 1136 |
-
"loss": 0.
|
| 1137 |
"step": 18800
|
| 1138 |
},
|
| 1139 |
{
|
| 1140 |
-
"epoch": 0.
|
| 1141 |
"learning_rate": 7e-05,
|
| 1142 |
-
"loss": 0.
|
| 1143 |
"step": 18900
|
| 1144 |
},
|
| 1145 |
{
|
| 1146 |
-
"epoch": 0.
|
| 1147 |
"learning_rate": 7e-05,
|
| 1148 |
-
"loss": 0.
|
| 1149 |
"step": 19000
|
| 1150 |
},
|
| 1151 |
{
|
| 1152 |
-
"epoch": 0.
|
| 1153 |
"learning_rate": 7e-05,
|
| 1154 |
-
"loss": 0.
|
| 1155 |
"step": 19100
|
| 1156 |
},
|
| 1157 |
{
|
| 1158 |
-
"epoch": 0.
|
| 1159 |
"learning_rate": 7e-05,
|
| 1160 |
-
"loss": 0.
|
| 1161 |
"step": 19200
|
| 1162 |
},
|
| 1163 |
{
|
| 1164 |
-
"epoch": 0.
|
| 1165 |
"learning_rate": 7e-05,
|
| 1166 |
-
"loss": 0.
|
| 1167 |
"step": 19300
|
| 1168 |
},
|
| 1169 |
{
|
| 1170 |
-
"epoch": 0.
|
| 1171 |
"learning_rate": 7e-05,
|
| 1172 |
-
"loss": 0.
|
| 1173 |
"step": 19400
|
| 1174 |
},
|
| 1175 |
{
|
| 1176 |
-
"epoch": 0.
|
| 1177 |
"learning_rate": 7e-05,
|
| 1178 |
-
"loss": 0.
|
| 1179 |
"step": 19500
|
| 1180 |
},
|
| 1181 |
{
|
| 1182 |
-
"epoch": 0.
|
| 1183 |
"learning_rate": 7e-05,
|
| 1184 |
-
"loss": 0.
|
| 1185 |
"step": 19600
|
| 1186 |
},
|
| 1187 |
{
|
| 1188 |
-
"epoch": 0.
|
| 1189 |
"learning_rate": 7e-05,
|
| 1190 |
-
"loss": 0.
|
| 1191 |
"step": 19700
|
| 1192 |
},
|
| 1193 |
{
|
| 1194 |
-
"epoch": 0.
|
| 1195 |
"learning_rate": 7e-05,
|
| 1196 |
-
"loss": 0.
|
| 1197 |
"step": 19800
|
| 1198 |
},
|
| 1199 |
{
|
| 1200 |
-
"epoch": 0.
|
| 1201 |
"learning_rate": 7e-05,
|
| 1202 |
-
"loss": 0.
|
| 1203 |
"step": 19900
|
| 1204 |
},
|
| 1205 |
{
|
| 1206 |
-
"epoch": 0.
|
| 1207 |
"learning_rate": 7e-05,
|
| 1208 |
-
"loss": 0.
|
| 1209 |
"step": 20000
|
| 1210 |
},
|
| 1211 |
{
|
| 1212 |
-
"epoch": 0.
|
| 1213 |
"learning_rate": 7e-05,
|
| 1214 |
-
"loss": 0.
|
| 1215 |
"step": 20100
|
| 1216 |
},
|
| 1217 |
{
|
| 1218 |
-
"epoch": 0.
|
| 1219 |
"learning_rate": 7e-05,
|
| 1220 |
-
"loss": 0.
|
| 1221 |
"step": 20200
|
| 1222 |
},
|
| 1223 |
{
|
| 1224 |
-
"epoch": 0.
|
| 1225 |
"learning_rate": 7e-05,
|
| 1226 |
-
"loss": 0.
|
| 1227 |
"step": 20300
|
| 1228 |
},
|
| 1229 |
{
|
| 1230 |
-
"epoch": 0.
|
| 1231 |
"learning_rate": 7e-05,
|
| 1232 |
-
"loss": 0.
|
| 1233 |
"step": 20400
|
| 1234 |
},
|
| 1235 |
{
|
| 1236 |
-
"epoch": 0.
|
| 1237 |
"learning_rate": 7e-05,
|
| 1238 |
-
"loss": 0.
|
| 1239 |
"step": 20500
|
| 1240 |
},
|
| 1241 |
{
|
| 1242 |
-
"epoch": 0.
|
| 1243 |
"learning_rate": 7e-05,
|
| 1244 |
-
"loss": 0.
|
| 1245 |
"step": 20600
|
| 1246 |
},
|
| 1247 |
{
|
| 1248 |
-
"epoch": 0.
|
| 1249 |
"learning_rate": 7e-05,
|
| 1250 |
-
"loss": 0.
|
| 1251 |
"step": 20700
|
| 1252 |
},
|
| 1253 |
{
|
| 1254 |
-
"epoch": 0.
|
| 1255 |
"learning_rate": 7e-05,
|
| 1256 |
-
"loss": 0.
|
| 1257 |
"step": 20800
|
| 1258 |
},
|
| 1259 |
{
|
| 1260 |
-
"epoch": 0.
|
| 1261 |
"learning_rate": 7e-05,
|
| 1262 |
-
"loss": 0.
|
| 1263 |
"step": 20900
|
| 1264 |
},
|
| 1265 |
{
|
| 1266 |
-
"epoch": 0.
|
| 1267 |
"learning_rate": 7e-05,
|
| 1268 |
-
"loss": 0.
|
| 1269 |
"step": 21000
|
| 1270 |
},
|
| 1271 |
{
|
| 1272 |
-
"epoch": 0.
|
| 1273 |
"learning_rate": 7e-05,
|
| 1274 |
-
"loss": 0.
|
| 1275 |
"step": 21100
|
| 1276 |
},
|
| 1277 |
{
|
| 1278 |
-
"epoch": 0.
|
| 1279 |
"learning_rate": 7e-05,
|
| 1280 |
-
"loss": 0.
|
| 1281 |
"step": 21200
|
| 1282 |
},
|
| 1283 |
{
|
| 1284 |
-
"epoch": 0.
|
| 1285 |
"learning_rate": 7e-05,
|
| 1286 |
-
"loss": 0.
|
| 1287 |
"step": 21300
|
| 1288 |
},
|
| 1289 |
{
|
| 1290 |
-
"epoch": 0.
|
| 1291 |
"learning_rate": 7e-05,
|
| 1292 |
-
"loss": 0.
|
| 1293 |
"step": 21400
|
| 1294 |
},
|
| 1295 |
{
|
| 1296 |
-
"epoch": 0.
|
| 1297 |
"learning_rate": 7e-05,
|
| 1298 |
-
"loss": 0.
|
| 1299 |
"step": 21500
|
| 1300 |
},
|
| 1301 |
{
|
| 1302 |
-
"epoch": 0.
|
| 1303 |
"learning_rate": 7e-05,
|
| 1304 |
-
"loss": 0.
|
| 1305 |
"step": 21600
|
| 1306 |
},
|
| 1307 |
{
|
| 1308 |
-
"epoch": 0.
|
| 1309 |
"learning_rate": 7e-05,
|
| 1310 |
-
"loss": 0.
|
| 1311 |
"step": 21700
|
| 1312 |
},
|
| 1313 |
{
|
| 1314 |
-
"epoch": 0.
|
| 1315 |
"learning_rate": 7e-05,
|
| 1316 |
-
"loss": 0.
|
| 1317 |
"step": 21800
|
| 1318 |
},
|
| 1319 |
{
|
| 1320 |
-
"epoch": 0.
|
| 1321 |
"learning_rate": 7e-05,
|
| 1322 |
-
"loss": 0.
|
| 1323 |
"step": 21900
|
| 1324 |
},
|
| 1325 |
{
|
| 1326 |
-
"epoch": 0.
|
| 1327 |
"learning_rate": 7e-05,
|
| 1328 |
-
"loss": 0.
|
| 1329 |
"step": 22000
|
| 1330 |
},
|
| 1331 |
{
|
| 1332 |
-
"epoch": 0.
|
| 1333 |
"learning_rate": 7e-05,
|
| 1334 |
-
"loss": 0.
|
| 1335 |
"step": 22100
|
| 1336 |
},
|
| 1337 |
{
|
| 1338 |
-
"epoch": 0.
|
| 1339 |
"learning_rate": 7e-05,
|
| 1340 |
-
"loss": 0.
|
| 1341 |
"step": 22200
|
| 1342 |
},
|
| 1343 |
{
|
| 1344 |
-
"epoch": 0.
|
| 1345 |
"learning_rate": 7e-05,
|
| 1346 |
-
"loss": 0.
|
| 1347 |
"step": 22300
|
| 1348 |
},
|
| 1349 |
{
|
| 1350 |
-
"epoch": 0.
|
| 1351 |
"learning_rate": 7e-05,
|
| 1352 |
-
"loss": 0.
|
| 1353 |
"step": 22400
|
| 1354 |
},
|
| 1355 |
{
|
| 1356 |
-
"epoch": 0.
|
| 1357 |
"learning_rate": 7e-05,
|
| 1358 |
-
"loss": 0.
|
| 1359 |
"step": 22500
|
| 1360 |
},
|
| 1361 |
{
|
| 1362 |
-
"epoch": 0.
|
| 1363 |
"learning_rate": 7e-05,
|
| 1364 |
-
"loss": 0.
|
| 1365 |
"step": 22600
|
| 1366 |
},
|
| 1367 |
{
|
| 1368 |
-
"epoch": 0.
|
| 1369 |
"learning_rate": 7e-05,
|
| 1370 |
-
"loss": 0.
|
| 1371 |
"step": 22700
|
| 1372 |
},
|
| 1373 |
{
|
| 1374 |
-
"epoch": 0.
|
| 1375 |
"learning_rate": 7e-05,
|
| 1376 |
-
"loss": 0.
|
| 1377 |
"step": 22800
|
| 1378 |
},
|
| 1379 |
{
|
| 1380 |
-
"epoch": 0.
|
| 1381 |
"learning_rate": 7e-05,
|
| 1382 |
-
"loss": 0.
|
| 1383 |
"step": 22900
|
| 1384 |
},
|
| 1385 |
{
|
| 1386 |
-
"epoch": 0.
|
| 1387 |
"learning_rate": 7e-05,
|
| 1388 |
-
"loss": 0.
|
| 1389 |
"step": 23000
|
| 1390 |
},
|
| 1391 |
{
|
| 1392 |
-
"epoch": 0.
|
| 1393 |
"learning_rate": 7e-05,
|
| 1394 |
-
"loss": 0.
|
| 1395 |
"step": 23100
|
| 1396 |
},
|
| 1397 |
{
|
| 1398 |
-
"epoch": 0.
|
| 1399 |
"learning_rate": 7e-05,
|
| 1400 |
-
"loss": 0.
|
| 1401 |
"step": 23200
|
| 1402 |
},
|
| 1403 |
{
|
| 1404 |
-
"epoch": 0.
|
| 1405 |
"learning_rate": 7e-05,
|
| 1406 |
-
"loss": 0.
|
| 1407 |
"step": 23300
|
| 1408 |
},
|
| 1409 |
{
|
| 1410 |
-
"epoch": 0.
|
| 1411 |
"learning_rate": 7e-05,
|
| 1412 |
-
"loss": 0.
|
| 1413 |
"step": 23400
|
| 1414 |
},
|
| 1415 |
{
|
| 1416 |
-
"epoch": 0.
|
| 1417 |
"learning_rate": 7e-05,
|
| 1418 |
-
"loss": 0.
|
| 1419 |
"step": 23500
|
| 1420 |
},
|
| 1421 |
{
|
| 1422 |
-
"epoch": 0.
|
| 1423 |
"learning_rate": 7e-05,
|
| 1424 |
-
"loss": 0.
|
| 1425 |
"step": 23600
|
| 1426 |
},
|
| 1427 |
{
|
| 1428 |
-
"epoch": 0.
|
| 1429 |
"learning_rate": 7e-05,
|
| 1430 |
-
"loss": 0.
|
| 1431 |
"step": 23700
|
| 1432 |
},
|
| 1433 |
{
|
| 1434 |
-
"epoch": 0.
|
| 1435 |
"learning_rate": 7e-05,
|
| 1436 |
-
"loss": 0.
|
| 1437 |
"step": 23800
|
| 1438 |
},
|
| 1439 |
{
|
| 1440 |
-
"epoch": 0.
|
| 1441 |
"learning_rate": 7e-05,
|
| 1442 |
-
"loss": 0.
|
| 1443 |
"step": 23900
|
| 1444 |
},
|
| 1445 |
{
|
| 1446 |
-
"epoch": 0.
|
| 1447 |
"learning_rate": 7e-05,
|
| 1448 |
-
"loss": 0.
|
| 1449 |
"step": 24000
|
| 1450 |
},
|
| 1451 |
{
|
| 1452 |
-
"epoch": 0.
|
| 1453 |
"learning_rate": 7e-05,
|
| 1454 |
-
"loss": 0.
|
| 1455 |
"step": 24100
|
| 1456 |
},
|
| 1457 |
{
|
| 1458 |
-
"epoch": 0.
|
| 1459 |
"learning_rate": 7e-05,
|
| 1460 |
-
"loss": 0.
|
| 1461 |
"step": 24200
|
| 1462 |
},
|
| 1463 |
{
|
| 1464 |
-
"epoch": 0.
|
| 1465 |
"learning_rate": 7e-05,
|
| 1466 |
-
"loss": 0.
|
| 1467 |
"step": 24300
|
| 1468 |
},
|
| 1469 |
{
|
| 1470 |
-
"epoch": 0.
|
| 1471 |
"learning_rate": 7e-05,
|
| 1472 |
-
"loss": 0.
|
| 1473 |
"step": 24400
|
| 1474 |
},
|
| 1475 |
{
|
| 1476 |
-
"epoch": 0.
|
| 1477 |
"learning_rate": 7e-05,
|
| 1478 |
-
"loss": 0.
|
| 1479 |
"step": 24500
|
| 1480 |
},
|
| 1481 |
{
|
| 1482 |
-
"epoch": 0.
|
| 1483 |
"learning_rate": 7e-05,
|
| 1484 |
-
"loss": 0.
|
| 1485 |
"step": 24600
|
| 1486 |
},
|
| 1487 |
{
|
| 1488 |
-
"epoch": 0.
|
| 1489 |
"learning_rate": 7e-05,
|
| 1490 |
-
"loss": 0.
|
| 1491 |
"step": 24700
|
| 1492 |
},
|
| 1493 |
{
|
| 1494 |
-
"epoch": 0.
|
| 1495 |
"learning_rate": 7e-05,
|
| 1496 |
-
"loss": 0.
|
| 1497 |
"step": 24800
|
| 1498 |
},
|
| 1499 |
{
|
| 1500 |
-
"epoch": 0.
|
| 1501 |
"learning_rate": 7e-05,
|
| 1502 |
-
"loss": 0.
|
| 1503 |
"step": 24900
|
| 1504 |
},
|
| 1505 |
{
|
| 1506 |
-
"epoch": 0.
|
| 1507 |
"learning_rate": 7e-05,
|
| 1508 |
-
"loss": 0.
|
| 1509 |
"step": 25000
|
| 1510 |
},
|
| 1511 |
{
|
| 1512 |
-
"epoch": 0.
|
| 1513 |
"learning_rate": 7e-05,
|
| 1514 |
-
"loss": 0.
|
| 1515 |
"step": 25100
|
| 1516 |
},
|
| 1517 |
{
|
| 1518 |
-
"epoch": 0.
|
| 1519 |
"learning_rate": 7e-05,
|
| 1520 |
-
"loss": 0.
|
| 1521 |
"step": 25200
|
| 1522 |
},
|
| 1523 |
{
|
| 1524 |
-
"epoch": 0.
|
| 1525 |
-
"learning_rate": 7e-05,
|
| 1526 |
-
"loss": 0.3616,
|
| 1527 |
-
"step": 25300
|
| 1528 |
-
},
|
| 1529 |
-
{
|
| 1530 |
-
"epoch": 0.72,
|
| 1531 |
"learning_rate": 7e-05,
|
| 1532 |
"loss": 0.3624,
|
| 1533 |
-
"step":
|
| 1534 |
-
},
|
| 1535 |
-
{
|
| 1536 |
-
"epoch": 0.73,
|
| 1537 |
-
"learning_rate": 7e-05,
|
| 1538 |
-
"loss": 0.3717,
|
| 1539 |
-
"step": 25500
|
| 1540 |
-
},
|
| 1541 |
-
{
|
| 1542 |
-
"epoch": 0.73,
|
| 1543 |
-
"learning_rate": 7e-05,
|
| 1544 |
-
"loss": 0.3606,
|
| 1545 |
-
"step": 25600
|
| 1546 |
-
},
|
| 1547 |
-
{
|
| 1548 |
-
"epoch": 0.73,
|
| 1549 |
-
"learning_rate": 7e-05,
|
| 1550 |
-
"loss": 0.3859,
|
| 1551 |
-
"step": 25700
|
| 1552 |
-
},
|
| 1553 |
-
{
|
| 1554 |
-
"epoch": 0.73,
|
| 1555 |
-
"learning_rate": 7e-05,
|
| 1556 |
-
"loss": 0.3844,
|
| 1557 |
-
"step": 25800
|
| 1558 |
-
},
|
| 1559 |
-
{
|
| 1560 |
-
"epoch": 0.74,
|
| 1561 |
-
"learning_rate": 7e-05,
|
| 1562 |
-
"loss": 0.3696,
|
| 1563 |
-
"step": 25900
|
| 1564 |
-
},
|
| 1565 |
-
{
|
| 1566 |
-
"epoch": 0.74,
|
| 1567 |
-
"learning_rate": 7e-05,
|
| 1568 |
-
"loss": 0.379,
|
| 1569 |
-
"step": 26000
|
| 1570 |
-
},
|
| 1571 |
-
{
|
| 1572 |
-
"epoch": 0.74,
|
| 1573 |
-
"learning_rate": 7e-05,
|
| 1574 |
-
"loss": 0.3798,
|
| 1575 |
-
"step": 26100
|
| 1576 |
-
},
|
| 1577 |
-
{
|
| 1578 |
-
"epoch": 0.75,
|
| 1579 |
-
"learning_rate": 7e-05,
|
| 1580 |
-
"loss": 0.3671,
|
| 1581 |
-
"step": 26200
|
| 1582 |
-
},
|
| 1583 |
-
{
|
| 1584 |
-
"epoch": 0.75,
|
| 1585 |
-
"learning_rate": 7e-05,
|
| 1586 |
-
"loss": 0.3764,
|
| 1587 |
-
"step": 26300
|
| 1588 |
-
},
|
| 1589 |
-
{
|
| 1590 |
-
"epoch": 0.75,
|
| 1591 |
-
"learning_rate": 7e-05,
|
| 1592 |
-
"loss": 0.3651,
|
| 1593 |
-
"step": 26400
|
| 1594 |
-
},
|
| 1595 |
-
{
|
| 1596 |
-
"epoch": 0.75,
|
| 1597 |
-
"learning_rate": 7e-05,
|
| 1598 |
-
"loss": 0.3756,
|
| 1599 |
-
"step": 26500
|
| 1600 |
-
},
|
| 1601 |
-
{
|
| 1602 |
-
"epoch": 0.76,
|
| 1603 |
-
"learning_rate": 7e-05,
|
| 1604 |
-
"loss": 0.3603,
|
| 1605 |
-
"step": 26600
|
| 1606 |
-
},
|
| 1607 |
-
{
|
| 1608 |
-
"epoch": 0.76,
|
| 1609 |
-
"learning_rate": 7e-05,
|
| 1610 |
-
"loss": 0.355,
|
| 1611 |
-
"step": 26700
|
| 1612 |
-
},
|
| 1613 |
-
{
|
| 1614 |
-
"epoch": 0.76,
|
| 1615 |
-
"learning_rate": 7e-05,
|
| 1616 |
-
"loss": 0.3756,
|
| 1617 |
-
"step": 26800
|
| 1618 |
-
},
|
| 1619 |
-
{
|
| 1620 |
-
"epoch": 0.77,
|
| 1621 |
-
"learning_rate": 7e-05,
|
| 1622 |
-
"loss": 0.3784,
|
| 1623 |
-
"step": 26900
|
| 1624 |
-
},
|
| 1625 |
-
{
|
| 1626 |
-
"epoch": 0.77,
|
| 1627 |
-
"learning_rate": 7e-05,
|
| 1628 |
-
"loss": 0.3592,
|
| 1629 |
-
"step": 27000
|
| 1630 |
-
},
|
| 1631 |
-
{
|
| 1632 |
-
"epoch": 0.77,
|
| 1633 |
-
"learning_rate": 7e-05,
|
| 1634 |
-
"loss": 0.353,
|
| 1635 |
-
"step": 27100
|
| 1636 |
-
},
|
| 1637 |
-
{
|
| 1638 |
-
"epoch": 0.77,
|
| 1639 |
-
"learning_rate": 7e-05,
|
| 1640 |
-
"loss": 0.3617,
|
| 1641 |
-
"step": 27200
|
| 1642 |
-
},
|
| 1643 |
-
{
|
| 1644 |
-
"epoch": 0.78,
|
| 1645 |
-
"learning_rate": 7e-05,
|
| 1646 |
-
"loss": 0.3548,
|
| 1647 |
-
"step": 27300
|
| 1648 |
-
},
|
| 1649 |
-
{
|
| 1650 |
-
"epoch": 0.78,
|
| 1651 |
-
"learning_rate": 7e-05,
|
| 1652 |
-
"loss": 0.3867,
|
| 1653 |
-
"step": 27400
|
| 1654 |
-
},
|
| 1655 |
-
{
|
| 1656 |
-
"epoch": 0.78,
|
| 1657 |
-
"learning_rate": 7e-05,
|
| 1658 |
-
"loss": 0.3676,
|
| 1659 |
-
"step": 27500
|
| 1660 |
-
},
|
| 1661 |
-
{
|
| 1662 |
-
"epoch": 0.79,
|
| 1663 |
-
"learning_rate": 7e-05,
|
| 1664 |
-
"loss": 0.3761,
|
| 1665 |
-
"step": 27600
|
| 1666 |
-
},
|
| 1667 |
-
{
|
| 1668 |
-
"epoch": 0.79,
|
| 1669 |
-
"learning_rate": 7e-05,
|
| 1670 |
-
"loss": 0.3783,
|
| 1671 |
-
"step": 27700
|
| 1672 |
-
},
|
| 1673 |
-
{
|
| 1674 |
-
"epoch": 0.79,
|
| 1675 |
-
"learning_rate": 7e-05,
|
| 1676 |
-
"loss": 0.3749,
|
| 1677 |
-
"step": 27800
|
| 1678 |
-
},
|
| 1679 |
-
{
|
| 1680 |
-
"epoch": 0.79,
|
| 1681 |
-
"learning_rate": 7e-05,
|
| 1682 |
-
"loss": 0.3588,
|
| 1683 |
-
"step": 27900
|
| 1684 |
-
},
|
| 1685 |
-
{
|
| 1686 |
-
"epoch": 0.8,
|
| 1687 |
-
"learning_rate": 7e-05,
|
| 1688 |
-
"loss": 0.3794,
|
| 1689 |
-
"step": 28000
|
| 1690 |
-
},
|
| 1691 |
-
{
|
| 1692 |
-
"epoch": 0.8,
|
| 1693 |
-
"learning_rate": 7e-05,
|
| 1694 |
-
"loss": 0.3742,
|
| 1695 |
-
"step": 28100
|
| 1696 |
-
},
|
| 1697 |
-
{
|
| 1698 |
-
"epoch": 0.8,
|
| 1699 |
-
"learning_rate": 7e-05,
|
| 1700 |
-
"loss": 0.3627,
|
| 1701 |
-
"step": 28200
|
| 1702 |
-
},
|
| 1703 |
-
{
|
| 1704 |
-
"epoch": 0.81,
|
| 1705 |
-
"learning_rate": 7e-05,
|
| 1706 |
-
"loss": 0.3796,
|
| 1707 |
-
"step": 28300
|
| 1708 |
-
},
|
| 1709 |
-
{
|
| 1710 |
-
"epoch": 0.81,
|
| 1711 |
-
"learning_rate": 7e-05,
|
| 1712 |
-
"loss": 0.3591,
|
| 1713 |
-
"step": 28400
|
| 1714 |
-
},
|
| 1715 |
-
{
|
| 1716 |
-
"epoch": 0.81,
|
| 1717 |
-
"learning_rate": 7e-05,
|
| 1718 |
-
"loss": 0.3696,
|
| 1719 |
-
"step": 28500
|
| 1720 |
-
},
|
| 1721 |
-
{
|
| 1722 |
-
"epoch": 0.81,
|
| 1723 |
-
"learning_rate": 7e-05,
|
| 1724 |
-
"loss": 0.3709,
|
| 1725 |
-
"step": 28600
|
| 1726 |
},
|
| 1727 |
{
|
| 1728 |
"epoch": 0.82,
|
| 1729 |
"learning_rate": 7e-05,
|
| 1730 |
-
"loss": 0.
|
| 1731 |
-
"step":
|
| 1732 |
},
|
| 1733 |
{
|
| 1734 |
"epoch": 0.82,
|
| 1735 |
"learning_rate": 7e-05,
|
| 1736 |
-
"loss": 0.
|
| 1737 |
-
"step":
|
| 1738 |
},
|
| 1739 |
{
|
| 1740 |
"epoch": 0.82,
|
| 1741 |
"learning_rate": 7e-05,
|
| 1742 |
-
"loss": 0.
|
| 1743 |
-
"step":
|
| 1744 |
-
},
|
| 1745 |
-
{
|
| 1746 |
-
"epoch": 0.83,
|
| 1747 |
-
"learning_rate": 7e-05,
|
| 1748 |
-
"loss": 0.3475,
|
| 1749 |
-
"step": 29000
|
| 1750 |
},
|
| 1751 |
{
|
| 1752 |
"epoch": 0.83,
|
| 1753 |
"learning_rate": 7e-05,
|
| 1754 |
-
"loss": 0.
|
| 1755 |
-
"step":
|
| 1756 |
},
|
| 1757 |
{
|
| 1758 |
"epoch": 0.83,
|
| 1759 |
"learning_rate": 7e-05,
|
| 1760 |
-
"loss": 0.
|
| 1761 |
-
"step":
|
| 1762 |
},
|
| 1763 |
{
|
| 1764 |
"epoch": 0.83,
|
| 1765 |
"learning_rate": 7e-05,
|
| 1766 |
-
"loss": 0.
|
| 1767 |
-
"step":
|
| 1768 |
},
|
| 1769 |
{
|
| 1770 |
"epoch": 0.84,
|
| 1771 |
"learning_rate": 7e-05,
|
| 1772 |
-
"loss": 0.
|
| 1773 |
-
"step":
|
| 1774 |
},
|
| 1775 |
{
|
| 1776 |
"epoch": 0.84,
|
| 1777 |
"learning_rate": 7e-05,
|
| 1778 |
-
"loss": 0.
|
| 1779 |
-
"step":
|
| 1780 |
},
|
| 1781 |
{
|
| 1782 |
"epoch": 0.84,
|
| 1783 |
"learning_rate": 7e-05,
|
| 1784 |
-
"loss": 0.
|
| 1785 |
-
"step":
|
| 1786 |
-
},
|
| 1787 |
-
{
|
| 1788 |
-
"epoch": 0.85,
|
| 1789 |
-
"learning_rate": 7e-05,
|
| 1790 |
-
"loss": 0.3689,
|
| 1791 |
-
"step": 29700
|
| 1792 |
},
|
| 1793 |
{
|
| 1794 |
"epoch": 0.85,
|
| 1795 |
"learning_rate": 7e-05,
|
| 1796 |
-
"loss": 0.
|
| 1797 |
-
"step":
|
| 1798 |
},
|
| 1799 |
{
|
| 1800 |
"epoch": 0.85,
|
| 1801 |
"learning_rate": 7e-05,
|
| 1802 |
-
"loss": 0.
|
| 1803 |
-
"step":
|
| 1804 |
},
|
| 1805 |
{
|
| 1806 |
"epoch": 0.85,
|
| 1807 |
"learning_rate": 7e-05,
|
| 1808 |
-
"loss": 0.
|
| 1809 |
-
"step":
|
| 1810 |
},
|
| 1811 |
{
|
| 1812 |
"epoch": 0.86,
|
| 1813 |
"learning_rate": 7e-05,
|
| 1814 |
-
"loss": 0.
|
| 1815 |
-
"step":
|
| 1816 |
},
|
| 1817 |
{
|
| 1818 |
"epoch": 0.86,
|
| 1819 |
"learning_rate": 7e-05,
|
| 1820 |
-
"loss": 0.
|
| 1821 |
-
"step":
|
| 1822 |
},
|
| 1823 |
{
|
| 1824 |
"epoch": 0.86,
|
| 1825 |
"learning_rate": 7e-05,
|
| 1826 |
-
"loss": 0.
|
| 1827 |
-
"step":
|
| 1828 |
-
},
|
| 1829 |
-
{
|
| 1830 |
-
"epoch": 0.87,
|
| 1831 |
-
"learning_rate": 7e-05,
|
| 1832 |
-
"loss": 0.3671,
|
| 1833 |
-
"step": 30400
|
| 1834 |
},
|
| 1835 |
{
|
| 1836 |
"epoch": 0.87,
|
| 1837 |
"learning_rate": 7e-05,
|
| 1838 |
-
"loss": 0.
|
| 1839 |
-
"step":
|
| 1840 |
},
|
| 1841 |
{
|
| 1842 |
"epoch": 0.87,
|
| 1843 |
"learning_rate": 7e-05,
|
| 1844 |
-
"loss": 0.
|
| 1845 |
-
"step":
|
| 1846 |
},
|
| 1847 |
{
|
| 1848 |
"epoch": 0.87,
|
| 1849 |
"learning_rate": 7e-05,
|
| 1850 |
-
"loss": 0.
|
| 1851 |
-
"step":
|
| 1852 |
},
|
| 1853 |
{
|
| 1854 |
"epoch": 0.88,
|
| 1855 |
"learning_rate": 7e-05,
|
| 1856 |
-
"loss": 0.
|
| 1857 |
-
"step":
|
| 1858 |
},
|
| 1859 |
{
|
| 1860 |
"epoch": 0.88,
|
| 1861 |
"learning_rate": 7e-05,
|
| 1862 |
-
"loss": 0.
|
| 1863 |
-
"step":
|
| 1864 |
},
|
| 1865 |
{
|
| 1866 |
"epoch": 0.88,
|
| 1867 |
"learning_rate": 7e-05,
|
| 1868 |
-
"loss": 0.
|
| 1869 |
-
"step":
|
| 1870 |
-
},
|
| 1871 |
-
{
|
| 1872 |
-
"epoch": 0.89,
|
| 1873 |
-
"learning_rate": 7e-05,
|
| 1874 |
-
"loss": 0.3672,
|
| 1875 |
-
"step": 31100
|
| 1876 |
},
|
| 1877 |
{
|
| 1878 |
"epoch": 0.89,
|
| 1879 |
"learning_rate": 7e-05,
|
| 1880 |
-
"loss": 0.
|
| 1881 |
-
"step":
|
| 1882 |
},
|
| 1883 |
{
|
| 1884 |
"epoch": 0.89,
|
| 1885 |
"learning_rate": 7e-05,
|
| 1886 |
-
"loss": 0.
|
| 1887 |
-
"step":
|
| 1888 |
},
|
| 1889 |
{
|
| 1890 |
"epoch": 0.89,
|
| 1891 |
"learning_rate": 7e-05,
|
| 1892 |
-
"loss": 0.
|
| 1893 |
-
"step":
|
| 1894 |
},
|
| 1895 |
{
|
| 1896 |
"epoch": 0.9,
|
| 1897 |
"learning_rate": 7e-05,
|
| 1898 |
-
"loss": 0.
|
| 1899 |
-
"step":
|
| 1900 |
},
|
| 1901 |
{
|
| 1902 |
"epoch": 0.9,
|
| 1903 |
"learning_rate": 7e-05,
|
| 1904 |
-
"loss": 0.
|
| 1905 |
-
"step":
|
| 1906 |
},
|
| 1907 |
{
|
| 1908 |
"epoch": 0.9,
|
| 1909 |
"learning_rate": 7e-05,
|
| 1910 |
-
"loss": 0.
|
| 1911 |
-
"step":
|
| 1912 |
},
|
| 1913 |
{
|
| 1914 |
"epoch": 0.91,
|
| 1915 |
"learning_rate": 7e-05,
|
| 1916 |
-
"loss": 0.
|
| 1917 |
-
"step":
|
| 1918 |
},
|
| 1919 |
{
|
| 1920 |
"epoch": 0.91,
|
| 1921 |
"learning_rate": 7e-05,
|
| 1922 |
-
"loss": 0.
|
| 1923 |
-
"step":
|
| 1924 |
},
|
| 1925 |
{
|
| 1926 |
"epoch": 0.91,
|
| 1927 |
"learning_rate": 7e-05,
|
| 1928 |
-
"loss": 0.
|
| 1929 |
-
"step":
|
| 1930 |
},
|
| 1931 |
{
|
| 1932 |
"epoch": 0.91,
|
| 1933 |
"learning_rate": 7e-05,
|
| 1934 |
-
"loss": 0.
|
| 1935 |
-
"step":
|
| 1936 |
},
|
| 1937 |
{
|
| 1938 |
"epoch": 0.92,
|
| 1939 |
"learning_rate": 7e-05,
|
| 1940 |
-
"loss": 0.
|
| 1941 |
-
"step":
|
| 1942 |
},
|
| 1943 |
{
|
| 1944 |
"epoch": 0.92,
|
| 1945 |
"learning_rate": 7e-05,
|
| 1946 |
-
"loss": 0.
|
| 1947 |
-
"step":
|
| 1948 |
},
|
| 1949 |
{
|
| 1950 |
"epoch": 0.92,
|
| 1951 |
"learning_rate": 7e-05,
|
| 1952 |
-
"loss": 0.
|
| 1953 |
-
"step":
|
| 1954 |
-
},
|
| 1955 |
-
{
|
| 1956 |
-
"epoch": 0.93,
|
| 1957 |
-
"learning_rate": 7e-05,
|
| 1958 |
-
"loss": 0.3603,
|
| 1959 |
-
"step": 32500
|
| 1960 |
},
|
| 1961 |
{
|
| 1962 |
"epoch": 0.93,
|
| 1963 |
"learning_rate": 7e-05,
|
| 1964 |
-
"loss": 0.
|
| 1965 |
-
"step":
|
| 1966 |
},
|
| 1967 |
{
|
| 1968 |
"epoch": 0.93,
|
| 1969 |
"learning_rate": 7e-05,
|
| 1970 |
-
"loss": 0.
|
| 1971 |
-
"step":
|
| 1972 |
},
|
| 1973 |
{
|
| 1974 |
"epoch": 0.93,
|
| 1975 |
"learning_rate": 7e-05,
|
| 1976 |
-
"loss": 0.
|
| 1977 |
-
"step":
|
| 1978 |
-
},
|
| 1979 |
-
{
|
| 1980 |
-
"epoch": 0.94,
|
| 1981 |
-
"learning_rate": 7e-05,
|
| 1982 |
-
"loss": 0.3643,
|
| 1983 |
-
"step": 32900
|
| 1984 |
},
|
| 1985 |
{
|
| 1986 |
"epoch": 0.94,
|
| 1987 |
"learning_rate": 7e-05,
|
| 1988 |
-
"loss": 0.
|
| 1989 |
-
"step":
|
| 1990 |
},
|
| 1991 |
{
|
| 1992 |
"epoch": 0.94,
|
| 1993 |
"learning_rate": 7e-05,
|
| 1994 |
-
"loss": 0.
|
| 1995 |
-
"step":
|
| 1996 |
},
|
| 1997 |
{
|
| 1998 |
"epoch": 0.94,
|
| 1999 |
"learning_rate": 7e-05,
|
| 2000 |
-
"loss": 0.
|
| 2001 |
-
"step":
|
| 2002 |
},
|
| 2003 |
{
|
| 2004 |
"epoch": 0.95,
|
| 2005 |
"learning_rate": 7e-05,
|
| 2006 |
-
"loss": 0.
|
| 2007 |
-
"step":
|
| 2008 |
},
|
| 2009 |
{
|
| 2010 |
"epoch": 0.95,
|
| 2011 |
"learning_rate": 7e-05,
|
| 2012 |
-
"loss": 0.
|
| 2013 |
-
"step":
|
| 2014 |
},
|
| 2015 |
{
|
| 2016 |
"epoch": 0.95,
|
| 2017 |
"learning_rate": 7e-05,
|
| 2018 |
-
"loss": 0.
|
| 2019 |
-
"step":
|
| 2020 |
-
},
|
| 2021 |
-
{
|
| 2022 |
-
"epoch": 0.96,
|
| 2023 |
-
"learning_rate": 7e-05,
|
| 2024 |
-
"loss": 0.3704,
|
| 2025 |
-
"step": 33600
|
| 2026 |
},
|
| 2027 |
{
|
| 2028 |
"epoch": 0.96,
|
| 2029 |
"learning_rate": 7e-05,
|
| 2030 |
-
"loss": 0.
|
| 2031 |
-
"step":
|
| 2032 |
},
|
| 2033 |
{
|
| 2034 |
"epoch": 0.96,
|
| 2035 |
"learning_rate": 7e-05,
|
| 2036 |
-
"loss": 0.
|
| 2037 |
-
"step":
|
| 2038 |
},
|
| 2039 |
{
|
| 2040 |
"epoch": 0.96,
|
| 2041 |
"learning_rate": 7e-05,
|
| 2042 |
-
"loss": 0.
|
| 2043 |
-
"step":
|
| 2044 |
},
|
| 2045 |
{
|
| 2046 |
"epoch": 0.97,
|
| 2047 |
"learning_rate": 7e-05,
|
| 2048 |
-
"loss": 0.
|
| 2049 |
-
"step":
|
| 2050 |
},
|
| 2051 |
{
|
| 2052 |
"epoch": 0.97,
|
| 2053 |
"learning_rate": 7e-05,
|
| 2054 |
-
"loss": 0.
|
| 2055 |
-
"step":
|
| 2056 |
},
|
| 2057 |
{
|
| 2058 |
"epoch": 0.97,
|
| 2059 |
"learning_rate": 7e-05,
|
| 2060 |
-
"loss": 0.
|
| 2061 |
-
"step":
|
| 2062 |
-
},
|
| 2063 |
-
{
|
| 2064 |
-
"epoch": 0.98,
|
| 2065 |
-
"learning_rate": 7e-05,
|
| 2066 |
-
"loss": 0.3795,
|
| 2067 |
-
"step": 34300
|
| 2068 |
},
|
| 2069 |
{
|
| 2070 |
"epoch": 0.98,
|
| 2071 |
"learning_rate": 7e-05,
|
| 2072 |
-
"loss": 0.
|
| 2073 |
-
"step":
|
| 2074 |
},
|
| 2075 |
{
|
| 2076 |
"epoch": 0.98,
|
| 2077 |
"learning_rate": 7e-05,
|
| 2078 |
-
"loss": 0.
|
| 2079 |
-
"step":
|
| 2080 |
},
|
| 2081 |
{
|
| 2082 |
"epoch": 0.98,
|
| 2083 |
"learning_rate": 7e-05,
|
| 2084 |
-
"loss": 0.
|
| 2085 |
-
"step":
|
| 2086 |
},
|
| 2087 |
{
|
| 2088 |
"epoch": 0.99,
|
| 2089 |
"learning_rate": 7e-05,
|
| 2090 |
-
"loss": 0.
|
| 2091 |
-
"step":
|
| 2092 |
},
|
| 2093 |
{
|
| 2094 |
"epoch": 0.99,
|
| 2095 |
"learning_rate": 7e-05,
|
| 2096 |
-
"loss": 0.
|
| 2097 |
-
"step":
|
| 2098 |
},
|
| 2099 |
{
|
| 2100 |
"epoch": 0.99,
|
| 2101 |
"learning_rate": 7e-05,
|
| 2102 |
-
"loss": 0.
|
| 2103 |
-
"step":
|
| 2104 |
},
|
| 2105 |
{
|
| 2106 |
"epoch": 1.0,
|
| 2107 |
"learning_rate": 7e-05,
|
| 2108 |
-
"loss": 0.
|
| 2109 |
-
"step":
|
| 2110 |
},
|
| 2111 |
{
|
| 2112 |
"epoch": 1.0,
|
| 2113 |
"learning_rate": 7e-05,
|
| 2114 |
-
"loss": 0.
|
| 2115 |
-
"step":
|
| 2116 |
},
|
| 2117 |
{
|
| 2118 |
"epoch": 1.0,
|
| 2119 |
-
"eval_gen_len": 24.
|
| 2120 |
-
"eval_loss": 0.
|
| 2121 |
-
"eval_rouge1": 86.
|
| 2122 |
-
"eval_rouge2": 75.
|
| 2123 |
-
"eval_rougeL": 85.
|
| 2124 |
-
"eval_rougeLsum": 85.
|
| 2125 |
-
"eval_runtime":
|
| 2126 |
-
"eval_samples_per_second": 4.
|
| 2127 |
-
"eval_steps_per_second": 1.
|
| 2128 |
-
"step":
|
| 2129 |
},
|
| 2130 |
{
|
| 2131 |
"epoch": 1.0,
|
| 2132 |
-
"step":
|
| 2133 |
-
"total_flos":
|
| 2134 |
-
"train_loss": 0.
|
| 2135 |
-
"train_runtime":
|
| 2136 |
-
"train_samples_per_second": 25.
|
| 2137 |
-
"train_steps_per_second": 0.
|
| 2138 |
}
|
| 2139 |
],
|
| 2140 |
"logging_steps": 100,
|
| 2141 |
-
"max_steps":
|
| 2142 |
"num_input_tokens_seen": 0,
|
| 2143 |
"num_train_epochs": 1,
|
| 2144 |
"save_steps": 2500,
|
| 2145 |
-
"total_flos":
|
| 2146 |
"trial_name": null,
|
| 2147 |
"trial_params": null
|
| 2148 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.0,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 31042,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 11 |
{
|
| 12 |
"epoch": 0.0,
|
| 13 |
"learning_rate": 7e-05,
|
| 14 |
+
"loss": 0.3828,
|
| 15 |
"step": 100
|
| 16 |
},
|
| 17 |
{
|
| 18 |
"epoch": 0.01,
|
| 19 |
"learning_rate": 7e-05,
|
| 20 |
+
"loss": 0.3667,
|
| 21 |
"step": 200
|
| 22 |
},
|
| 23 |
{
|
| 24 |
"epoch": 0.01,
|
| 25 |
"learning_rate": 7e-05,
|
| 26 |
+
"loss": 0.3574,
|
| 27 |
"step": 300
|
| 28 |
},
|
| 29 |
{
|
| 30 |
"epoch": 0.01,
|
| 31 |
"learning_rate": 7e-05,
|
| 32 |
+
"loss": 0.3835,
|
| 33 |
"step": 400
|
| 34 |
},
|
| 35 |
{
|
| 36 |
+
"epoch": 0.02,
|
| 37 |
"learning_rate": 7e-05,
|
| 38 |
+
"loss": 0.3657,
|
| 39 |
"step": 500
|
| 40 |
},
|
| 41 |
{
|
| 42 |
"epoch": 0.02,
|
| 43 |
"learning_rate": 7e-05,
|
| 44 |
+
"loss": 0.3542,
|
| 45 |
"step": 600
|
| 46 |
},
|
| 47 |
{
|
| 48 |
"epoch": 0.02,
|
| 49 |
"learning_rate": 7e-05,
|
| 50 |
+
"loss": 0.3592,
|
| 51 |
"step": 700
|
| 52 |
},
|
| 53 |
{
|
| 54 |
+
"epoch": 0.03,
|
| 55 |
"learning_rate": 7e-05,
|
| 56 |
+
"loss": 0.3429,
|
| 57 |
"step": 800
|
| 58 |
},
|
| 59 |
{
|
| 60 |
"epoch": 0.03,
|
| 61 |
"learning_rate": 7e-05,
|
| 62 |
+
"loss": 0.3618,
|
| 63 |
"step": 900
|
| 64 |
},
|
| 65 |
{
|
| 66 |
"epoch": 0.03,
|
| 67 |
"learning_rate": 7e-05,
|
| 68 |
+
"loss": 0.3658,
|
| 69 |
"step": 1000
|
| 70 |
},
|
| 71 |
{
|
| 72 |
+
"epoch": 0.04,
|
| 73 |
"learning_rate": 7e-05,
|
| 74 |
+
"loss": 0.3608,
|
| 75 |
"step": 1100
|
| 76 |
},
|
| 77 |
{
|
| 78 |
+
"epoch": 0.04,
|
| 79 |
"learning_rate": 7e-05,
|
| 80 |
+
"loss": 0.3488,
|
| 81 |
"step": 1200
|
| 82 |
},
|
| 83 |
{
|
| 84 |
"epoch": 0.04,
|
| 85 |
"learning_rate": 7e-05,
|
| 86 |
+
"loss": 0.3714,
|
| 87 |
"step": 1300
|
| 88 |
},
|
| 89 |
{
|
| 90 |
+
"epoch": 0.05,
|
| 91 |
"learning_rate": 7e-05,
|
| 92 |
+
"loss": 0.3791,
|
| 93 |
"step": 1400
|
| 94 |
},
|
| 95 |
{
|
| 96 |
+
"epoch": 0.05,
|
| 97 |
"learning_rate": 7e-05,
|
| 98 |
+
"loss": 0.3425,
|
| 99 |
"step": 1500
|
| 100 |
},
|
| 101 |
{
|
| 102 |
"epoch": 0.05,
|
| 103 |
"learning_rate": 7e-05,
|
| 104 |
+
"loss": 0.3717,
|
| 105 |
"step": 1600
|
| 106 |
},
|
| 107 |
{
|
| 108 |
"epoch": 0.05,
|
| 109 |
"learning_rate": 7e-05,
|
| 110 |
+
"loss": 0.3593,
|
| 111 |
"step": 1700
|
| 112 |
},
|
| 113 |
{
|
| 114 |
+
"epoch": 0.06,
|
| 115 |
"learning_rate": 7e-05,
|
| 116 |
+
"loss": 0.3619,
|
| 117 |
"step": 1800
|
| 118 |
},
|
| 119 |
{
|
| 120 |
+
"epoch": 0.06,
|
| 121 |
"learning_rate": 7e-05,
|
| 122 |
+
"loss": 0.3634,
|
| 123 |
"step": 1900
|
| 124 |
},
|
| 125 |
{
|
| 126 |
"epoch": 0.06,
|
| 127 |
"learning_rate": 7e-05,
|
| 128 |
+
"loss": 0.3694,
|
| 129 |
"step": 2000
|
| 130 |
},
|
| 131 |
{
|
| 132 |
+
"epoch": 0.07,
|
| 133 |
"learning_rate": 7e-05,
|
| 134 |
+
"loss": 0.3664,
|
| 135 |
"step": 2100
|
| 136 |
},
|
| 137 |
{
|
| 138 |
+
"epoch": 0.07,
|
| 139 |
"learning_rate": 7e-05,
|
| 140 |
+
"loss": 0.3553,
|
| 141 |
"step": 2200
|
| 142 |
},
|
| 143 |
{
|
| 144 |
"epoch": 0.07,
|
| 145 |
"learning_rate": 7e-05,
|
| 146 |
+
"loss": 0.372,
|
| 147 |
"step": 2300
|
| 148 |
},
|
| 149 |
{
|
| 150 |
+
"epoch": 0.08,
|
| 151 |
"learning_rate": 7e-05,
|
| 152 |
+
"loss": 0.3901,
|
| 153 |
"step": 2400
|
| 154 |
},
|
| 155 |
{
|
| 156 |
+
"epoch": 0.08,
|
| 157 |
"learning_rate": 7e-05,
|
| 158 |
+
"loss": 0.3713,
|
| 159 |
"step": 2500
|
| 160 |
},
|
| 161 |
{
|
| 162 |
+
"epoch": 0.08,
|
| 163 |
"learning_rate": 7e-05,
|
| 164 |
+
"loss": 0.3659,
|
| 165 |
"step": 2600
|
| 166 |
},
|
| 167 |
{
|
| 168 |
+
"epoch": 0.09,
|
| 169 |
"learning_rate": 7e-05,
|
| 170 |
+
"loss": 0.3568,
|
| 171 |
"step": 2700
|
| 172 |
},
|
| 173 |
{
|
| 174 |
+
"epoch": 0.09,
|
| 175 |
"learning_rate": 7e-05,
|
| 176 |
+
"loss": 0.3533,
|
| 177 |
"step": 2800
|
| 178 |
},
|
| 179 |
{
|
| 180 |
+
"epoch": 0.09,
|
| 181 |
"learning_rate": 7e-05,
|
| 182 |
+
"loss": 0.365,
|
| 183 |
"step": 2900
|
| 184 |
},
|
| 185 |
{
|
| 186 |
+
"epoch": 0.1,
|
| 187 |
"learning_rate": 7e-05,
|
| 188 |
+
"loss": 0.3647,
|
| 189 |
"step": 3000
|
| 190 |
},
|
| 191 |
{
|
| 192 |
+
"epoch": 0.1,
|
| 193 |
"learning_rate": 7e-05,
|
| 194 |
+
"loss": 0.3671,
|
| 195 |
"step": 3100
|
| 196 |
},
|
| 197 |
{
|
| 198 |
+
"epoch": 0.1,
|
| 199 |
"learning_rate": 7e-05,
|
| 200 |
+
"loss": 0.3423,
|
| 201 |
"step": 3200
|
| 202 |
},
|
| 203 |
{
|
| 204 |
+
"epoch": 0.11,
|
| 205 |
"learning_rate": 7e-05,
|
| 206 |
+
"loss": 0.3591,
|
| 207 |
"step": 3300
|
| 208 |
},
|
| 209 |
{
|
| 210 |
+
"epoch": 0.11,
|
| 211 |
"learning_rate": 7e-05,
|
| 212 |
+
"loss": 0.3715,
|
| 213 |
"step": 3400
|
| 214 |
},
|
| 215 |
{
|
| 216 |
+
"epoch": 0.11,
|
| 217 |
"learning_rate": 7e-05,
|
| 218 |
+
"loss": 0.3561,
|
| 219 |
"step": 3500
|
| 220 |
},
|
| 221 |
{
|
| 222 |
+
"epoch": 0.12,
|
| 223 |
"learning_rate": 7e-05,
|
| 224 |
+
"loss": 0.369,
|
| 225 |
"step": 3600
|
| 226 |
},
|
| 227 |
{
|
| 228 |
+
"epoch": 0.12,
|
| 229 |
"learning_rate": 7e-05,
|
| 230 |
+
"loss": 0.3725,
|
| 231 |
"step": 3700
|
| 232 |
},
|
| 233 |
{
|
| 234 |
+
"epoch": 0.12,
|
| 235 |
"learning_rate": 7e-05,
|
| 236 |
+
"loss": 0.3563,
|
| 237 |
"step": 3800
|
| 238 |
},
|
| 239 |
{
|
| 240 |
+
"epoch": 0.13,
|
| 241 |
"learning_rate": 7e-05,
|
| 242 |
+
"loss": 0.3738,
|
| 243 |
"step": 3900
|
| 244 |
},
|
| 245 |
{
|
| 246 |
+
"epoch": 0.13,
|
| 247 |
"learning_rate": 7e-05,
|
| 248 |
+
"loss": 0.355,
|
| 249 |
"step": 4000
|
| 250 |
},
|
| 251 |
{
|
| 252 |
+
"epoch": 0.13,
|
| 253 |
"learning_rate": 7e-05,
|
| 254 |
+
"loss": 0.3486,
|
| 255 |
"step": 4100
|
| 256 |
},
|
| 257 |
{
|
| 258 |
+
"epoch": 0.14,
|
| 259 |
"learning_rate": 7e-05,
|
| 260 |
+
"loss": 0.3627,
|
| 261 |
"step": 4200
|
| 262 |
},
|
| 263 |
{
|
| 264 |
+
"epoch": 0.14,
|
| 265 |
"learning_rate": 7e-05,
|
| 266 |
+
"loss": 0.3324,
|
| 267 |
"step": 4300
|
| 268 |
},
|
| 269 |
{
|
| 270 |
+
"epoch": 0.14,
|
| 271 |
"learning_rate": 7e-05,
|
| 272 |
+
"loss": 0.3921,
|
| 273 |
"step": 4400
|
| 274 |
},
|
| 275 |
{
|
| 276 |
+
"epoch": 0.14,
|
| 277 |
"learning_rate": 7e-05,
|
| 278 |
+
"loss": 0.3509,
|
| 279 |
"step": 4500
|
| 280 |
},
|
| 281 |
{
|
| 282 |
+
"epoch": 0.15,
|
| 283 |
"learning_rate": 7e-05,
|
| 284 |
+
"loss": 0.3671,
|
| 285 |
"step": 4600
|
| 286 |
},
|
| 287 |
{
|
| 288 |
+
"epoch": 0.15,
|
| 289 |
"learning_rate": 7e-05,
|
| 290 |
+
"loss": 0.363,
|
| 291 |
"step": 4700
|
| 292 |
},
|
| 293 |
{
|
| 294 |
+
"epoch": 0.15,
|
| 295 |
"learning_rate": 7e-05,
|
| 296 |
+
"loss": 0.3565,
|
| 297 |
"step": 4800
|
| 298 |
},
|
| 299 |
{
|
| 300 |
+
"epoch": 0.16,
|
| 301 |
"learning_rate": 7e-05,
|
| 302 |
+
"loss": 0.3709,
|
| 303 |
"step": 4900
|
| 304 |
},
|
| 305 |
{
|
| 306 |
+
"epoch": 0.16,
|
| 307 |
"learning_rate": 7e-05,
|
| 308 |
+
"loss": 0.3825,
|
| 309 |
"step": 5000
|
| 310 |
},
|
| 311 |
{
|
| 312 |
+
"epoch": 0.16,
|
| 313 |
"learning_rate": 7e-05,
|
| 314 |
+
"loss": 0.3565,
|
| 315 |
"step": 5100
|
| 316 |
},
|
| 317 |
{
|
| 318 |
+
"epoch": 0.17,
|
| 319 |
"learning_rate": 7e-05,
|
| 320 |
+
"loss": 0.3665,
|
| 321 |
"step": 5200
|
| 322 |
},
|
| 323 |
{
|
| 324 |
+
"epoch": 0.17,
|
| 325 |
"learning_rate": 7e-05,
|
| 326 |
+
"loss": 0.364,
|
| 327 |
"step": 5300
|
| 328 |
},
|
| 329 |
{
|
| 330 |
+
"epoch": 0.17,
|
| 331 |
"learning_rate": 7e-05,
|
| 332 |
+
"loss": 0.3676,
|
| 333 |
"step": 5400
|
| 334 |
},
|
| 335 |
{
|
| 336 |
+
"epoch": 0.18,
|
| 337 |
"learning_rate": 7e-05,
|
| 338 |
+
"loss": 0.3629,
|
| 339 |
"step": 5500
|
| 340 |
},
|
| 341 |
{
|
| 342 |
+
"epoch": 0.18,
|
| 343 |
"learning_rate": 7e-05,
|
| 344 |
+
"loss": 0.3677,
|
| 345 |
"step": 5600
|
| 346 |
},
|
| 347 |
{
|
| 348 |
+
"epoch": 0.18,
|
| 349 |
"learning_rate": 7e-05,
|
| 350 |
+
"loss": 0.3749,
|
| 351 |
"step": 5700
|
| 352 |
},
|
| 353 |
{
|
| 354 |
+
"epoch": 0.19,
|
| 355 |
"learning_rate": 7e-05,
|
| 356 |
+
"loss": 0.3595,
|
| 357 |
"step": 5800
|
| 358 |
},
|
| 359 |
{
|
| 360 |
+
"epoch": 0.19,
|
| 361 |
"learning_rate": 7e-05,
|
| 362 |
+
"loss": 0.3583,
|
| 363 |
"step": 5900
|
| 364 |
},
|
| 365 |
{
|
| 366 |
+
"epoch": 0.19,
|
| 367 |
"learning_rate": 7e-05,
|
| 368 |
+
"loss": 0.3677,
|
| 369 |
"step": 6000
|
| 370 |
},
|
| 371 |
{
|
| 372 |
+
"epoch": 0.2,
|
| 373 |
"learning_rate": 7e-05,
|
| 374 |
+
"loss": 0.3529,
|
| 375 |
"step": 6100
|
| 376 |
},
|
| 377 |
{
|
| 378 |
+
"epoch": 0.2,
|
| 379 |
"learning_rate": 7e-05,
|
| 380 |
+
"loss": 0.368,
|
| 381 |
"step": 6200
|
| 382 |
},
|
| 383 |
{
|
| 384 |
+
"epoch": 0.2,
|
| 385 |
"learning_rate": 7e-05,
|
| 386 |
+
"loss": 0.3578,
|
| 387 |
"step": 6300
|
| 388 |
},
|
| 389 |
{
|
| 390 |
+
"epoch": 0.21,
|
| 391 |
"learning_rate": 7e-05,
|
| 392 |
+
"loss": 0.3463,
|
| 393 |
"step": 6400
|
| 394 |
},
|
| 395 |
{
|
| 396 |
+
"epoch": 0.21,
|
| 397 |
"learning_rate": 7e-05,
|
| 398 |
+
"loss": 0.3579,
|
| 399 |
"step": 6500
|
| 400 |
},
|
| 401 |
{
|
| 402 |
+
"epoch": 0.21,
|
| 403 |
"learning_rate": 7e-05,
|
| 404 |
+
"loss": 0.3628,
|
| 405 |
"step": 6600
|
| 406 |
},
|
| 407 |
{
|
| 408 |
+
"epoch": 0.22,
|
| 409 |
"learning_rate": 7e-05,
|
| 410 |
+
"loss": 0.3662,
|
| 411 |
"step": 6700
|
| 412 |
},
|
| 413 |
{
|
| 414 |
+
"epoch": 0.22,
|
| 415 |
"learning_rate": 7e-05,
|
| 416 |
+
"loss": 0.3689,
|
| 417 |
"step": 6800
|
| 418 |
},
|
| 419 |
{
|
| 420 |
+
"epoch": 0.22,
|
| 421 |
"learning_rate": 7e-05,
|
| 422 |
+
"loss": 0.3758,
|
| 423 |
"step": 6900
|
| 424 |
},
|
| 425 |
{
|
| 426 |
+
"epoch": 0.23,
|
| 427 |
"learning_rate": 7e-05,
|
| 428 |
+
"loss": 0.347,
|
| 429 |
"step": 7000
|
| 430 |
},
|
| 431 |
{
|
| 432 |
+
"epoch": 0.23,
|
| 433 |
"learning_rate": 7e-05,
|
| 434 |
+
"loss": 0.3525,
|
| 435 |
"step": 7100
|
| 436 |
},
|
| 437 |
{
|
| 438 |
+
"epoch": 0.23,
|
| 439 |
"learning_rate": 7e-05,
|
| 440 |
+
"loss": 0.3622,
|
| 441 |
"step": 7200
|
| 442 |
},
|
| 443 |
{
|
| 444 |
+
"epoch": 0.24,
|
| 445 |
"learning_rate": 7e-05,
|
| 446 |
+
"loss": 0.3462,
|
| 447 |
"step": 7300
|
| 448 |
},
|
| 449 |
{
|
| 450 |
+
"epoch": 0.24,
|
| 451 |
"learning_rate": 7e-05,
|
| 452 |
+
"loss": 0.3592,
|
| 453 |
"step": 7400
|
| 454 |
},
|
| 455 |
{
|
| 456 |
+
"epoch": 0.24,
|
| 457 |
"learning_rate": 7e-05,
|
| 458 |
+
"loss": 0.3551,
|
| 459 |
"step": 7500
|
| 460 |
},
|
| 461 |
{
|
| 462 |
+
"epoch": 0.24,
|
| 463 |
"learning_rate": 7e-05,
|
| 464 |
+
"loss": 0.3624,
|
| 465 |
"step": 7600
|
| 466 |
},
|
| 467 |
{
|
| 468 |
+
"epoch": 0.25,
|
| 469 |
"learning_rate": 7e-05,
|
| 470 |
+
"loss": 0.3676,
|
| 471 |
"step": 7700
|
| 472 |
},
|
| 473 |
{
|
| 474 |
+
"epoch": 0.25,
|
| 475 |
"learning_rate": 7e-05,
|
| 476 |
+
"loss": 0.3589,
|
| 477 |
"step": 7800
|
| 478 |
},
|
| 479 |
{
|
| 480 |
+
"epoch": 0.25,
|
| 481 |
"learning_rate": 7e-05,
|
| 482 |
+
"loss": 0.3474,
|
| 483 |
"step": 7900
|
| 484 |
},
|
| 485 |
{
|
| 486 |
+
"epoch": 0.26,
|
| 487 |
"learning_rate": 7e-05,
|
| 488 |
+
"loss": 0.3467,
|
| 489 |
"step": 8000
|
| 490 |
},
|
| 491 |
{
|
| 492 |
+
"epoch": 0.26,
|
| 493 |
"learning_rate": 7e-05,
|
| 494 |
+
"loss": 0.3568,
|
| 495 |
"step": 8100
|
| 496 |
},
|
| 497 |
{
|
| 498 |
+
"epoch": 0.26,
|
| 499 |
"learning_rate": 7e-05,
|
| 500 |
+
"loss": 0.37,
|
| 501 |
"step": 8200
|
| 502 |
},
|
| 503 |
{
|
| 504 |
+
"epoch": 0.27,
|
| 505 |
"learning_rate": 7e-05,
|
| 506 |
+
"loss": 0.358,
|
| 507 |
"step": 8300
|
| 508 |
},
|
| 509 |
{
|
| 510 |
+
"epoch": 0.27,
|
| 511 |
"learning_rate": 7e-05,
|
| 512 |
+
"loss": 0.3691,
|
| 513 |
"step": 8400
|
| 514 |
},
|
| 515 |
{
|
| 516 |
+
"epoch": 0.27,
|
| 517 |
"learning_rate": 7e-05,
|
| 518 |
+
"loss": 0.36,
|
| 519 |
"step": 8500
|
| 520 |
},
|
| 521 |
{
|
| 522 |
+
"epoch": 0.28,
|
| 523 |
"learning_rate": 7e-05,
|
| 524 |
+
"loss": 0.3666,
|
| 525 |
"step": 8600
|
| 526 |
},
|
| 527 |
{
|
| 528 |
+
"epoch": 0.28,
|
| 529 |
"learning_rate": 7e-05,
|
| 530 |
+
"loss": 0.3508,
|
| 531 |
"step": 8700
|
| 532 |
},
|
| 533 |
{
|
| 534 |
+
"epoch": 0.28,
|
| 535 |
"learning_rate": 7e-05,
|
| 536 |
+
"loss": 0.3457,
|
| 537 |
"step": 8800
|
| 538 |
},
|
| 539 |
{
|
| 540 |
+
"epoch": 0.29,
|
| 541 |
"learning_rate": 7e-05,
|
| 542 |
+
"loss": 0.3658,
|
| 543 |
"step": 8900
|
| 544 |
},
|
| 545 |
{
|
| 546 |
+
"epoch": 0.29,
|
| 547 |
"learning_rate": 7e-05,
|
| 548 |
+
"loss": 0.37,
|
| 549 |
"step": 9000
|
| 550 |
},
|
| 551 |
{
|
| 552 |
+
"epoch": 0.29,
|
| 553 |
"learning_rate": 7e-05,
|
| 554 |
+
"loss": 0.3657,
|
| 555 |
"step": 9100
|
| 556 |
},
|
| 557 |
{
|
| 558 |
+
"epoch": 0.3,
|
| 559 |
"learning_rate": 7e-05,
|
| 560 |
+
"loss": 0.3677,
|
| 561 |
"step": 9200
|
| 562 |
},
|
| 563 |
{
|
| 564 |
+
"epoch": 0.3,
|
| 565 |
"learning_rate": 7e-05,
|
| 566 |
+
"loss": 0.3692,
|
| 567 |
"step": 9300
|
| 568 |
},
|
| 569 |
{
|
| 570 |
+
"epoch": 0.3,
|
| 571 |
"learning_rate": 7e-05,
|
| 572 |
+
"loss": 0.366,
|
| 573 |
"step": 9400
|
| 574 |
},
|
| 575 |
{
|
| 576 |
+
"epoch": 0.31,
|
| 577 |
"learning_rate": 7e-05,
|
| 578 |
+
"loss": 0.3782,
|
| 579 |
"step": 9500
|
| 580 |
},
|
| 581 |
{
|
| 582 |
+
"epoch": 0.31,
|
| 583 |
"learning_rate": 7e-05,
|
| 584 |
+
"loss": 0.355,
|
| 585 |
"step": 9600
|
| 586 |
},
|
| 587 |
{
|
| 588 |
+
"epoch": 0.31,
|
| 589 |
"learning_rate": 7e-05,
|
| 590 |
+
"loss": 0.3562,
|
| 591 |
"step": 9700
|
| 592 |
},
|
| 593 |
{
|
| 594 |
+
"epoch": 0.32,
|
| 595 |
"learning_rate": 7e-05,
|
| 596 |
+
"loss": 0.3615,
|
| 597 |
"step": 9800
|
| 598 |
},
|
| 599 |
{
|
| 600 |
+
"epoch": 0.32,
|
| 601 |
"learning_rate": 7e-05,
|
| 602 |
+
"loss": 0.3597,
|
| 603 |
"step": 9900
|
| 604 |
},
|
| 605 |
{
|
| 606 |
+
"epoch": 0.32,
|
| 607 |
"learning_rate": 7e-05,
|
| 608 |
+
"loss": 0.3682,
|
| 609 |
"step": 10000
|
| 610 |
},
|
| 611 |
{
|
| 612 |
+
"epoch": 0.33,
|
| 613 |
"learning_rate": 7e-05,
|
| 614 |
+
"loss": 0.3818,
|
| 615 |
"step": 10100
|
| 616 |
},
|
| 617 |
{
|
| 618 |
+
"epoch": 0.33,
|
| 619 |
"learning_rate": 7e-05,
|
| 620 |
+
"loss": 0.3741,
|
| 621 |
"step": 10200
|
| 622 |
},
|
| 623 |
{
|
| 624 |
+
"epoch": 0.33,
|
| 625 |
"learning_rate": 7e-05,
|
| 626 |
+
"loss": 0.3695,
|
| 627 |
"step": 10300
|
| 628 |
},
|
| 629 |
{
|
| 630 |
+
"epoch": 0.34,
|
| 631 |
"learning_rate": 7e-05,
|
| 632 |
+
"loss": 0.3604,
|
| 633 |
"step": 10400
|
| 634 |
},
|
| 635 |
{
|
| 636 |
+
"epoch": 0.34,
|
| 637 |
"learning_rate": 7e-05,
|
| 638 |
+
"loss": 0.3679,
|
| 639 |
"step": 10500
|
| 640 |
},
|
| 641 |
{
|
| 642 |
+
"epoch": 0.34,
|
| 643 |
"learning_rate": 7e-05,
|
| 644 |
+
"loss": 0.3571,
|
| 645 |
"step": 10600
|
| 646 |
},
|
| 647 |
{
|
| 648 |
+
"epoch": 0.34,
|
| 649 |
"learning_rate": 7e-05,
|
| 650 |
+
"loss": 0.367,
|
| 651 |
"step": 10700
|
| 652 |
},
|
| 653 |
{
|
| 654 |
+
"epoch": 0.35,
|
| 655 |
"learning_rate": 7e-05,
|
| 656 |
+
"loss": 0.3479,
|
| 657 |
"step": 10800
|
| 658 |
},
|
| 659 |
{
|
| 660 |
+
"epoch": 0.35,
|
| 661 |
"learning_rate": 7e-05,
|
| 662 |
+
"loss": 0.3574,
|
| 663 |
"step": 10900
|
| 664 |
},
|
| 665 |
{
|
| 666 |
+
"epoch": 0.35,
|
| 667 |
"learning_rate": 7e-05,
|
| 668 |
+
"loss": 0.3602,
|
| 669 |
"step": 11000
|
| 670 |
},
|
| 671 |
{
|
| 672 |
+
"epoch": 0.36,
|
| 673 |
"learning_rate": 7e-05,
|
| 674 |
+
"loss": 0.3455,
|
| 675 |
"step": 11100
|
| 676 |
},
|
| 677 |
{
|
| 678 |
+
"epoch": 0.36,
|
| 679 |
"learning_rate": 7e-05,
|
| 680 |
+
"loss": 0.3559,
|
| 681 |
"step": 11200
|
| 682 |
},
|
| 683 |
{
|
| 684 |
+
"epoch": 0.36,
|
| 685 |
"learning_rate": 7e-05,
|
| 686 |
+
"loss": 0.3476,
|
| 687 |
"step": 11300
|
| 688 |
},
|
| 689 |
{
|
| 690 |
+
"epoch": 0.37,
|
| 691 |
"learning_rate": 7e-05,
|
| 692 |
+
"loss": 0.3659,
|
| 693 |
"step": 11400
|
| 694 |
},
|
| 695 |
{
|
| 696 |
+
"epoch": 0.37,
|
| 697 |
"learning_rate": 7e-05,
|
| 698 |
+
"loss": 0.3612,
|
| 699 |
"step": 11500
|
| 700 |
},
|
| 701 |
{
|
| 702 |
+
"epoch": 0.37,
|
| 703 |
"learning_rate": 7e-05,
|
| 704 |
+
"loss": 0.3524,
|
| 705 |
"step": 11600
|
| 706 |
},
|
| 707 |
{
|
| 708 |
+
"epoch": 0.38,
|
| 709 |
"learning_rate": 7e-05,
|
| 710 |
+
"loss": 0.3663,
|
| 711 |
"step": 11700
|
| 712 |
},
|
| 713 |
{
|
| 714 |
+
"epoch": 0.38,
|
| 715 |
"learning_rate": 7e-05,
|
| 716 |
+
"loss": 0.346,
|
| 717 |
"step": 11800
|
| 718 |
},
|
| 719 |
{
|
| 720 |
+
"epoch": 0.38,
|
| 721 |
"learning_rate": 7e-05,
|
| 722 |
+
"loss": 0.3569,
|
| 723 |
"step": 11900
|
| 724 |
},
|
| 725 |
{
|
| 726 |
+
"epoch": 0.39,
|
| 727 |
"learning_rate": 7e-05,
|
| 728 |
+
"loss": 0.3549,
|
| 729 |
"step": 12000
|
| 730 |
},
|
| 731 |
{
|
| 732 |
+
"epoch": 0.39,
|
| 733 |
"learning_rate": 7e-05,
|
| 734 |
+
"loss": 0.3698,
|
| 735 |
"step": 12100
|
| 736 |
},
|
| 737 |
{
|
| 738 |
+
"epoch": 0.39,
|
| 739 |
"learning_rate": 7e-05,
|
| 740 |
+
"loss": 0.3494,
|
| 741 |
"step": 12200
|
| 742 |
},
|
| 743 |
{
|
| 744 |
+
"epoch": 0.4,
|
| 745 |
"learning_rate": 7e-05,
|
| 746 |
+
"loss": 0.3592,
|
| 747 |
"step": 12300
|
| 748 |
},
|
| 749 |
{
|
| 750 |
+
"epoch": 0.4,
|
| 751 |
"learning_rate": 7e-05,
|
| 752 |
+
"loss": 0.3365,
|
| 753 |
"step": 12400
|
| 754 |
},
|
| 755 |
{
|
| 756 |
+
"epoch": 0.4,
|
| 757 |
"learning_rate": 7e-05,
|
| 758 |
+
"loss": 0.3527,
|
| 759 |
"step": 12500
|
| 760 |
},
|
| 761 |
{
|
| 762 |
+
"epoch": 0.41,
|
| 763 |
"learning_rate": 7e-05,
|
| 764 |
+
"loss": 0.3749,
|
| 765 |
"step": 12600
|
| 766 |
},
|
| 767 |
{
|
| 768 |
+
"epoch": 0.41,
|
| 769 |
"learning_rate": 7e-05,
|
| 770 |
+
"loss": 0.3476,
|
| 771 |
"step": 12700
|
| 772 |
},
|
| 773 |
{
|
| 774 |
+
"epoch": 0.41,
|
| 775 |
"learning_rate": 7e-05,
|
| 776 |
+
"loss": 0.377,
|
| 777 |
"step": 12800
|
| 778 |
},
|
| 779 |
{
|
| 780 |
+
"epoch": 0.42,
|
| 781 |
"learning_rate": 7e-05,
|
| 782 |
+
"loss": 0.3615,
|
| 783 |
"step": 12900
|
| 784 |
},
|
| 785 |
{
|
| 786 |
+
"epoch": 0.42,
|
| 787 |
"learning_rate": 7e-05,
|
| 788 |
+
"loss": 0.3604,
|
| 789 |
"step": 13000
|
| 790 |
},
|
| 791 |
{
|
| 792 |
+
"epoch": 0.42,
|
| 793 |
"learning_rate": 7e-05,
|
| 794 |
+
"loss": 0.3534,
|
| 795 |
"step": 13100
|
| 796 |
},
|
| 797 |
{
|
| 798 |
+
"epoch": 0.43,
|
| 799 |
"learning_rate": 7e-05,
|
| 800 |
+
"loss": 0.3597,
|
| 801 |
"step": 13200
|
| 802 |
},
|
| 803 |
{
|
| 804 |
+
"epoch": 0.43,
|
| 805 |
"learning_rate": 7e-05,
|
| 806 |
+
"loss": 0.3376,
|
| 807 |
"step": 13300
|
| 808 |
},
|
| 809 |
{
|
| 810 |
+
"epoch": 0.43,
|
| 811 |
"learning_rate": 7e-05,
|
| 812 |
+
"loss": 0.3576,
|
| 813 |
"step": 13400
|
| 814 |
},
|
| 815 |
{
|
| 816 |
+
"epoch": 0.43,
|
| 817 |
"learning_rate": 7e-05,
|
| 818 |
+
"loss": 0.357,
|
| 819 |
"step": 13500
|
| 820 |
},
|
| 821 |
{
|
| 822 |
+
"epoch": 0.44,
|
| 823 |
"learning_rate": 7e-05,
|
| 824 |
+
"loss": 0.3584,
|
| 825 |
"step": 13600
|
| 826 |
},
|
| 827 |
{
|
| 828 |
+
"epoch": 0.44,
|
| 829 |
"learning_rate": 7e-05,
|
| 830 |
+
"loss": 0.3394,
|
| 831 |
"step": 13700
|
| 832 |
},
|
| 833 |
{
|
| 834 |
+
"epoch": 0.44,
|
| 835 |
"learning_rate": 7e-05,
|
| 836 |
+
"loss": 0.3629,
|
| 837 |
"step": 13800
|
| 838 |
},
|
| 839 |
{
|
| 840 |
+
"epoch": 0.45,
|
| 841 |
"learning_rate": 7e-05,
|
| 842 |
+
"loss": 0.3707,
|
| 843 |
"step": 13900
|
| 844 |
},
|
| 845 |
{
|
| 846 |
+
"epoch": 0.45,
|
| 847 |
"learning_rate": 7e-05,
|
| 848 |
+
"loss": 0.3613,
|
| 849 |
"step": 14000
|
| 850 |
},
|
| 851 |
{
|
| 852 |
+
"epoch": 0.45,
|
| 853 |
"learning_rate": 7e-05,
|
| 854 |
+
"loss": 0.3557,
|
| 855 |
"step": 14100
|
| 856 |
},
|
| 857 |
{
|
| 858 |
+
"epoch": 0.46,
|
| 859 |
"learning_rate": 7e-05,
|
| 860 |
+
"loss": 0.3499,
|
| 861 |
"step": 14200
|
| 862 |
},
|
| 863 |
{
|
| 864 |
+
"epoch": 0.46,
|
| 865 |
"learning_rate": 7e-05,
|
| 866 |
+
"loss": 0.3644,
|
| 867 |
"step": 14300
|
| 868 |
},
|
| 869 |
{
|
| 870 |
+
"epoch": 0.46,
|
| 871 |
"learning_rate": 7e-05,
|
| 872 |
+
"loss": 0.3624,
|
| 873 |
"step": 14400
|
| 874 |
},
|
| 875 |
{
|
| 876 |
+
"epoch": 0.47,
|
| 877 |
"learning_rate": 7e-05,
|
| 878 |
+
"loss": 0.3672,
|
| 879 |
"step": 14500
|
| 880 |
},
|
| 881 |
{
|
| 882 |
+
"epoch": 0.47,
|
| 883 |
"learning_rate": 7e-05,
|
| 884 |
+
"loss": 0.3573,
|
| 885 |
"step": 14600
|
| 886 |
},
|
| 887 |
{
|
| 888 |
+
"epoch": 0.47,
|
| 889 |
"learning_rate": 7e-05,
|
| 890 |
+
"loss": 0.3572,
|
| 891 |
"step": 14700
|
| 892 |
},
|
| 893 |
{
|
| 894 |
+
"epoch": 0.48,
|
| 895 |
"learning_rate": 7e-05,
|
| 896 |
+
"loss": 0.3698,
|
| 897 |
"step": 14800
|
| 898 |
},
|
| 899 |
{
|
| 900 |
+
"epoch": 0.48,
|
| 901 |
"learning_rate": 7e-05,
|
| 902 |
+
"loss": 0.3638,
|
| 903 |
"step": 14900
|
| 904 |
},
|
| 905 |
{
|
| 906 |
+
"epoch": 0.48,
|
| 907 |
"learning_rate": 7e-05,
|
| 908 |
+
"loss": 0.3541,
|
| 909 |
"step": 15000
|
| 910 |
},
|
| 911 |
{
|
| 912 |
+
"epoch": 0.49,
|
| 913 |
"learning_rate": 7e-05,
|
| 914 |
+
"loss": 0.3533,
|
| 915 |
"step": 15100
|
| 916 |
},
|
| 917 |
{
|
| 918 |
+
"epoch": 0.49,
|
| 919 |
"learning_rate": 7e-05,
|
| 920 |
+
"loss": 0.3522,
|
| 921 |
"step": 15200
|
| 922 |
},
|
| 923 |
{
|
| 924 |
+
"epoch": 0.49,
|
| 925 |
"learning_rate": 7e-05,
|
| 926 |
+
"loss": 0.3643,
|
| 927 |
"step": 15300
|
| 928 |
},
|
| 929 |
{
|
| 930 |
+
"epoch": 0.5,
|
| 931 |
"learning_rate": 7e-05,
|
| 932 |
+
"loss": 0.355,
|
| 933 |
"step": 15400
|
| 934 |
},
|
| 935 |
{
|
| 936 |
+
"epoch": 0.5,
|
| 937 |
"learning_rate": 7e-05,
|
| 938 |
+
"loss": 0.3587,
|
| 939 |
"step": 15500
|
| 940 |
},
|
| 941 |
{
|
| 942 |
+
"epoch": 0.5,
|
| 943 |
"learning_rate": 7e-05,
|
| 944 |
+
"loss": 0.377,
|
| 945 |
"step": 15600
|
| 946 |
},
|
| 947 |
{
|
| 948 |
+
"epoch": 0.51,
|
| 949 |
"learning_rate": 7e-05,
|
| 950 |
+
"loss": 0.3594,
|
| 951 |
"step": 15700
|
| 952 |
},
|
| 953 |
{
|
| 954 |
+
"epoch": 0.51,
|
| 955 |
"learning_rate": 7e-05,
|
| 956 |
+
"loss": 0.3567,
|
| 957 |
"step": 15800
|
| 958 |
},
|
| 959 |
{
|
| 960 |
+
"epoch": 0.51,
|
| 961 |
"learning_rate": 7e-05,
|
| 962 |
+
"loss": 0.3627,
|
| 963 |
"step": 15900
|
| 964 |
},
|
| 965 |
{
|
| 966 |
+
"epoch": 0.52,
|
| 967 |
"learning_rate": 7e-05,
|
| 968 |
+
"loss": 0.3429,
|
| 969 |
"step": 16000
|
| 970 |
},
|
| 971 |
{
|
| 972 |
+
"epoch": 0.52,
|
| 973 |
"learning_rate": 7e-05,
|
| 974 |
+
"loss": 0.3555,
|
| 975 |
"step": 16100
|
| 976 |
},
|
| 977 |
{
|
| 978 |
+
"epoch": 0.52,
|
| 979 |
"learning_rate": 7e-05,
|
| 980 |
+
"loss": 0.3558,
|
| 981 |
"step": 16200
|
| 982 |
},
|
| 983 |
{
|
| 984 |
+
"epoch": 0.53,
|
| 985 |
"learning_rate": 7e-05,
|
| 986 |
+
"loss": 0.374,
|
| 987 |
"step": 16300
|
| 988 |
},
|
| 989 |
{
|
| 990 |
+
"epoch": 0.53,
|
| 991 |
"learning_rate": 7e-05,
|
| 992 |
+
"loss": 0.3602,
|
| 993 |
"step": 16400
|
| 994 |
},
|
| 995 |
{
|
| 996 |
+
"epoch": 0.53,
|
| 997 |
"learning_rate": 7e-05,
|
| 998 |
+
"loss": 0.3603,
|
| 999 |
"step": 16500
|
| 1000 |
},
|
| 1001 |
{
|
| 1002 |
+
"epoch": 0.53,
|
| 1003 |
"learning_rate": 7e-05,
|
| 1004 |
+
"loss": 0.3642,
|
| 1005 |
"step": 16600
|
| 1006 |
},
|
| 1007 |
{
|
| 1008 |
+
"epoch": 0.54,
|
| 1009 |
"learning_rate": 7e-05,
|
| 1010 |
+
"loss": 0.3631,
|
| 1011 |
"step": 16700
|
| 1012 |
},
|
| 1013 |
{
|
| 1014 |
+
"epoch": 0.54,
|
| 1015 |
"learning_rate": 7e-05,
|
| 1016 |
+
"loss": 0.3665,
|
| 1017 |
"step": 16800
|
| 1018 |
},
|
| 1019 |
{
|
| 1020 |
+
"epoch": 0.54,
|
| 1021 |
"learning_rate": 7e-05,
|
| 1022 |
+
"loss": 0.3725,
|
| 1023 |
"step": 16900
|
| 1024 |
},
|
| 1025 |
{
|
| 1026 |
+
"epoch": 0.55,
|
| 1027 |
"learning_rate": 7e-05,
|
| 1028 |
+
"loss": 0.3726,
|
| 1029 |
"step": 17000
|
| 1030 |
},
|
| 1031 |
{
|
| 1032 |
+
"epoch": 0.55,
|
| 1033 |
"learning_rate": 7e-05,
|
| 1034 |
+
"loss": 0.3464,
|
| 1035 |
"step": 17100
|
| 1036 |
},
|
| 1037 |
{
|
| 1038 |
+
"epoch": 0.55,
|
| 1039 |
"learning_rate": 7e-05,
|
| 1040 |
+
"loss": 0.3689,
|
| 1041 |
"step": 17200
|
| 1042 |
},
|
| 1043 |
{
|
| 1044 |
+
"epoch": 0.56,
|
| 1045 |
"learning_rate": 7e-05,
|
| 1046 |
+
"loss": 0.3635,
|
| 1047 |
"step": 17300
|
| 1048 |
},
|
| 1049 |
{
|
| 1050 |
+
"epoch": 0.56,
|
| 1051 |
"learning_rate": 7e-05,
|
| 1052 |
+
"loss": 0.3618,
|
| 1053 |
"step": 17400
|
| 1054 |
},
|
| 1055 |
{
|
| 1056 |
+
"epoch": 0.56,
|
| 1057 |
"learning_rate": 7e-05,
|
| 1058 |
+
"loss": 0.3706,
|
| 1059 |
"step": 17500
|
| 1060 |
},
|
| 1061 |
{
|
| 1062 |
+
"epoch": 0.57,
|
| 1063 |
"learning_rate": 7e-05,
|
| 1064 |
+
"loss": 0.3636,
|
| 1065 |
"step": 17600
|
| 1066 |
},
|
| 1067 |
{
|
| 1068 |
+
"epoch": 0.57,
|
| 1069 |
"learning_rate": 7e-05,
|
| 1070 |
+
"loss": 0.3652,
|
| 1071 |
"step": 17700
|
| 1072 |
},
|
| 1073 |
{
|
| 1074 |
+
"epoch": 0.57,
|
| 1075 |
"learning_rate": 7e-05,
|
| 1076 |
+
"loss": 0.3519,
|
| 1077 |
"step": 17800
|
| 1078 |
},
|
| 1079 |
{
|
| 1080 |
+
"epoch": 0.58,
|
| 1081 |
"learning_rate": 7e-05,
|
| 1082 |
+
"loss": 0.3633,
|
| 1083 |
"step": 17900
|
| 1084 |
},
|
| 1085 |
{
|
| 1086 |
+
"epoch": 0.58,
|
| 1087 |
"learning_rate": 7e-05,
|
| 1088 |
+
"loss": 0.3798,
|
| 1089 |
"step": 18000
|
| 1090 |
},
|
| 1091 |
{
|
| 1092 |
+
"epoch": 0.58,
|
| 1093 |
"learning_rate": 7e-05,
|
| 1094 |
+
"loss": 0.3414,
|
| 1095 |
"step": 18100
|
| 1096 |
},
|
| 1097 |
{
|
| 1098 |
+
"epoch": 0.59,
|
| 1099 |
"learning_rate": 7e-05,
|
| 1100 |
+
"loss": 0.3405,
|
| 1101 |
"step": 18200
|
| 1102 |
},
|
| 1103 |
{
|
| 1104 |
+
"epoch": 0.59,
|
| 1105 |
"learning_rate": 7e-05,
|
| 1106 |
+
"loss": 0.361,
|
| 1107 |
"step": 18300
|
| 1108 |
},
|
| 1109 |
{
|
| 1110 |
+
"epoch": 0.59,
|
| 1111 |
"learning_rate": 7e-05,
|
| 1112 |
+
"loss": 0.3777,
|
| 1113 |
"step": 18400
|
| 1114 |
},
|
| 1115 |
{
|
| 1116 |
+
"epoch": 0.6,
|
| 1117 |
"learning_rate": 7e-05,
|
| 1118 |
+
"loss": 0.3564,
|
| 1119 |
"step": 18500
|
| 1120 |
},
|
| 1121 |
{
|
| 1122 |
+
"epoch": 0.6,
|
| 1123 |
"learning_rate": 7e-05,
|
| 1124 |
+
"loss": 0.3567,
|
| 1125 |
"step": 18600
|
| 1126 |
},
|
| 1127 |
{
|
| 1128 |
+
"epoch": 0.6,
|
| 1129 |
"learning_rate": 7e-05,
|
| 1130 |
+
"loss": 0.3654,
|
| 1131 |
"step": 18700
|
| 1132 |
},
|
| 1133 |
{
|
| 1134 |
+
"epoch": 0.61,
|
| 1135 |
"learning_rate": 7e-05,
|
| 1136 |
+
"loss": 0.3418,
|
| 1137 |
"step": 18800
|
| 1138 |
},
|
| 1139 |
{
|
| 1140 |
+
"epoch": 0.61,
|
| 1141 |
"learning_rate": 7e-05,
|
| 1142 |
+
"loss": 0.3553,
|
| 1143 |
"step": 18900
|
| 1144 |
},
|
| 1145 |
{
|
| 1146 |
+
"epoch": 0.61,
|
| 1147 |
"learning_rate": 7e-05,
|
| 1148 |
+
"loss": 0.3686,
|
| 1149 |
"step": 19000
|
| 1150 |
},
|
| 1151 |
{
|
| 1152 |
+
"epoch": 0.62,
|
| 1153 |
"learning_rate": 7e-05,
|
| 1154 |
+
"loss": 0.3579,
|
| 1155 |
"step": 19100
|
| 1156 |
},
|
| 1157 |
{
|
| 1158 |
+
"epoch": 0.62,
|
| 1159 |
"learning_rate": 7e-05,
|
| 1160 |
+
"loss": 0.3597,
|
| 1161 |
"step": 19200
|
| 1162 |
},
|
| 1163 |
{
|
| 1164 |
+
"epoch": 0.62,
|
| 1165 |
"learning_rate": 7e-05,
|
| 1166 |
+
"loss": 0.3524,
|
| 1167 |
"step": 19300
|
| 1168 |
},
|
| 1169 |
{
|
| 1170 |
+
"epoch": 0.62,
|
| 1171 |
"learning_rate": 7e-05,
|
| 1172 |
+
"loss": 0.3459,
|
| 1173 |
"step": 19400
|
| 1174 |
},
|
| 1175 |
{
|
| 1176 |
+
"epoch": 0.63,
|
| 1177 |
"learning_rate": 7e-05,
|
| 1178 |
+
"loss": 0.3647,
|
| 1179 |
"step": 19500
|
| 1180 |
},
|
| 1181 |
{
|
| 1182 |
+
"epoch": 0.63,
|
| 1183 |
"learning_rate": 7e-05,
|
| 1184 |
+
"loss": 0.3553,
|
| 1185 |
"step": 19600
|
| 1186 |
},
|
| 1187 |
{
|
| 1188 |
+
"epoch": 0.63,
|
| 1189 |
"learning_rate": 7e-05,
|
| 1190 |
+
"loss": 0.3351,
|
| 1191 |
"step": 19700
|
| 1192 |
},
|
| 1193 |
{
|
| 1194 |
+
"epoch": 0.64,
|
| 1195 |
"learning_rate": 7e-05,
|
| 1196 |
+
"loss": 0.3583,
|
| 1197 |
"step": 19800
|
| 1198 |
},
|
| 1199 |
{
|
| 1200 |
+
"epoch": 0.64,
|
| 1201 |
"learning_rate": 7e-05,
|
| 1202 |
+
"loss": 0.3648,
|
| 1203 |
"step": 19900
|
| 1204 |
},
|
| 1205 |
{
|
| 1206 |
+
"epoch": 0.64,
|
| 1207 |
"learning_rate": 7e-05,
|
| 1208 |
+
"loss": 0.3533,
|
| 1209 |
"step": 20000
|
| 1210 |
},
|
| 1211 |
{
|
| 1212 |
+
"epoch": 0.65,
|
| 1213 |
"learning_rate": 7e-05,
|
| 1214 |
+
"loss": 0.3432,
|
| 1215 |
"step": 20100
|
| 1216 |
},
|
| 1217 |
{
|
| 1218 |
+
"epoch": 0.65,
|
| 1219 |
"learning_rate": 7e-05,
|
| 1220 |
+
"loss": 0.3592,
|
| 1221 |
"step": 20200
|
| 1222 |
},
|
| 1223 |
{
|
| 1224 |
+
"epoch": 0.65,
|
| 1225 |
"learning_rate": 7e-05,
|
| 1226 |
+
"loss": 0.3663,
|
| 1227 |
"step": 20300
|
| 1228 |
},
|
| 1229 |
{
|
| 1230 |
+
"epoch": 0.66,
|
| 1231 |
"learning_rate": 7e-05,
|
| 1232 |
+
"loss": 0.3716,
|
| 1233 |
"step": 20400
|
| 1234 |
},
|
| 1235 |
{
|
| 1236 |
+
"epoch": 0.66,
|
| 1237 |
"learning_rate": 7e-05,
|
| 1238 |
+
"loss": 0.3461,
|
| 1239 |
"step": 20500
|
| 1240 |
},
|
| 1241 |
{
|
| 1242 |
+
"epoch": 0.66,
|
| 1243 |
"learning_rate": 7e-05,
|
| 1244 |
+
"loss": 0.3683,
|
| 1245 |
"step": 20600
|
| 1246 |
},
|
| 1247 |
{
|
| 1248 |
+
"epoch": 0.67,
|
| 1249 |
"learning_rate": 7e-05,
|
| 1250 |
+
"loss": 0.3581,
|
| 1251 |
"step": 20700
|
| 1252 |
},
|
| 1253 |
{
|
| 1254 |
+
"epoch": 0.67,
|
| 1255 |
"learning_rate": 7e-05,
|
| 1256 |
+
"loss": 0.358,
|
| 1257 |
"step": 20800
|
| 1258 |
},
|
| 1259 |
{
|
| 1260 |
+
"epoch": 0.67,
|
| 1261 |
"learning_rate": 7e-05,
|
| 1262 |
+
"loss": 0.3629,
|
| 1263 |
"step": 20900
|
| 1264 |
},
|
| 1265 |
{
|
| 1266 |
+
"epoch": 0.68,
|
| 1267 |
"learning_rate": 7e-05,
|
| 1268 |
+
"loss": 0.3568,
|
| 1269 |
"step": 21000
|
| 1270 |
},
|
| 1271 |
{
|
| 1272 |
+
"epoch": 0.68,
|
| 1273 |
"learning_rate": 7e-05,
|
| 1274 |
+
"loss": 0.3432,
|
| 1275 |
"step": 21100
|
| 1276 |
},
|
| 1277 |
{
|
| 1278 |
+
"epoch": 0.68,
|
| 1279 |
"learning_rate": 7e-05,
|
| 1280 |
+
"loss": 0.3688,
|
| 1281 |
"step": 21200
|
| 1282 |
},
|
| 1283 |
{
|
| 1284 |
+
"epoch": 0.69,
|
| 1285 |
"learning_rate": 7e-05,
|
| 1286 |
+
"loss": 0.347,
|
| 1287 |
"step": 21300
|
| 1288 |
},
|
| 1289 |
{
|
| 1290 |
+
"epoch": 0.69,
|
| 1291 |
"learning_rate": 7e-05,
|
| 1292 |
+
"loss": 0.3672,
|
| 1293 |
"step": 21400
|
| 1294 |
},
|
| 1295 |
{
|
| 1296 |
+
"epoch": 0.69,
|
| 1297 |
"learning_rate": 7e-05,
|
| 1298 |
+
"loss": 0.3509,
|
| 1299 |
"step": 21500
|
| 1300 |
},
|
| 1301 |
{
|
| 1302 |
+
"epoch": 0.7,
|
| 1303 |
"learning_rate": 7e-05,
|
| 1304 |
+
"loss": 0.3538,
|
| 1305 |
"step": 21600
|
| 1306 |
},
|
| 1307 |
{
|
| 1308 |
+
"epoch": 0.7,
|
| 1309 |
"learning_rate": 7e-05,
|
| 1310 |
+
"loss": 0.3493,
|
| 1311 |
"step": 21700
|
| 1312 |
},
|
| 1313 |
{
|
| 1314 |
+
"epoch": 0.7,
|
| 1315 |
"learning_rate": 7e-05,
|
| 1316 |
+
"loss": 0.3354,
|
| 1317 |
"step": 21800
|
| 1318 |
},
|
| 1319 |
{
|
| 1320 |
+
"epoch": 0.71,
|
| 1321 |
"learning_rate": 7e-05,
|
| 1322 |
+
"loss": 0.3656,
|
| 1323 |
"step": 21900
|
| 1324 |
},
|
| 1325 |
{
|
| 1326 |
+
"epoch": 0.71,
|
| 1327 |
"learning_rate": 7e-05,
|
| 1328 |
+
"loss": 0.3581,
|
| 1329 |
"step": 22000
|
| 1330 |
},
|
| 1331 |
{
|
| 1332 |
+
"epoch": 0.71,
|
| 1333 |
"learning_rate": 7e-05,
|
| 1334 |
+
"loss": 0.3547,
|
| 1335 |
"step": 22100
|
| 1336 |
},
|
| 1337 |
{
|
| 1338 |
+
"epoch": 0.72,
|
| 1339 |
"learning_rate": 7e-05,
|
| 1340 |
+
"loss": 0.3526,
|
| 1341 |
"step": 22200
|
| 1342 |
},
|
| 1343 |
{
|
| 1344 |
+
"epoch": 0.72,
|
| 1345 |
"learning_rate": 7e-05,
|
| 1346 |
+
"loss": 0.35,
|
| 1347 |
"step": 22300
|
| 1348 |
},
|
| 1349 |
{
|
| 1350 |
+
"epoch": 0.72,
|
| 1351 |
"learning_rate": 7e-05,
|
| 1352 |
+
"loss": 0.3452,
|
| 1353 |
"step": 22400
|
| 1354 |
},
|
| 1355 |
{
|
| 1356 |
+
"epoch": 0.72,
|
| 1357 |
"learning_rate": 7e-05,
|
| 1358 |
+
"loss": 0.3586,
|
| 1359 |
"step": 22500
|
| 1360 |
},
|
| 1361 |
{
|
| 1362 |
+
"epoch": 0.73,
|
| 1363 |
"learning_rate": 7e-05,
|
| 1364 |
+
"loss": 0.3598,
|
| 1365 |
"step": 22600
|
| 1366 |
},
|
| 1367 |
{
|
| 1368 |
+
"epoch": 0.73,
|
| 1369 |
"learning_rate": 7e-05,
|
| 1370 |
+
"loss": 0.3608,
|
| 1371 |
"step": 22700
|
| 1372 |
},
|
| 1373 |
{
|
| 1374 |
+
"epoch": 0.73,
|
| 1375 |
"learning_rate": 7e-05,
|
| 1376 |
+
"loss": 0.3521,
|
| 1377 |
"step": 22800
|
| 1378 |
},
|
| 1379 |
{
|
| 1380 |
+
"epoch": 0.74,
|
| 1381 |
"learning_rate": 7e-05,
|
| 1382 |
+
"loss": 0.3393,
|
| 1383 |
"step": 22900
|
| 1384 |
},
|
| 1385 |
{
|
| 1386 |
+
"epoch": 0.74,
|
| 1387 |
"learning_rate": 7e-05,
|
| 1388 |
+
"loss": 0.3425,
|
| 1389 |
"step": 23000
|
| 1390 |
},
|
| 1391 |
{
|
| 1392 |
+
"epoch": 0.74,
|
| 1393 |
"learning_rate": 7e-05,
|
| 1394 |
+
"loss": 0.3533,
|
| 1395 |
"step": 23100
|
| 1396 |
},
|
| 1397 |
{
|
| 1398 |
+
"epoch": 0.75,
|
| 1399 |
"learning_rate": 7e-05,
|
| 1400 |
+
"loss": 0.3337,
|
| 1401 |
"step": 23200
|
| 1402 |
},
|
| 1403 |
{
|
| 1404 |
+
"epoch": 0.75,
|
| 1405 |
"learning_rate": 7e-05,
|
| 1406 |
+
"loss": 0.3693,
|
| 1407 |
"step": 23300
|
| 1408 |
},
|
| 1409 |
{
|
| 1410 |
+
"epoch": 0.75,
|
| 1411 |
"learning_rate": 7e-05,
|
| 1412 |
+
"loss": 0.3425,
|
| 1413 |
"step": 23400
|
| 1414 |
},
|
| 1415 |
{
|
| 1416 |
+
"epoch": 0.76,
|
| 1417 |
"learning_rate": 7e-05,
|
| 1418 |
+
"loss": 0.3775,
|
| 1419 |
"step": 23500
|
| 1420 |
},
|
| 1421 |
{
|
| 1422 |
+
"epoch": 0.76,
|
| 1423 |
"learning_rate": 7e-05,
|
| 1424 |
+
"loss": 0.3319,
|
| 1425 |
"step": 23600
|
| 1426 |
},
|
| 1427 |
{
|
| 1428 |
+
"epoch": 0.76,
|
| 1429 |
"learning_rate": 7e-05,
|
| 1430 |
+
"loss": 0.3546,
|
| 1431 |
"step": 23700
|
| 1432 |
},
|
| 1433 |
{
|
| 1434 |
+
"epoch": 0.77,
|
| 1435 |
"learning_rate": 7e-05,
|
| 1436 |
+
"loss": 0.356,
|
| 1437 |
"step": 23800
|
| 1438 |
},
|
| 1439 |
{
|
| 1440 |
+
"epoch": 0.77,
|
| 1441 |
"learning_rate": 7e-05,
|
| 1442 |
+
"loss": 0.3533,
|
| 1443 |
"step": 23900
|
| 1444 |
},
|
| 1445 |
{
|
| 1446 |
+
"epoch": 0.77,
|
| 1447 |
"learning_rate": 7e-05,
|
| 1448 |
+
"loss": 0.3589,
|
| 1449 |
"step": 24000
|
| 1450 |
},
|
| 1451 |
{
|
| 1452 |
+
"epoch": 0.78,
|
| 1453 |
"learning_rate": 7e-05,
|
| 1454 |
+
"loss": 0.3502,
|
| 1455 |
"step": 24100
|
| 1456 |
},
|
| 1457 |
{
|
| 1458 |
+
"epoch": 0.78,
|
| 1459 |
"learning_rate": 7e-05,
|
| 1460 |
+
"loss": 0.3625,
|
| 1461 |
"step": 24200
|
| 1462 |
},
|
| 1463 |
{
|
| 1464 |
+
"epoch": 0.78,
|
| 1465 |
"learning_rate": 7e-05,
|
| 1466 |
+
"loss": 0.3531,
|
| 1467 |
"step": 24300
|
| 1468 |
},
|
| 1469 |
{
|
| 1470 |
+
"epoch": 0.79,
|
| 1471 |
"learning_rate": 7e-05,
|
| 1472 |
+
"loss": 0.3592,
|
| 1473 |
"step": 24400
|
| 1474 |
},
|
| 1475 |
{
|
| 1476 |
+
"epoch": 0.79,
|
| 1477 |
"learning_rate": 7e-05,
|
| 1478 |
+
"loss": 0.3519,
|
| 1479 |
"step": 24500
|
| 1480 |
},
|
| 1481 |
{
|
| 1482 |
+
"epoch": 0.79,
|
| 1483 |
"learning_rate": 7e-05,
|
| 1484 |
+
"loss": 0.3729,
|
| 1485 |
"step": 24600
|
| 1486 |
},
|
| 1487 |
{
|
| 1488 |
+
"epoch": 0.8,
|
| 1489 |
"learning_rate": 7e-05,
|
| 1490 |
+
"loss": 0.3504,
|
| 1491 |
"step": 24700
|
| 1492 |
},
|
| 1493 |
{
|
| 1494 |
+
"epoch": 0.8,
|
| 1495 |
"learning_rate": 7e-05,
|
| 1496 |
+
"loss": 0.3532,
|
| 1497 |
"step": 24800
|
| 1498 |
},
|
| 1499 |
{
|
| 1500 |
+
"epoch": 0.8,
|
| 1501 |
"learning_rate": 7e-05,
|
| 1502 |
+
"loss": 0.3552,
|
| 1503 |
"step": 24900
|
| 1504 |
},
|
| 1505 |
{
|
| 1506 |
+
"epoch": 0.81,
|
| 1507 |
"learning_rate": 7e-05,
|
| 1508 |
+
"loss": 0.3739,
|
| 1509 |
"step": 25000
|
| 1510 |
},
|
| 1511 |
{
|
| 1512 |
+
"epoch": 0.81,
|
| 1513 |
"learning_rate": 7e-05,
|
| 1514 |
+
"loss": 0.3445,
|
| 1515 |
"step": 25100
|
| 1516 |
},
|
| 1517 |
{
|
| 1518 |
+
"epoch": 0.81,
|
| 1519 |
"learning_rate": 7e-05,
|
| 1520 |
+
"loss": 0.3482,
|
| 1521 |
"step": 25200
|
| 1522 |
},
|
| 1523 |
{
|
| 1524 |
+
"epoch": 0.82,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1525 |
"learning_rate": 7e-05,
|
| 1526 |
"loss": 0.3624,
|
| 1527 |
+
"step": 25300
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1528 |
},
|
| 1529 |
{
|
| 1530 |
"epoch": 0.82,
|
| 1531 |
"learning_rate": 7e-05,
|
| 1532 |
+
"loss": 0.343,
|
| 1533 |
+
"step": 25400
|
| 1534 |
},
|
| 1535 |
{
|
| 1536 |
"epoch": 0.82,
|
| 1537 |
"learning_rate": 7e-05,
|
| 1538 |
+
"loss": 0.3656,
|
| 1539 |
+
"step": 25500
|
| 1540 |
},
|
| 1541 |
{
|
| 1542 |
"epoch": 0.82,
|
| 1543 |
"learning_rate": 7e-05,
|
| 1544 |
+
"loss": 0.3474,
|
| 1545 |
+
"step": 25600
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1546 |
},
|
| 1547 |
{
|
| 1548 |
"epoch": 0.83,
|
| 1549 |
"learning_rate": 7e-05,
|
| 1550 |
+
"loss": 0.3582,
|
| 1551 |
+
"step": 25700
|
| 1552 |
},
|
| 1553 |
{
|
| 1554 |
"epoch": 0.83,
|
| 1555 |
"learning_rate": 7e-05,
|
| 1556 |
+
"loss": 0.3675,
|
| 1557 |
+
"step": 25800
|
| 1558 |
},
|
| 1559 |
{
|
| 1560 |
"epoch": 0.83,
|
| 1561 |
"learning_rate": 7e-05,
|
| 1562 |
+
"loss": 0.3547,
|
| 1563 |
+
"step": 25900
|
| 1564 |
},
|
| 1565 |
{
|
| 1566 |
"epoch": 0.84,
|
| 1567 |
"learning_rate": 7e-05,
|
| 1568 |
+
"loss": 0.3549,
|
| 1569 |
+
"step": 26000
|
| 1570 |
},
|
| 1571 |
{
|
| 1572 |
"epoch": 0.84,
|
| 1573 |
"learning_rate": 7e-05,
|
| 1574 |
+
"loss": 0.3592,
|
| 1575 |
+
"step": 26100
|
| 1576 |
},
|
| 1577 |
{
|
| 1578 |
"epoch": 0.84,
|
| 1579 |
"learning_rate": 7e-05,
|
| 1580 |
+
"loss": 0.3546,
|
| 1581 |
+
"step": 26200
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1582 |
},
|
| 1583 |
{
|
| 1584 |
"epoch": 0.85,
|
| 1585 |
"learning_rate": 7e-05,
|
| 1586 |
+
"loss": 0.3508,
|
| 1587 |
+
"step": 26300
|
| 1588 |
},
|
| 1589 |
{
|
| 1590 |
"epoch": 0.85,
|
| 1591 |
"learning_rate": 7e-05,
|
| 1592 |
+
"loss": 0.339,
|
| 1593 |
+
"step": 26400
|
| 1594 |
},
|
| 1595 |
{
|
| 1596 |
"epoch": 0.85,
|
| 1597 |
"learning_rate": 7e-05,
|
| 1598 |
+
"loss": 0.3557,
|
| 1599 |
+
"step": 26500
|
| 1600 |
},
|
| 1601 |
{
|
| 1602 |
"epoch": 0.86,
|
| 1603 |
"learning_rate": 7e-05,
|
| 1604 |
+
"loss": 0.3742,
|
| 1605 |
+
"step": 26600
|
| 1606 |
},
|
| 1607 |
{
|
| 1608 |
"epoch": 0.86,
|
| 1609 |
"learning_rate": 7e-05,
|
| 1610 |
+
"loss": 0.3491,
|
| 1611 |
+
"step": 26700
|
| 1612 |
},
|
| 1613 |
{
|
| 1614 |
"epoch": 0.86,
|
| 1615 |
"learning_rate": 7e-05,
|
| 1616 |
+
"loss": 0.3663,
|
| 1617 |
+
"step": 26800
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1618 |
},
|
| 1619 |
{
|
| 1620 |
"epoch": 0.87,
|
| 1621 |
"learning_rate": 7e-05,
|
| 1622 |
+
"loss": 0.3625,
|
| 1623 |
+
"step": 26900
|
| 1624 |
},
|
| 1625 |
{
|
| 1626 |
"epoch": 0.87,
|
| 1627 |
"learning_rate": 7e-05,
|
| 1628 |
+
"loss": 0.3421,
|
| 1629 |
+
"step": 27000
|
| 1630 |
},
|
| 1631 |
{
|
| 1632 |
"epoch": 0.87,
|
| 1633 |
"learning_rate": 7e-05,
|
| 1634 |
+
"loss": 0.3412,
|
| 1635 |
+
"step": 27100
|
| 1636 |
},
|
| 1637 |
{
|
| 1638 |
"epoch": 0.88,
|
| 1639 |
"learning_rate": 7e-05,
|
| 1640 |
+
"loss": 0.3576,
|
| 1641 |
+
"step": 27200
|
| 1642 |
},
|
| 1643 |
{
|
| 1644 |
"epoch": 0.88,
|
| 1645 |
"learning_rate": 7e-05,
|
| 1646 |
+
"loss": 0.3635,
|
| 1647 |
+
"step": 27300
|
| 1648 |
},
|
| 1649 |
{
|
| 1650 |
"epoch": 0.88,
|
| 1651 |
"learning_rate": 7e-05,
|
| 1652 |
+
"loss": 0.3554,
|
| 1653 |
+
"step": 27400
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1654 |
},
|
| 1655 |
{
|
| 1656 |
"epoch": 0.89,
|
| 1657 |
"learning_rate": 7e-05,
|
| 1658 |
+
"loss": 0.368,
|
| 1659 |
+
"step": 27500
|
| 1660 |
},
|
| 1661 |
{
|
| 1662 |
"epoch": 0.89,
|
| 1663 |
"learning_rate": 7e-05,
|
| 1664 |
+
"loss": 0.3524,
|
| 1665 |
+
"step": 27600
|
| 1666 |
},
|
| 1667 |
{
|
| 1668 |
"epoch": 0.89,
|
| 1669 |
"learning_rate": 7e-05,
|
| 1670 |
+
"loss": 0.3484,
|
| 1671 |
+
"step": 27700
|
| 1672 |
},
|
| 1673 |
{
|
| 1674 |
"epoch": 0.9,
|
| 1675 |
"learning_rate": 7e-05,
|
| 1676 |
+
"loss": 0.3473,
|
| 1677 |
+
"step": 27800
|
| 1678 |
},
|
| 1679 |
{
|
| 1680 |
"epoch": 0.9,
|
| 1681 |
"learning_rate": 7e-05,
|
| 1682 |
+
"loss": 0.3328,
|
| 1683 |
+
"step": 27900
|
| 1684 |
},
|
| 1685 |
{
|
| 1686 |
"epoch": 0.9,
|
| 1687 |
"learning_rate": 7e-05,
|
| 1688 |
+
"loss": 0.3554,
|
| 1689 |
+
"step": 28000
|
| 1690 |
},
|
| 1691 |
{
|
| 1692 |
"epoch": 0.91,
|
| 1693 |
"learning_rate": 7e-05,
|
| 1694 |
+
"loss": 0.3536,
|
| 1695 |
+
"step": 28100
|
| 1696 |
},
|
| 1697 |
{
|
| 1698 |
"epoch": 0.91,
|
| 1699 |
"learning_rate": 7e-05,
|
| 1700 |
+
"loss": 0.3424,
|
| 1701 |
+
"step": 28200
|
| 1702 |
},
|
| 1703 |
{
|
| 1704 |
"epoch": 0.91,
|
| 1705 |
"learning_rate": 7e-05,
|
| 1706 |
+
"loss": 0.3631,
|
| 1707 |
+
"step": 28300
|
| 1708 |
},
|
| 1709 |
{
|
| 1710 |
"epoch": 0.91,
|
| 1711 |
"learning_rate": 7e-05,
|
| 1712 |
+
"loss": 0.333,
|
| 1713 |
+
"step": 28400
|
| 1714 |
},
|
| 1715 |
{
|
| 1716 |
"epoch": 0.92,
|
| 1717 |
"learning_rate": 7e-05,
|
| 1718 |
+
"loss": 0.3468,
|
| 1719 |
+
"step": 28500
|
| 1720 |
},
|
| 1721 |
{
|
| 1722 |
"epoch": 0.92,
|
| 1723 |
"learning_rate": 7e-05,
|
| 1724 |
+
"loss": 0.3487,
|
| 1725 |
+
"step": 28600
|
| 1726 |
},
|
| 1727 |
{
|
| 1728 |
"epoch": 0.92,
|
| 1729 |
"learning_rate": 7e-05,
|
| 1730 |
+
"loss": 0.3542,
|
| 1731 |
+
"step": 28700
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1732 |
},
|
| 1733 |
{
|
| 1734 |
"epoch": 0.93,
|
| 1735 |
"learning_rate": 7e-05,
|
| 1736 |
+
"loss": 0.3579,
|
| 1737 |
+
"step": 28800
|
| 1738 |
},
|
| 1739 |
{
|
| 1740 |
"epoch": 0.93,
|
| 1741 |
"learning_rate": 7e-05,
|
| 1742 |
+
"loss": 0.3641,
|
| 1743 |
+
"step": 28900
|
| 1744 |
},
|
| 1745 |
{
|
| 1746 |
"epoch": 0.93,
|
| 1747 |
"learning_rate": 7e-05,
|
| 1748 |
+
"loss": 0.3729,
|
| 1749 |
+
"step": 29000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1750 |
},
|
| 1751 |
{
|
| 1752 |
"epoch": 0.94,
|
| 1753 |
"learning_rate": 7e-05,
|
| 1754 |
+
"loss": 0.3395,
|
| 1755 |
+
"step": 29100
|
| 1756 |
},
|
| 1757 |
{
|
| 1758 |
"epoch": 0.94,
|
| 1759 |
"learning_rate": 7e-05,
|
| 1760 |
+
"loss": 0.3572,
|
| 1761 |
+
"step": 29200
|
| 1762 |
},
|
| 1763 |
{
|
| 1764 |
"epoch": 0.94,
|
| 1765 |
"learning_rate": 7e-05,
|
| 1766 |
+
"loss": 0.3525,
|
| 1767 |
+
"step": 29300
|
| 1768 |
},
|
| 1769 |
{
|
| 1770 |
"epoch": 0.95,
|
| 1771 |
"learning_rate": 7e-05,
|
| 1772 |
+
"loss": 0.357,
|
| 1773 |
+
"step": 29400
|
| 1774 |
},
|
| 1775 |
{
|
| 1776 |
"epoch": 0.95,
|
| 1777 |
"learning_rate": 7e-05,
|
| 1778 |
+
"loss": 0.3647,
|
| 1779 |
+
"step": 29500
|
| 1780 |
},
|
| 1781 |
{
|
| 1782 |
"epoch": 0.95,
|
| 1783 |
"learning_rate": 7e-05,
|
| 1784 |
+
"loss": 0.3675,
|
| 1785 |
+
"step": 29600
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1786 |
},
|
| 1787 |
{
|
| 1788 |
"epoch": 0.96,
|
| 1789 |
"learning_rate": 7e-05,
|
| 1790 |
+
"loss": 0.365,
|
| 1791 |
+
"step": 29700
|
| 1792 |
},
|
| 1793 |
{
|
| 1794 |
"epoch": 0.96,
|
| 1795 |
"learning_rate": 7e-05,
|
| 1796 |
+
"loss": 0.3439,
|
| 1797 |
+
"step": 29800
|
| 1798 |
},
|
| 1799 |
{
|
| 1800 |
"epoch": 0.96,
|
| 1801 |
"learning_rate": 7e-05,
|
| 1802 |
+
"loss": 0.3501,
|
| 1803 |
+
"step": 29900
|
| 1804 |
},
|
| 1805 |
{
|
| 1806 |
"epoch": 0.97,
|
| 1807 |
"learning_rate": 7e-05,
|
| 1808 |
+
"loss": 0.3552,
|
| 1809 |
+
"step": 30000
|
| 1810 |
},
|
| 1811 |
{
|
| 1812 |
"epoch": 0.97,
|
| 1813 |
"learning_rate": 7e-05,
|
| 1814 |
+
"loss": 0.351,
|
| 1815 |
+
"step": 30100
|
| 1816 |
},
|
| 1817 |
{
|
| 1818 |
"epoch": 0.97,
|
| 1819 |
"learning_rate": 7e-05,
|
| 1820 |
+
"loss": 0.3591,
|
| 1821 |
+
"step": 30200
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1822 |
},
|
| 1823 |
{
|
| 1824 |
"epoch": 0.98,
|
| 1825 |
"learning_rate": 7e-05,
|
| 1826 |
+
"loss": 0.3548,
|
| 1827 |
+
"step": 30300
|
| 1828 |
},
|
| 1829 |
{
|
| 1830 |
"epoch": 0.98,
|
| 1831 |
"learning_rate": 7e-05,
|
| 1832 |
+
"loss": 0.3608,
|
| 1833 |
+
"step": 30400
|
| 1834 |
},
|
| 1835 |
{
|
| 1836 |
"epoch": 0.98,
|
| 1837 |
"learning_rate": 7e-05,
|
| 1838 |
+
"loss": 0.3504,
|
| 1839 |
+
"step": 30500
|
| 1840 |
},
|
| 1841 |
{
|
| 1842 |
"epoch": 0.99,
|
| 1843 |
"learning_rate": 7e-05,
|
| 1844 |
+
"loss": 0.3514,
|
| 1845 |
+
"step": 30600
|
| 1846 |
},
|
| 1847 |
{
|
| 1848 |
"epoch": 0.99,
|
| 1849 |
"learning_rate": 7e-05,
|
| 1850 |
+
"loss": 0.3456,
|
| 1851 |
+
"step": 30700
|
| 1852 |
},
|
| 1853 |
{
|
| 1854 |
"epoch": 0.99,
|
| 1855 |
"learning_rate": 7e-05,
|
| 1856 |
+
"loss": 0.343,
|
| 1857 |
+
"step": 30800
|
| 1858 |
},
|
| 1859 |
{
|
| 1860 |
"epoch": 1.0,
|
| 1861 |
"learning_rate": 7e-05,
|
| 1862 |
+
"loss": 0.3548,
|
| 1863 |
+
"step": 30900
|
| 1864 |
},
|
| 1865 |
{
|
| 1866 |
"epoch": 1.0,
|
| 1867 |
"learning_rate": 7e-05,
|
| 1868 |
+
"loss": 0.3601,
|
| 1869 |
+
"step": 31000
|
| 1870 |
},
|
| 1871 |
{
|
| 1872 |
"epoch": 1.0,
|
| 1873 |
+
"eval_gen_len": 24.568762132298044,
|
| 1874 |
+
"eval_loss": 0.3032413423061371,
|
| 1875 |
+
"eval_rouge1": 86.5313,
|
| 1876 |
+
"eval_rouge2": 75.3351,
|
| 1877 |
+
"eval_rougeL": 85.9565,
|
| 1878 |
+
"eval_rougeLsum": 85.9785,
|
| 1879 |
+
"eval_runtime": 1478.6198,
|
| 1880 |
+
"eval_samples_per_second": 4.529,
|
| 1881 |
+
"eval_steps_per_second": 1.133,
|
| 1882 |
+
"step": 31042
|
| 1883 |
},
|
| 1884 |
{
|
| 1885 |
"epoch": 1.0,
|
| 1886 |
+
"step": 31042,
|
| 1887 |
+
"total_flos": 8.055925371907523e+17,
|
| 1888 |
+
"train_loss": 0.35858034201253897,
|
| 1889 |
+
"train_runtime": 38927.7212,
|
| 1890 |
+
"train_samples_per_second": 25.518,
|
| 1891 |
+
"train_steps_per_second": 0.797
|
| 1892 |
}
|
| 1893 |
],
|
| 1894 |
"logging_steps": 100,
|
| 1895 |
+
"max_steps": 31042,
|
| 1896 |
"num_input_tokens_seen": 0,
|
| 1897 |
"num_train_epochs": 1,
|
| 1898 |
"save_steps": 2500,
|
| 1899 |
+
"total_flos": 8.055925371907523e+17,
|
| 1900 |
"trial_name": null,
|
| 1901 |
"trial_params": null
|
| 1902 |
}
|