🍻 cheers
Browse files- README.md +7 -1
- all_results.json +11 -11
- eval_results.json +6 -6
- runs/Dec14_01-48-11_784ca4b30f2d/events.out.tfevents.1702519483.784ca4b30f2d.1711.1 +3 -0
- train_results.json +6 -6
- trainer_state.json +1200 -228
README.md
CHANGED
|
@@ -2,9 +2,12 @@
|
|
| 2 |
license: apache-2.0
|
| 3 |
base_model: google/vit-base-patch16-224-in21k
|
| 4 |
tags:
|
|
|
|
| 5 |
- generated_from_trainer
|
| 6 |
datasets:
|
| 7 |
- imagefolder
|
|
|
|
|
|
|
| 8 |
model-index:
|
| 9 |
- name: xyz
|
| 10 |
results: []
|
|
@@ -15,7 +18,10 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 15 |
|
| 16 |
# xyz
|
| 17 |
|
| 18 |
-
This model is a fine-tuned version of [google/vit-base-patch16-224-in21k](https://huggingface.co/google/vit-base-patch16-224-in21k) on the
|
|
|
|
|
|
|
|
|
|
| 19 |
|
| 20 |
## Model description
|
| 21 |
|
|
|
|
| 2 |
license: apache-2.0
|
| 3 |
base_model: google/vit-base-patch16-224-in21k
|
| 4 |
tags:
|
| 5 |
+
- image-classification
|
| 6 |
- generated_from_trainer
|
| 7 |
datasets:
|
| 8 |
- imagefolder
|
| 9 |
+
metrics:
|
| 10 |
+
- accuracy
|
| 11 |
model-index:
|
| 12 |
- name: xyz
|
| 13 |
results: []
|
|
|
|
| 18 |
|
| 19 |
# xyz
|
| 20 |
|
| 21 |
+
This model is a fine-tuned version of [google/vit-base-patch16-224-in21k](https://huggingface.co/google/vit-base-patch16-224-in21k) on the mclr dataset.
|
| 22 |
+
It achieves the following results on the evaluation set:
|
| 23 |
+
- Loss: 0.3957
|
| 24 |
+
- Accuracy: 0.9157
|
| 25 |
|
| 26 |
## Model description
|
| 27 |
|
all_results.json
CHANGED
|
@@ -1,13 +1,13 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"eval_accuracy": 0.
|
| 4 |
-
"eval_loss": 0.
|
| 5 |
-
"eval_runtime":
|
| 6 |
-
"eval_samples_per_second":
|
| 7 |
-
"eval_steps_per_second":
|
| 8 |
-
"total_flos":
|
| 9 |
-
"train_loss": 0.
|
| 10 |
-
"train_runtime":
|
| 11 |
-
"train_samples_per_second":
|
| 12 |
-
"train_steps_per_second": 2.
|
| 13 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 10.0,
|
| 3 |
+
"eval_accuracy": 0.9157407407407407,
|
| 4 |
+
"eval_loss": 0.39567169547080994,
|
| 5 |
+
"eval_runtime": 20.1724,
|
| 6 |
+
"eval_samples_per_second": 53.539,
|
| 7 |
+
"eval_steps_per_second": 6.692,
|
| 8 |
+
"total_flos": 3.347863979094835e+18,
|
| 9 |
+
"train_loss": 0.2644773021671507,
|
| 10 |
+
"train_runtime": 949.4602,
|
| 11 |
+
"train_samples_per_second": 45.5,
|
| 12 |
+
"train_steps_per_second": 2.844
|
| 13 |
}
|
eval_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"eval_accuracy": 0.
|
| 4 |
-
"eval_loss": 0.
|
| 5 |
-
"eval_runtime":
|
| 6 |
-
"eval_samples_per_second":
|
| 7 |
-
"eval_steps_per_second":
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 10.0,
|
| 3 |
+
"eval_accuracy": 0.9157407407407407,
|
| 4 |
+
"eval_loss": 0.39567169547080994,
|
| 5 |
+
"eval_runtime": 20.1724,
|
| 6 |
+
"eval_samples_per_second": 53.539,
|
| 7 |
+
"eval_steps_per_second": 6.692
|
| 8 |
}
|
runs/Dec14_01-48-11_784ca4b30f2d/events.out.tfevents.1702519483.784ca4b30f2d.1711.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9f5c2e3004e91687627e6ef69597c2aefd563841c7225a04a592a3473930b241
|
| 3 |
+
size 734
|
train_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"total_flos":
|
| 4 |
-
"train_loss": 0.
|
| 5 |
-
"train_runtime":
|
| 6 |
-
"train_samples_per_second":
|
| 7 |
-
"train_steps_per_second": 2.
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 10.0,
|
| 3 |
+
"total_flos": 3.347863979094835e+18,
|
| 4 |
+
"train_loss": 0.2644773021671507,
|
| 5 |
+
"train_runtime": 949.4602,
|
| 6 |
+
"train_samples_per_second": 45.5,
|
| 7 |
+
"train_steps_per_second": 2.844
|
| 8 |
}
|
trainer_state.json
CHANGED
|
@@ -1,676 +1,1648 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch":
|
| 5 |
"eval_steps": 3000,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
"epoch": 0.04,
|
| 13 |
-
"learning_rate": 0.
|
| 14 |
-
"loss": 2.
|
| 15 |
"step": 10
|
| 16 |
},
|
| 17 |
{
|
| 18 |
"epoch": 0.07,
|
| 19 |
-
"learning_rate": 0.
|
| 20 |
-
"loss":
|
| 21 |
"step": 20
|
| 22 |
},
|
| 23 |
{
|
| 24 |
"epoch": 0.11,
|
| 25 |
-
"learning_rate": 0.
|
| 26 |
-
"loss": 1.
|
| 27 |
"step": 30
|
| 28 |
},
|
| 29 |
{
|
| 30 |
"epoch": 0.15,
|
| 31 |
-
"learning_rate": 0.
|
| 32 |
-
"loss": 1.
|
| 33 |
"step": 40
|
| 34 |
},
|
| 35 |
{
|
| 36 |
"epoch": 0.19,
|
| 37 |
-
"learning_rate": 0.
|
| 38 |
-
"loss": 1.
|
| 39 |
"step": 50
|
| 40 |
},
|
| 41 |
{
|
| 42 |
"epoch": 0.22,
|
| 43 |
-
"learning_rate": 0.
|
| 44 |
-
"loss": 1.
|
| 45 |
"step": 60
|
| 46 |
},
|
| 47 |
{
|
| 48 |
"epoch": 0.26,
|
| 49 |
-
"learning_rate": 0.
|
| 50 |
-
"loss": 1.
|
| 51 |
"step": 70
|
| 52 |
},
|
| 53 |
{
|
| 54 |
"epoch": 0.3,
|
| 55 |
-
"learning_rate": 0.
|
| 56 |
-
"loss": 1.
|
| 57 |
"step": 80
|
| 58 |
},
|
| 59 |
{
|
| 60 |
"epoch": 0.33,
|
| 61 |
-
"learning_rate": 0.
|
| 62 |
-
"loss": 1.
|
| 63 |
"step": 90
|
| 64 |
},
|
| 65 |
{
|
| 66 |
"epoch": 0.37,
|
| 67 |
-
"learning_rate": 0.
|
| 68 |
-
"loss": 1.
|
| 69 |
"step": 100
|
| 70 |
},
|
| 71 |
{
|
| 72 |
"epoch": 0.41,
|
| 73 |
-
"learning_rate": 0.
|
| 74 |
-
"loss": 1.
|
| 75 |
"step": 110
|
| 76 |
},
|
| 77 |
{
|
| 78 |
"epoch": 0.44,
|
| 79 |
-
"learning_rate": 0.
|
| 80 |
-
"loss": 1.
|
| 81 |
"step": 120
|
| 82 |
},
|
| 83 |
{
|
| 84 |
"epoch": 0.48,
|
| 85 |
-
"learning_rate": 0.
|
| 86 |
-
"loss": 1.
|
| 87 |
"step": 130
|
| 88 |
},
|
| 89 |
{
|
| 90 |
"epoch": 0.52,
|
| 91 |
-
"learning_rate": 0.
|
| 92 |
-
"loss": 1.
|
| 93 |
"step": 140
|
| 94 |
},
|
| 95 |
{
|
| 96 |
"epoch": 0.56,
|
| 97 |
-
"learning_rate": 0.
|
| 98 |
-
"loss": 1.
|
| 99 |
"step": 150
|
| 100 |
},
|
| 101 |
{
|
| 102 |
"epoch": 0.59,
|
| 103 |
-
"learning_rate": 0.
|
| 104 |
-
"loss": 1.
|
| 105 |
"step": 160
|
| 106 |
},
|
| 107 |
{
|
| 108 |
"epoch": 0.63,
|
| 109 |
-
"learning_rate": 0.
|
| 110 |
-
"loss": 1.
|
| 111 |
"step": 170
|
| 112 |
},
|
| 113 |
{
|
| 114 |
"epoch": 0.67,
|
| 115 |
-
"learning_rate": 0.
|
| 116 |
-
"loss":
|
| 117 |
"step": 180
|
| 118 |
},
|
| 119 |
{
|
| 120 |
"epoch": 0.7,
|
| 121 |
-
"learning_rate": 0.
|
| 122 |
-
"loss":
|
| 123 |
"step": 190
|
| 124 |
},
|
| 125 |
{
|
| 126 |
"epoch": 0.74,
|
| 127 |
-
"learning_rate": 0.
|
| 128 |
-
"loss": 0.
|
| 129 |
"step": 200
|
| 130 |
},
|
| 131 |
{
|
| 132 |
"epoch": 0.78,
|
| 133 |
-
"learning_rate": 0.
|
| 134 |
-
"loss": 0.
|
| 135 |
"step": 210
|
| 136 |
},
|
| 137 |
{
|
| 138 |
"epoch": 0.81,
|
| 139 |
-
"learning_rate": 0.
|
| 140 |
-
"loss":
|
| 141 |
"step": 220
|
| 142 |
},
|
| 143 |
{
|
| 144 |
"epoch": 0.85,
|
| 145 |
-
"learning_rate": 0.
|
| 146 |
-
"loss": 0.
|
| 147 |
"step": 230
|
| 148 |
},
|
| 149 |
{
|
| 150 |
"epoch": 0.89,
|
| 151 |
-
"learning_rate": 0.
|
| 152 |
-
"loss":
|
| 153 |
"step": 240
|
| 154 |
},
|
| 155 |
{
|
| 156 |
"epoch": 0.93,
|
| 157 |
-
"learning_rate": 0.
|
| 158 |
-
"loss": 0.
|
| 159 |
"step": 250
|
| 160 |
},
|
| 161 |
{
|
| 162 |
"epoch": 0.96,
|
| 163 |
-
"learning_rate": 0.
|
| 164 |
-
"loss": 0.
|
| 165 |
"step": 260
|
| 166 |
},
|
| 167 |
{
|
| 168 |
"epoch": 1.0,
|
| 169 |
-
"learning_rate": 0.
|
| 170 |
-
"loss": 1.
|
| 171 |
"step": 270
|
| 172 |
},
|
| 173 |
{
|
| 174 |
"epoch": 1.04,
|
| 175 |
-
"learning_rate": 0.
|
| 176 |
-
"loss": 0.
|
| 177 |
"step": 280
|
| 178 |
},
|
| 179 |
{
|
| 180 |
"epoch": 1.07,
|
| 181 |
-
"learning_rate": 0.
|
| 182 |
-
"loss": 0.
|
| 183 |
"step": 290
|
| 184 |
},
|
| 185 |
{
|
| 186 |
"epoch": 1.11,
|
| 187 |
-
"learning_rate": 0.
|
| 188 |
-
"loss": 0.
|
| 189 |
"step": 300
|
| 190 |
},
|
| 191 |
{
|
| 192 |
"epoch": 1.15,
|
| 193 |
-
"learning_rate": 0.
|
| 194 |
-
"loss": 0.
|
| 195 |
"step": 310
|
| 196 |
},
|
| 197 |
{
|
| 198 |
"epoch": 1.19,
|
| 199 |
-
"learning_rate": 0.
|
| 200 |
-
"loss": 0.
|
| 201 |
"step": 320
|
| 202 |
},
|
| 203 |
{
|
| 204 |
"epoch": 1.22,
|
| 205 |
-
"learning_rate": 0.
|
| 206 |
-
"loss": 0.
|
| 207 |
"step": 330
|
| 208 |
},
|
| 209 |
{
|
| 210 |
"epoch": 1.26,
|
| 211 |
-
"learning_rate": 0.
|
| 212 |
-
"loss": 0.
|
| 213 |
"step": 340
|
| 214 |
},
|
| 215 |
{
|
| 216 |
"epoch": 1.3,
|
| 217 |
-
"learning_rate": 0.
|
| 218 |
-
"loss": 0.
|
| 219 |
"step": 350
|
| 220 |
},
|
| 221 |
{
|
| 222 |
"epoch": 1.33,
|
| 223 |
-
"learning_rate": 0.
|
| 224 |
-
"loss": 0.
|
| 225 |
"step": 360
|
| 226 |
},
|
| 227 |
{
|
| 228 |
"epoch": 1.37,
|
| 229 |
-
"learning_rate": 0.
|
| 230 |
-
"loss": 0.
|
| 231 |
"step": 370
|
| 232 |
},
|
| 233 |
{
|
| 234 |
"epoch": 1.41,
|
| 235 |
-
"learning_rate": 0.
|
| 236 |
-
"loss": 0.
|
| 237 |
"step": 380
|
| 238 |
},
|
| 239 |
{
|
| 240 |
"epoch": 1.44,
|
| 241 |
-
"learning_rate": 0.
|
| 242 |
-
"loss": 0.
|
| 243 |
"step": 390
|
| 244 |
},
|
| 245 |
{
|
| 246 |
"epoch": 1.48,
|
| 247 |
-
"learning_rate": 0.
|
| 248 |
-
"loss": 0.
|
| 249 |
"step": 400
|
| 250 |
},
|
| 251 |
{
|
| 252 |
"epoch": 1.52,
|
| 253 |
-
"learning_rate": 0.
|
| 254 |
-
"loss": 0.
|
| 255 |
"step": 410
|
| 256 |
},
|
| 257 |
{
|
| 258 |
"epoch": 1.56,
|
| 259 |
-
"learning_rate": 0.
|
| 260 |
-
"loss": 0.
|
| 261 |
"step": 420
|
| 262 |
},
|
| 263 |
{
|
| 264 |
"epoch": 1.59,
|
| 265 |
-
"learning_rate": 0.
|
| 266 |
-
"loss": 0.
|
| 267 |
"step": 430
|
| 268 |
},
|
| 269 |
{
|
| 270 |
"epoch": 1.63,
|
| 271 |
-
"learning_rate": 0.
|
| 272 |
-
"loss": 0.
|
| 273 |
"step": 440
|
| 274 |
},
|
| 275 |
{
|
| 276 |
"epoch": 1.67,
|
| 277 |
-
"learning_rate": 0.
|
| 278 |
-
"loss": 0.
|
| 279 |
"step": 450
|
| 280 |
},
|
| 281 |
{
|
| 282 |
"epoch": 1.7,
|
| 283 |
-
"learning_rate": 0.
|
| 284 |
-
"loss": 0.
|
| 285 |
"step": 460
|
| 286 |
},
|
| 287 |
{
|
| 288 |
"epoch": 1.74,
|
| 289 |
-
"learning_rate": 0.
|
| 290 |
-
"loss": 0.
|
| 291 |
"step": 470
|
| 292 |
},
|
| 293 |
{
|
| 294 |
"epoch": 1.78,
|
| 295 |
-
"learning_rate": 0.
|
| 296 |
-
"loss": 0.
|
| 297 |
"step": 480
|
| 298 |
},
|
| 299 |
{
|
| 300 |
"epoch": 1.81,
|
| 301 |
-
"learning_rate": 0.
|
| 302 |
-
"loss": 0.
|
| 303 |
"step": 490
|
| 304 |
},
|
| 305 |
{
|
| 306 |
"epoch": 1.85,
|
| 307 |
-
"learning_rate": 0.
|
| 308 |
-
"loss": 0.
|
| 309 |
"step": 500
|
| 310 |
},
|
| 311 |
{
|
| 312 |
"epoch": 1.89,
|
| 313 |
-
"learning_rate": 0.
|
| 314 |
-
"loss": 0.
|
| 315 |
"step": 510
|
| 316 |
},
|
| 317 |
{
|
| 318 |
"epoch": 1.93,
|
| 319 |
-
"learning_rate": 0.
|
| 320 |
-
"loss": 0.
|
| 321 |
"step": 520
|
| 322 |
},
|
| 323 |
{
|
| 324 |
"epoch": 1.96,
|
| 325 |
-
"learning_rate": 0.
|
| 326 |
-
"loss": 0.
|
| 327 |
"step": 530
|
| 328 |
},
|
| 329 |
{
|
| 330 |
"epoch": 2.0,
|
| 331 |
-
"learning_rate": 0.
|
| 332 |
-
"loss": 0.
|
| 333 |
"step": 540
|
| 334 |
},
|
| 335 |
{
|
| 336 |
"epoch": 2.04,
|
| 337 |
-
"learning_rate":
|
| 338 |
-
"loss": 0.
|
| 339 |
"step": 550
|
| 340 |
},
|
| 341 |
{
|
| 342 |
"epoch": 2.07,
|
| 343 |
-
"learning_rate":
|
| 344 |
-
"loss": 0.
|
| 345 |
"step": 560
|
| 346 |
},
|
| 347 |
{
|
| 348 |
"epoch": 2.11,
|
| 349 |
-
"learning_rate":
|
| 350 |
-
"loss": 0.
|
| 351 |
"step": 570
|
| 352 |
},
|
| 353 |
{
|
| 354 |
"epoch": 2.15,
|
| 355 |
-
"learning_rate":
|
| 356 |
-
"loss": 0.
|
| 357 |
"step": 580
|
| 358 |
},
|
| 359 |
{
|
| 360 |
"epoch": 2.19,
|
| 361 |
-
"learning_rate":
|
| 362 |
-
"loss": 0.
|
| 363 |
"step": 590
|
| 364 |
},
|
| 365 |
{
|
| 366 |
"epoch": 2.22,
|
| 367 |
-
"learning_rate":
|
| 368 |
-
"loss": 0.
|
| 369 |
"step": 600
|
| 370 |
},
|
| 371 |
{
|
| 372 |
"epoch": 2.26,
|
| 373 |
-
"learning_rate":
|
| 374 |
-
"loss": 0.
|
| 375 |
"step": 610
|
| 376 |
},
|
| 377 |
{
|
| 378 |
"epoch": 2.3,
|
| 379 |
-
"learning_rate":
|
| 380 |
-
"loss": 0.
|
| 381 |
"step": 620
|
| 382 |
},
|
| 383 |
{
|
| 384 |
"epoch": 2.33,
|
| 385 |
-
"learning_rate":
|
| 386 |
-
"loss": 0.
|
| 387 |
"step": 630
|
| 388 |
},
|
| 389 |
{
|
| 390 |
"epoch": 2.37,
|
| 391 |
-
"learning_rate":
|
| 392 |
-
"loss": 0.
|
| 393 |
"step": 640
|
| 394 |
},
|
| 395 |
{
|
| 396 |
"epoch": 2.41,
|
| 397 |
-
"learning_rate":
|
| 398 |
-
"loss": 0.
|
| 399 |
"step": 650
|
| 400 |
},
|
| 401 |
{
|
| 402 |
"epoch": 2.44,
|
| 403 |
-
"learning_rate":
|
| 404 |
-
"loss": 0.
|
| 405 |
"step": 660
|
| 406 |
},
|
| 407 |
{
|
| 408 |
"epoch": 2.48,
|
| 409 |
-
"learning_rate":
|
| 410 |
-
"loss": 0.
|
| 411 |
"step": 670
|
| 412 |
},
|
| 413 |
{
|
| 414 |
"epoch": 2.52,
|
| 415 |
-
"learning_rate":
|
| 416 |
-
"loss": 0.
|
| 417 |
"step": 680
|
| 418 |
},
|
| 419 |
{
|
| 420 |
"epoch": 2.56,
|
| 421 |
-
"learning_rate":
|
| 422 |
-
"loss": 0.
|
| 423 |
"step": 690
|
| 424 |
},
|
| 425 |
{
|
| 426 |
"epoch": 2.59,
|
| 427 |
-
"learning_rate":
|
| 428 |
-
"loss": 0.
|
| 429 |
"step": 700
|
| 430 |
},
|
| 431 |
{
|
| 432 |
"epoch": 2.63,
|
| 433 |
-
"learning_rate":
|
| 434 |
-
"loss": 0.
|
| 435 |
"step": 710
|
| 436 |
},
|
| 437 |
{
|
| 438 |
"epoch": 2.67,
|
| 439 |
-
"learning_rate":
|
| 440 |
-
"loss": 0.
|
| 441 |
"step": 720
|
| 442 |
},
|
| 443 |
{
|
| 444 |
"epoch": 2.7,
|
| 445 |
-
"learning_rate":
|
| 446 |
-
"loss": 0.
|
| 447 |
"step": 730
|
| 448 |
},
|
| 449 |
{
|
| 450 |
"epoch": 2.74,
|
| 451 |
-
"learning_rate":
|
| 452 |
-
"loss": 0.
|
| 453 |
"step": 740
|
| 454 |
},
|
| 455 |
{
|
| 456 |
"epoch": 2.78,
|
| 457 |
-
"learning_rate":
|
| 458 |
-
"loss": 0.
|
| 459 |
"step": 750
|
| 460 |
},
|
| 461 |
{
|
| 462 |
"epoch": 2.81,
|
| 463 |
-
"learning_rate":
|
| 464 |
-
"loss": 0.
|
| 465 |
"step": 760
|
| 466 |
},
|
| 467 |
{
|
| 468 |
"epoch": 2.85,
|
| 469 |
-
"learning_rate":
|
| 470 |
-
"loss": 0.
|
| 471 |
"step": 770
|
| 472 |
},
|
| 473 |
{
|
| 474 |
"epoch": 2.89,
|
| 475 |
-
"learning_rate":
|
| 476 |
-
"loss": 0.
|
| 477 |
"step": 780
|
| 478 |
},
|
| 479 |
{
|
| 480 |
"epoch": 2.93,
|
| 481 |
-
"learning_rate":
|
| 482 |
-
"loss": 0.
|
| 483 |
"step": 790
|
| 484 |
},
|
| 485 |
{
|
| 486 |
"epoch": 2.96,
|
| 487 |
-
"learning_rate":
|
| 488 |
-
"loss": 0.
|
| 489 |
"step": 800
|
| 490 |
},
|
| 491 |
{
|
| 492 |
"epoch": 3.0,
|
| 493 |
-
"learning_rate":
|
| 494 |
-
"loss": 0.
|
| 495 |
"step": 810
|
| 496 |
},
|
| 497 |
{
|
| 498 |
"epoch": 3.04,
|
| 499 |
-
"learning_rate":
|
| 500 |
-
"loss": 0.
|
| 501 |
"step": 820
|
| 502 |
},
|
| 503 |
{
|
| 504 |
"epoch": 3.07,
|
| 505 |
-
"learning_rate":
|
| 506 |
-
"loss": 0.
|
| 507 |
"step": 830
|
| 508 |
},
|
| 509 |
{
|
| 510 |
"epoch": 3.11,
|
| 511 |
-
"learning_rate":
|
| 512 |
-
"loss": 0.
|
| 513 |
"step": 840
|
| 514 |
},
|
| 515 |
{
|
| 516 |
"epoch": 3.15,
|
| 517 |
-
"learning_rate":
|
| 518 |
-
"loss": 0.
|
| 519 |
"step": 850
|
| 520 |
},
|
| 521 |
{
|
| 522 |
"epoch": 3.19,
|
| 523 |
-
"learning_rate":
|
| 524 |
-
"loss": 0.
|
| 525 |
"step": 860
|
| 526 |
},
|
| 527 |
{
|
| 528 |
"epoch": 3.22,
|
| 529 |
-
"learning_rate":
|
| 530 |
-
"loss": 0.
|
| 531 |
"step": 870
|
| 532 |
},
|
| 533 |
{
|
| 534 |
"epoch": 3.26,
|
| 535 |
-
"learning_rate":
|
| 536 |
-
"loss": 0.
|
| 537 |
"step": 880
|
| 538 |
},
|
| 539 |
{
|
| 540 |
"epoch": 3.3,
|
| 541 |
-
"learning_rate":
|
| 542 |
-
"loss": 0.
|
| 543 |
"step": 890
|
| 544 |
},
|
| 545 |
{
|
| 546 |
"epoch": 3.33,
|
| 547 |
-
"learning_rate":
|
| 548 |
-
"loss": 0.
|
| 549 |
"step": 900
|
| 550 |
},
|
| 551 |
{
|
| 552 |
"epoch": 3.37,
|
| 553 |
-
"learning_rate":
|
| 554 |
-
"loss": 0.
|
| 555 |
"step": 910
|
| 556 |
},
|
| 557 |
{
|
| 558 |
"epoch": 3.41,
|
| 559 |
-
"learning_rate":
|
| 560 |
-
"loss": 0.
|
| 561 |
"step": 920
|
| 562 |
},
|
| 563 |
{
|
| 564 |
"epoch": 3.44,
|
| 565 |
-
"learning_rate":
|
| 566 |
-
"loss": 0.
|
| 567 |
"step": 930
|
| 568 |
},
|
| 569 |
{
|
| 570 |
"epoch": 3.48,
|
| 571 |
-
"learning_rate":
|
| 572 |
-
"loss": 0.
|
| 573 |
"step": 940
|
| 574 |
},
|
| 575 |
{
|
| 576 |
"epoch": 3.52,
|
| 577 |
-
"learning_rate":
|
| 578 |
-
"loss": 0.
|
| 579 |
"step": 950
|
| 580 |
},
|
| 581 |
{
|
| 582 |
"epoch": 3.56,
|
| 583 |
-
"learning_rate":
|
| 584 |
-
"loss": 0.
|
| 585 |
"step": 960
|
| 586 |
},
|
| 587 |
{
|
| 588 |
"epoch": 3.59,
|
| 589 |
-
"learning_rate":
|
| 590 |
-
"loss": 0.
|
| 591 |
"step": 970
|
| 592 |
},
|
| 593 |
{
|
| 594 |
"epoch": 3.63,
|
| 595 |
-
"learning_rate":
|
| 596 |
-
"loss": 0.
|
| 597 |
"step": 980
|
| 598 |
},
|
| 599 |
{
|
| 600 |
"epoch": 3.67,
|
| 601 |
-
"learning_rate":
|
| 602 |
-
"loss": 0.
|
| 603 |
"step": 990
|
| 604 |
},
|
| 605 |
{
|
| 606 |
"epoch": 3.7,
|
| 607 |
-
"learning_rate":
|
| 608 |
-
"loss": 0.
|
| 609 |
"step": 1000
|
| 610 |
},
|
| 611 |
{
|
| 612 |
"epoch": 3.74,
|
| 613 |
-
"learning_rate":
|
| 614 |
-
"loss": 0.
|
| 615 |
"step": 1010
|
| 616 |
},
|
| 617 |
{
|
| 618 |
"epoch": 3.78,
|
| 619 |
-
"learning_rate":
|
| 620 |
-
"loss": 0.
|
| 621 |
"step": 1020
|
| 622 |
},
|
| 623 |
{
|
| 624 |
"epoch": 3.81,
|
| 625 |
-
"learning_rate":
|
| 626 |
-
"loss": 0.
|
| 627 |
"step": 1030
|
| 628 |
},
|
| 629 |
{
|
| 630 |
"epoch": 3.85,
|
| 631 |
-
"learning_rate":
|
| 632 |
-
"loss": 0.
|
| 633 |
"step": 1040
|
| 634 |
},
|
| 635 |
{
|
| 636 |
"epoch": 3.89,
|
| 637 |
-
"learning_rate":
|
| 638 |
-
"loss": 0.
|
| 639 |
"step": 1050
|
| 640 |
},
|
| 641 |
{
|
| 642 |
"epoch": 3.93,
|
| 643 |
-
"learning_rate":
|
| 644 |
-
"loss": 0.
|
| 645 |
"step": 1060
|
| 646 |
},
|
| 647 |
{
|
| 648 |
"epoch": 3.96,
|
| 649 |
-
"learning_rate":
|
| 650 |
-
"loss": 0.
|
| 651 |
"step": 1070
|
| 652 |
},
|
| 653 |
{
|
| 654 |
"epoch": 4.0,
|
| 655 |
-
"learning_rate": 0.
|
| 656 |
-
"loss": 0.
|
| 657 |
"step": 1080
|
| 658 |
},
|
| 659 |
{
|
| 660 |
-
"epoch": 4.
|
| 661 |
-
"
|
| 662 |
-
"
|
| 663 |
-
"
|
| 664 |
-
|
| 665 |
-
|
| 666 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 667 |
}
|
| 668 |
],
|
| 669 |
"logging_steps": 10,
|
| 670 |
-
"max_steps":
|
| 671 |
-
"num_train_epochs":
|
| 672 |
"save_steps": 3000,
|
| 673 |
-
"total_flos":
|
| 674 |
"trial_name": null,
|
| 675 |
"trial_params": null
|
| 676 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 10.0,
|
| 5 |
"eval_steps": 3000,
|
| 6 |
+
"global_step": 2700,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
"epoch": 0.04,
|
| 13 |
+
"learning_rate": 0.00019925925925925927,
|
| 14 |
+
"loss": 2.1757,
|
| 15 |
"step": 10
|
| 16 |
},
|
| 17 |
{
|
| 18 |
"epoch": 0.07,
|
| 19 |
+
"learning_rate": 0.00019851851851851853,
|
| 20 |
+
"loss": 1.9027,
|
| 21 |
"step": 20
|
| 22 |
},
|
| 23 |
{
|
| 24 |
"epoch": 0.11,
|
| 25 |
+
"learning_rate": 0.00019777777777777778,
|
| 26 |
+
"loss": 1.7624,
|
| 27 |
"step": 30
|
| 28 |
},
|
| 29 |
{
|
| 30 |
"epoch": 0.15,
|
| 31 |
+
"learning_rate": 0.00019703703703703704,
|
| 32 |
+
"loss": 1.6008,
|
| 33 |
"step": 40
|
| 34 |
},
|
| 35 |
{
|
| 36 |
"epoch": 0.19,
|
| 37 |
+
"learning_rate": 0.0001962962962962963,
|
| 38 |
+
"loss": 1.3749,
|
| 39 |
"step": 50
|
| 40 |
},
|
| 41 |
{
|
| 42 |
"epoch": 0.22,
|
| 43 |
+
"learning_rate": 0.00019555555555555556,
|
| 44 |
+
"loss": 1.494,
|
| 45 |
"step": 60
|
| 46 |
},
|
| 47 |
{
|
| 48 |
"epoch": 0.26,
|
| 49 |
+
"learning_rate": 0.00019481481481481482,
|
| 50 |
+
"loss": 1.3449,
|
| 51 |
"step": 70
|
| 52 |
},
|
| 53 |
{
|
| 54 |
"epoch": 0.3,
|
| 55 |
+
"learning_rate": 0.00019407407407407408,
|
| 56 |
+
"loss": 1.2193,
|
| 57 |
"step": 80
|
| 58 |
},
|
| 59 |
{
|
| 60 |
"epoch": 0.33,
|
| 61 |
+
"learning_rate": 0.00019333333333333333,
|
| 62 |
+
"loss": 1.3163,
|
| 63 |
"step": 90
|
| 64 |
},
|
| 65 |
{
|
| 66 |
"epoch": 0.37,
|
| 67 |
+
"learning_rate": 0.0001925925925925926,
|
| 68 |
+
"loss": 1.1955,
|
| 69 |
"step": 100
|
| 70 |
},
|
| 71 |
{
|
| 72 |
"epoch": 0.41,
|
| 73 |
+
"learning_rate": 0.00019185185185185185,
|
| 74 |
+
"loss": 1.3643,
|
| 75 |
"step": 110
|
| 76 |
},
|
| 77 |
{
|
| 78 |
"epoch": 0.44,
|
| 79 |
+
"learning_rate": 0.00019111111111111114,
|
| 80 |
+
"loss": 1.3792,
|
| 81 |
"step": 120
|
| 82 |
},
|
| 83 |
{
|
| 84 |
"epoch": 0.48,
|
| 85 |
+
"learning_rate": 0.00019037037037037037,
|
| 86 |
+
"loss": 1.1842,
|
| 87 |
"step": 130
|
| 88 |
},
|
| 89 |
{
|
| 90 |
"epoch": 0.52,
|
| 91 |
+
"learning_rate": 0.00018962962962962965,
|
| 92 |
+
"loss": 1.1366,
|
| 93 |
"step": 140
|
| 94 |
},
|
| 95 |
{
|
| 96 |
"epoch": 0.56,
|
| 97 |
+
"learning_rate": 0.00018888888888888888,
|
| 98 |
+
"loss": 1.2302,
|
| 99 |
"step": 150
|
| 100 |
},
|
| 101 |
{
|
| 102 |
"epoch": 0.59,
|
| 103 |
+
"learning_rate": 0.00018814814814814817,
|
| 104 |
+
"loss": 1.1672,
|
| 105 |
"step": 160
|
| 106 |
},
|
| 107 |
{
|
| 108 |
"epoch": 0.63,
|
| 109 |
+
"learning_rate": 0.0001874074074074074,
|
| 110 |
+
"loss": 1.1124,
|
| 111 |
"step": 170
|
| 112 |
},
|
| 113 |
{
|
| 114 |
"epoch": 0.67,
|
| 115 |
+
"learning_rate": 0.0001866666666666667,
|
| 116 |
+
"loss": 0.9382,
|
| 117 |
"step": 180
|
| 118 |
},
|
| 119 |
{
|
| 120 |
"epoch": 0.7,
|
| 121 |
+
"learning_rate": 0.00018592592592592592,
|
| 122 |
+
"loss": 0.9567,
|
| 123 |
"step": 190
|
| 124 |
},
|
| 125 |
{
|
| 126 |
"epoch": 0.74,
|
| 127 |
+
"learning_rate": 0.0001851851851851852,
|
| 128 |
+
"loss": 0.9075,
|
| 129 |
"step": 200
|
| 130 |
},
|
| 131 |
{
|
| 132 |
"epoch": 0.78,
|
| 133 |
+
"learning_rate": 0.00018444444444444446,
|
| 134 |
+
"loss": 0.9064,
|
| 135 |
"step": 210
|
| 136 |
},
|
| 137 |
{
|
| 138 |
"epoch": 0.81,
|
| 139 |
+
"learning_rate": 0.00018370370370370372,
|
| 140 |
+
"loss": 1.0113,
|
| 141 |
"step": 220
|
| 142 |
},
|
| 143 |
{
|
| 144 |
"epoch": 0.85,
|
| 145 |
+
"learning_rate": 0.00018296296296296298,
|
| 146 |
+
"loss": 0.863,
|
| 147 |
"step": 230
|
| 148 |
},
|
| 149 |
{
|
| 150 |
"epoch": 0.89,
|
| 151 |
+
"learning_rate": 0.00018222222222222224,
|
| 152 |
+
"loss": 0.83,
|
| 153 |
"step": 240
|
| 154 |
},
|
| 155 |
{
|
| 156 |
"epoch": 0.93,
|
| 157 |
+
"learning_rate": 0.0001814814814814815,
|
| 158 |
+
"loss": 0.9212,
|
| 159 |
"step": 250
|
| 160 |
},
|
| 161 |
{
|
| 162 |
"epoch": 0.96,
|
| 163 |
+
"learning_rate": 0.00018074074074074075,
|
| 164 |
+
"loss": 0.8929,
|
| 165 |
"step": 260
|
| 166 |
},
|
| 167 |
{
|
| 168 |
"epoch": 1.0,
|
| 169 |
+
"learning_rate": 0.00018,
|
| 170 |
+
"loss": 1.09,
|
| 171 |
"step": 270
|
| 172 |
},
|
| 173 |
{
|
| 174 |
"epoch": 1.04,
|
| 175 |
+
"learning_rate": 0.00017925925925925927,
|
| 176 |
+
"loss": 0.7487,
|
| 177 |
"step": 280
|
| 178 |
},
|
| 179 |
{
|
| 180 |
"epoch": 1.07,
|
| 181 |
+
"learning_rate": 0.00017851851851851853,
|
| 182 |
+
"loss": 0.5648,
|
| 183 |
"step": 290
|
| 184 |
},
|
| 185 |
{
|
| 186 |
"epoch": 1.11,
|
| 187 |
+
"learning_rate": 0.00017777777777777779,
|
| 188 |
+
"loss": 0.8104,
|
| 189 |
"step": 300
|
| 190 |
},
|
| 191 |
{
|
| 192 |
"epoch": 1.15,
|
| 193 |
+
"learning_rate": 0.00017703703703703704,
|
| 194 |
+
"loss": 0.8599,
|
| 195 |
"step": 310
|
| 196 |
},
|
| 197 |
{
|
| 198 |
"epoch": 1.19,
|
| 199 |
+
"learning_rate": 0.0001762962962962963,
|
| 200 |
+
"loss": 0.6207,
|
| 201 |
"step": 320
|
| 202 |
},
|
| 203 |
{
|
| 204 |
"epoch": 1.22,
|
| 205 |
+
"learning_rate": 0.00017555555555555556,
|
| 206 |
+
"loss": 0.6954,
|
| 207 |
"step": 330
|
| 208 |
},
|
| 209 |
{
|
| 210 |
"epoch": 1.26,
|
| 211 |
+
"learning_rate": 0.00017481481481481482,
|
| 212 |
+
"loss": 0.6896,
|
| 213 |
"step": 340
|
| 214 |
},
|
| 215 |
{
|
| 216 |
"epoch": 1.3,
|
| 217 |
+
"learning_rate": 0.00017407407407407408,
|
| 218 |
+
"loss": 0.6426,
|
| 219 |
"step": 350
|
| 220 |
},
|
| 221 |
{
|
| 222 |
"epoch": 1.33,
|
| 223 |
+
"learning_rate": 0.00017333333333333334,
|
| 224 |
+
"loss": 0.6755,
|
| 225 |
"step": 360
|
| 226 |
},
|
| 227 |
{
|
| 228 |
"epoch": 1.37,
|
| 229 |
+
"learning_rate": 0.0001725925925925926,
|
| 230 |
+
"loss": 0.5689,
|
| 231 |
"step": 370
|
| 232 |
},
|
| 233 |
{
|
| 234 |
"epoch": 1.41,
|
| 235 |
+
"learning_rate": 0.00017185185185185185,
|
| 236 |
+
"loss": 0.8554,
|
| 237 |
"step": 380
|
| 238 |
},
|
| 239 |
{
|
| 240 |
"epoch": 1.44,
|
| 241 |
+
"learning_rate": 0.0001711111111111111,
|
| 242 |
+
"loss": 0.5816,
|
| 243 |
"step": 390
|
| 244 |
},
|
| 245 |
{
|
| 246 |
"epoch": 1.48,
|
| 247 |
+
"learning_rate": 0.00017037037037037037,
|
| 248 |
+
"loss": 0.71,
|
| 249 |
"step": 400
|
| 250 |
},
|
| 251 |
{
|
| 252 |
"epoch": 1.52,
|
| 253 |
+
"learning_rate": 0.00016962962962962963,
|
| 254 |
+
"loss": 0.6038,
|
| 255 |
"step": 410
|
| 256 |
},
|
| 257 |
{
|
| 258 |
"epoch": 1.56,
|
| 259 |
+
"learning_rate": 0.00016888888888888889,
|
| 260 |
+
"loss": 0.5442,
|
| 261 |
"step": 420
|
| 262 |
},
|
| 263 |
{
|
| 264 |
"epoch": 1.59,
|
| 265 |
+
"learning_rate": 0.00016814814814814817,
|
| 266 |
+
"loss": 0.6354,
|
| 267 |
"step": 430
|
| 268 |
},
|
| 269 |
{
|
| 270 |
"epoch": 1.63,
|
| 271 |
+
"learning_rate": 0.0001674074074074074,
|
| 272 |
+
"loss": 0.8464,
|
| 273 |
"step": 440
|
| 274 |
},
|
| 275 |
{
|
| 276 |
"epoch": 1.67,
|
| 277 |
+
"learning_rate": 0.0001666666666666667,
|
| 278 |
+
"loss": 0.8248,
|
| 279 |
"step": 450
|
| 280 |
},
|
| 281 |
{
|
| 282 |
"epoch": 1.7,
|
| 283 |
+
"learning_rate": 0.00016592592592592592,
|
| 284 |
+
"loss": 0.6434,
|
| 285 |
"step": 460
|
| 286 |
},
|
| 287 |
{
|
| 288 |
"epoch": 1.74,
|
| 289 |
+
"learning_rate": 0.0001651851851851852,
|
| 290 |
+
"loss": 0.6549,
|
| 291 |
"step": 470
|
| 292 |
},
|
| 293 |
{
|
| 294 |
"epoch": 1.78,
|
| 295 |
+
"learning_rate": 0.00016444444444444444,
|
| 296 |
+
"loss": 0.5651,
|
| 297 |
"step": 480
|
| 298 |
},
|
| 299 |
{
|
| 300 |
"epoch": 1.81,
|
| 301 |
+
"learning_rate": 0.00016370370370370372,
|
| 302 |
+
"loss": 0.4748,
|
| 303 |
"step": 490
|
| 304 |
},
|
| 305 |
{
|
| 306 |
"epoch": 1.85,
|
| 307 |
+
"learning_rate": 0.00016296296296296295,
|
| 308 |
+
"loss": 0.6189,
|
| 309 |
"step": 500
|
| 310 |
},
|
| 311 |
{
|
| 312 |
"epoch": 1.89,
|
| 313 |
+
"learning_rate": 0.00016222222222222224,
|
| 314 |
+
"loss": 0.4894,
|
| 315 |
"step": 510
|
| 316 |
},
|
| 317 |
{
|
| 318 |
"epoch": 1.93,
|
| 319 |
+
"learning_rate": 0.0001614814814814815,
|
| 320 |
+
"loss": 0.6632,
|
| 321 |
"step": 520
|
| 322 |
},
|
| 323 |
{
|
| 324 |
"epoch": 1.96,
|
| 325 |
+
"learning_rate": 0.00016074074074074075,
|
| 326 |
+
"loss": 0.5738,
|
| 327 |
"step": 530
|
| 328 |
},
|
| 329 |
{
|
| 330 |
"epoch": 2.0,
|
| 331 |
+
"learning_rate": 0.00016,
|
| 332 |
+
"loss": 0.5349,
|
| 333 |
"step": 540
|
| 334 |
},
|
| 335 |
{
|
| 336 |
"epoch": 2.04,
|
| 337 |
+
"learning_rate": 0.00015925925925925927,
|
| 338 |
+
"loss": 0.3067,
|
| 339 |
"step": 550
|
| 340 |
},
|
| 341 |
{
|
| 342 |
"epoch": 2.07,
|
| 343 |
+
"learning_rate": 0.00015851851851851853,
|
| 344 |
+
"loss": 0.3444,
|
| 345 |
"step": 560
|
| 346 |
},
|
| 347 |
{
|
| 348 |
"epoch": 2.11,
|
| 349 |
+
"learning_rate": 0.0001577777777777778,
|
| 350 |
+
"loss": 0.3304,
|
| 351 |
"step": 570
|
| 352 |
},
|
| 353 |
{
|
| 354 |
"epoch": 2.15,
|
| 355 |
+
"learning_rate": 0.00015703703703703705,
|
| 356 |
+
"loss": 0.5832,
|
| 357 |
"step": 580
|
| 358 |
},
|
| 359 |
{
|
| 360 |
"epoch": 2.19,
|
| 361 |
+
"learning_rate": 0.0001562962962962963,
|
| 362 |
+
"loss": 0.5208,
|
| 363 |
"step": 590
|
| 364 |
},
|
| 365 |
{
|
| 366 |
"epoch": 2.22,
|
| 367 |
+
"learning_rate": 0.00015555555555555556,
|
| 368 |
+
"loss": 0.3367,
|
| 369 |
"step": 600
|
| 370 |
},
|
| 371 |
{
|
| 372 |
"epoch": 2.26,
|
| 373 |
+
"learning_rate": 0.00015481481481481482,
|
| 374 |
+
"loss": 0.3564,
|
| 375 |
"step": 610
|
| 376 |
},
|
| 377 |
{
|
| 378 |
"epoch": 2.3,
|
| 379 |
+
"learning_rate": 0.00015407407407407408,
|
| 380 |
+
"loss": 0.5481,
|
| 381 |
"step": 620
|
| 382 |
},
|
| 383 |
{
|
| 384 |
"epoch": 2.33,
|
| 385 |
+
"learning_rate": 0.00015333333333333334,
|
| 386 |
+
"loss": 0.3271,
|
| 387 |
"step": 630
|
| 388 |
},
|
| 389 |
{
|
| 390 |
"epoch": 2.37,
|
| 391 |
+
"learning_rate": 0.0001525925925925926,
|
| 392 |
+
"loss": 0.3379,
|
| 393 |
"step": 640
|
| 394 |
},
|
| 395 |
{
|
| 396 |
"epoch": 2.41,
|
| 397 |
+
"learning_rate": 0.00015185185185185185,
|
| 398 |
+
"loss": 0.4362,
|
| 399 |
"step": 650
|
| 400 |
},
|
| 401 |
{
|
| 402 |
"epoch": 2.44,
|
| 403 |
+
"learning_rate": 0.0001511111111111111,
|
| 404 |
+
"loss": 0.4654,
|
| 405 |
"step": 660
|
| 406 |
},
|
| 407 |
{
|
| 408 |
"epoch": 2.48,
|
| 409 |
+
"learning_rate": 0.00015037037037037037,
|
| 410 |
+
"loss": 0.4666,
|
| 411 |
"step": 670
|
| 412 |
},
|
| 413 |
{
|
| 414 |
"epoch": 2.52,
|
| 415 |
+
"learning_rate": 0.00014962962962962963,
|
| 416 |
+
"loss": 0.2851,
|
| 417 |
"step": 680
|
| 418 |
},
|
| 419 |
{
|
| 420 |
"epoch": 2.56,
|
| 421 |
+
"learning_rate": 0.0001488888888888889,
|
| 422 |
+
"loss": 0.3229,
|
| 423 |
"step": 690
|
| 424 |
},
|
| 425 |
{
|
| 426 |
"epoch": 2.59,
|
| 427 |
+
"learning_rate": 0.00014814814814814815,
|
| 428 |
+
"loss": 0.4301,
|
| 429 |
"step": 700
|
| 430 |
},
|
| 431 |
{
|
| 432 |
"epoch": 2.63,
|
| 433 |
+
"learning_rate": 0.0001474074074074074,
|
| 434 |
+
"loss": 0.22,
|
| 435 |
"step": 710
|
| 436 |
},
|
| 437 |
{
|
| 438 |
"epoch": 2.67,
|
| 439 |
+
"learning_rate": 0.00014666666666666666,
|
| 440 |
+
"loss": 0.3819,
|
| 441 |
"step": 720
|
| 442 |
},
|
| 443 |
{
|
| 444 |
"epoch": 2.7,
|
| 445 |
+
"learning_rate": 0.00014592592592592592,
|
| 446 |
+
"loss": 0.4382,
|
| 447 |
"step": 730
|
| 448 |
},
|
| 449 |
{
|
| 450 |
"epoch": 2.74,
|
| 451 |
+
"learning_rate": 0.0001451851851851852,
|
| 452 |
+
"loss": 0.3389,
|
| 453 |
"step": 740
|
| 454 |
},
|
| 455 |
{
|
| 456 |
"epoch": 2.78,
|
| 457 |
+
"learning_rate": 0.00014444444444444444,
|
| 458 |
+
"loss": 0.392,
|
| 459 |
"step": 750
|
| 460 |
},
|
| 461 |
{
|
| 462 |
"epoch": 2.81,
|
| 463 |
+
"learning_rate": 0.00014370370370370372,
|
| 464 |
+
"loss": 0.5607,
|
| 465 |
"step": 760
|
| 466 |
},
|
| 467 |
{
|
| 468 |
"epoch": 2.85,
|
| 469 |
+
"learning_rate": 0.00014296296296296295,
|
| 470 |
+
"loss": 0.2621,
|
| 471 |
"step": 770
|
| 472 |
},
|
| 473 |
{
|
| 474 |
"epoch": 2.89,
|
| 475 |
+
"learning_rate": 0.00014222222222222224,
|
| 476 |
+
"loss": 0.4369,
|
| 477 |
"step": 780
|
| 478 |
},
|
| 479 |
{
|
| 480 |
"epoch": 2.93,
|
| 481 |
+
"learning_rate": 0.00014148148148148147,
|
| 482 |
+
"loss": 0.3211,
|
| 483 |
"step": 790
|
| 484 |
},
|
| 485 |
{
|
| 486 |
"epoch": 2.96,
|
| 487 |
+
"learning_rate": 0.00014074074074074076,
|
| 488 |
+
"loss": 0.3269,
|
| 489 |
"step": 800
|
| 490 |
},
|
| 491 |
{
|
| 492 |
"epoch": 3.0,
|
| 493 |
+
"learning_rate": 0.00014,
|
| 494 |
+
"loss": 0.2869,
|
| 495 |
"step": 810
|
| 496 |
},
|
| 497 |
{
|
| 498 |
"epoch": 3.04,
|
| 499 |
+
"learning_rate": 0.00013925925925925927,
|
| 500 |
+
"loss": 0.2943,
|
| 501 |
"step": 820
|
| 502 |
},
|
| 503 |
{
|
| 504 |
"epoch": 3.07,
|
| 505 |
+
"learning_rate": 0.0001385185185185185,
|
| 506 |
+
"loss": 0.2445,
|
| 507 |
"step": 830
|
| 508 |
},
|
| 509 |
{
|
| 510 |
"epoch": 3.11,
|
| 511 |
+
"learning_rate": 0.0001377777777777778,
|
| 512 |
+
"loss": 0.1964,
|
| 513 |
"step": 840
|
| 514 |
},
|
| 515 |
{
|
| 516 |
"epoch": 3.15,
|
| 517 |
+
"learning_rate": 0.00013703703703703705,
|
| 518 |
+
"loss": 0.1399,
|
| 519 |
"step": 850
|
| 520 |
},
|
| 521 |
{
|
| 522 |
"epoch": 3.19,
|
| 523 |
+
"learning_rate": 0.0001362962962962963,
|
| 524 |
+
"loss": 0.1535,
|
| 525 |
"step": 860
|
| 526 |
},
|
| 527 |
{
|
| 528 |
"epoch": 3.22,
|
| 529 |
+
"learning_rate": 0.00013555555555555556,
|
| 530 |
+
"loss": 0.2293,
|
| 531 |
"step": 870
|
| 532 |
},
|
| 533 |
{
|
| 534 |
"epoch": 3.26,
|
| 535 |
+
"learning_rate": 0.00013481481481481482,
|
| 536 |
+
"loss": 0.2467,
|
| 537 |
"step": 880
|
| 538 |
},
|
| 539 |
{
|
| 540 |
"epoch": 3.3,
|
| 541 |
+
"learning_rate": 0.00013407407407407408,
|
| 542 |
+
"loss": 0.2009,
|
| 543 |
"step": 890
|
| 544 |
},
|
| 545 |
{
|
| 546 |
"epoch": 3.33,
|
| 547 |
+
"learning_rate": 0.00013333333333333334,
|
| 548 |
+
"loss": 0.2182,
|
| 549 |
"step": 900
|
| 550 |
},
|
| 551 |
{
|
| 552 |
"epoch": 3.37,
|
| 553 |
+
"learning_rate": 0.0001325925925925926,
|
| 554 |
+
"loss": 0.0813,
|
| 555 |
"step": 910
|
| 556 |
},
|
| 557 |
{
|
| 558 |
"epoch": 3.41,
|
| 559 |
+
"learning_rate": 0.00013185185185185186,
|
| 560 |
+
"loss": 0.1104,
|
| 561 |
"step": 920
|
| 562 |
},
|
| 563 |
{
|
| 564 |
"epoch": 3.44,
|
| 565 |
+
"learning_rate": 0.00013111111111111111,
|
| 566 |
+
"loss": 0.2386,
|
| 567 |
"step": 930
|
| 568 |
},
|
| 569 |
{
|
| 570 |
"epoch": 3.48,
|
| 571 |
+
"learning_rate": 0.0001303703703703704,
|
| 572 |
+
"loss": 0.228,
|
| 573 |
"step": 940
|
| 574 |
},
|
| 575 |
{
|
| 576 |
"epoch": 3.52,
|
| 577 |
+
"learning_rate": 0.00012962962962962963,
|
| 578 |
+
"loss": 0.2143,
|
| 579 |
"step": 950
|
| 580 |
},
|
| 581 |
{
|
| 582 |
"epoch": 3.56,
|
| 583 |
+
"learning_rate": 0.00012888888888888892,
|
| 584 |
+
"loss": 0.1438,
|
| 585 |
"step": 960
|
| 586 |
},
|
| 587 |
{
|
| 588 |
"epoch": 3.59,
|
| 589 |
+
"learning_rate": 0.00012814814814814815,
|
| 590 |
+
"loss": 0.1099,
|
| 591 |
"step": 970
|
| 592 |
},
|
| 593 |
{
|
| 594 |
"epoch": 3.63,
|
| 595 |
+
"learning_rate": 0.0001274074074074074,
|
| 596 |
+
"loss": 0.1817,
|
| 597 |
"step": 980
|
| 598 |
},
|
| 599 |
{
|
| 600 |
"epoch": 3.67,
|
| 601 |
+
"learning_rate": 0.00012666666666666666,
|
| 602 |
+
"loss": 0.1819,
|
| 603 |
"step": 990
|
| 604 |
},
|
| 605 |
{
|
| 606 |
"epoch": 3.7,
|
| 607 |
+
"learning_rate": 0.00012592592592592592,
|
| 608 |
+
"loss": 0.0993,
|
| 609 |
"step": 1000
|
| 610 |
},
|
| 611 |
{
|
| 612 |
"epoch": 3.74,
|
| 613 |
+
"learning_rate": 0.00012518518518518518,
|
| 614 |
+
"loss": 0.1726,
|
| 615 |
"step": 1010
|
| 616 |
},
|
| 617 |
{
|
| 618 |
"epoch": 3.78,
|
| 619 |
+
"learning_rate": 0.00012444444444444444,
|
| 620 |
+
"loss": 0.0822,
|
| 621 |
"step": 1020
|
| 622 |
},
|
| 623 |
{
|
| 624 |
"epoch": 3.81,
|
| 625 |
+
"learning_rate": 0.0001237037037037037,
|
| 626 |
+
"loss": 0.158,
|
| 627 |
"step": 1030
|
| 628 |
},
|
| 629 |
{
|
| 630 |
"epoch": 3.85,
|
| 631 |
+
"learning_rate": 0.00012296296296296296,
|
| 632 |
+
"loss": 0.1716,
|
| 633 |
"step": 1040
|
| 634 |
},
|
| 635 |
{
|
| 636 |
"epoch": 3.89,
|
| 637 |
+
"learning_rate": 0.00012222222222222224,
|
| 638 |
+
"loss": 0.1157,
|
| 639 |
"step": 1050
|
| 640 |
},
|
| 641 |
{
|
| 642 |
"epoch": 3.93,
|
| 643 |
+
"learning_rate": 0.00012148148148148149,
|
| 644 |
+
"loss": 0.1985,
|
| 645 |
"step": 1060
|
| 646 |
},
|
| 647 |
{
|
| 648 |
"epoch": 3.96,
|
| 649 |
+
"learning_rate": 0.00012074074074074076,
|
| 650 |
+
"loss": 0.1466,
|
| 651 |
"step": 1070
|
| 652 |
},
|
| 653 |
{
|
| 654 |
"epoch": 4.0,
|
| 655 |
+
"learning_rate": 0.00012,
|
| 656 |
+
"loss": 0.1043,
|
| 657 |
"step": 1080
|
| 658 |
},
|
| 659 |
{
|
| 660 |
+
"epoch": 4.04,
|
| 661 |
+
"learning_rate": 0.00011925925925925927,
|
| 662 |
+
"loss": 0.0642,
|
| 663 |
+
"step": 1090
|
| 664 |
+
},
|
| 665 |
+
{
|
| 666 |
+
"epoch": 4.07,
|
| 667 |
+
"learning_rate": 0.00011851851851851852,
|
| 668 |
+
"loss": 0.0815,
|
| 669 |
+
"step": 1100
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"epoch": 4.11,
|
| 673 |
+
"learning_rate": 0.00011777777777777779,
|
| 674 |
+
"loss": 0.1572,
|
| 675 |
+
"step": 1110
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 4.15,
|
| 679 |
+
"learning_rate": 0.00011703703703703704,
|
| 680 |
+
"loss": 0.1935,
|
| 681 |
+
"step": 1120
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 4.19,
|
| 685 |
+
"learning_rate": 0.00011629629629629631,
|
| 686 |
+
"loss": 0.183,
|
| 687 |
+
"step": 1130
|
| 688 |
+
},
|
| 689 |
+
{
|
| 690 |
+
"epoch": 4.22,
|
| 691 |
+
"learning_rate": 0.00011555555555555555,
|
| 692 |
+
"loss": 0.0969,
|
| 693 |
+
"step": 1140
|
| 694 |
+
},
|
| 695 |
+
{
|
| 696 |
+
"epoch": 4.26,
|
| 697 |
+
"learning_rate": 0.00011481481481481482,
|
| 698 |
+
"loss": 0.1347,
|
| 699 |
+
"step": 1150
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 4.3,
|
| 703 |
+
"learning_rate": 0.00011407407407407408,
|
| 704 |
+
"loss": 0.0702,
|
| 705 |
+
"step": 1160
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 4.33,
|
| 709 |
+
"learning_rate": 0.00011333333333333334,
|
| 710 |
+
"loss": 0.0834,
|
| 711 |
+
"step": 1170
|
| 712 |
+
},
|
| 713 |
+
{
|
| 714 |
+
"epoch": 4.37,
|
| 715 |
+
"learning_rate": 0.0001125925925925926,
|
| 716 |
+
"loss": 0.144,
|
| 717 |
+
"step": 1180
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 4.41,
|
| 721 |
+
"learning_rate": 0.00011185185185185186,
|
| 722 |
+
"loss": 0.2074,
|
| 723 |
+
"step": 1190
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 4.44,
|
| 727 |
+
"learning_rate": 0.00011111111111111112,
|
| 728 |
+
"loss": 0.1733,
|
| 729 |
+
"step": 1200
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 4.48,
|
| 733 |
+
"learning_rate": 0.00011037037037037037,
|
| 734 |
+
"loss": 0.1018,
|
| 735 |
+
"step": 1210
|
| 736 |
+
},
|
| 737 |
+
{
|
| 738 |
+
"epoch": 4.52,
|
| 739 |
+
"learning_rate": 0.00010962962962962963,
|
| 740 |
+
"loss": 0.1497,
|
| 741 |
+
"step": 1220
|
| 742 |
+
},
|
| 743 |
+
{
|
| 744 |
+
"epoch": 4.56,
|
| 745 |
+
"learning_rate": 0.00010888888888888889,
|
| 746 |
+
"loss": 0.1547,
|
| 747 |
+
"step": 1230
|
| 748 |
+
},
|
| 749 |
+
{
|
| 750 |
+
"epoch": 4.59,
|
| 751 |
+
"learning_rate": 0.00010814814814814815,
|
| 752 |
+
"loss": 0.0763,
|
| 753 |
+
"step": 1240
|
| 754 |
+
},
|
| 755 |
+
{
|
| 756 |
+
"epoch": 4.63,
|
| 757 |
+
"learning_rate": 0.00010740740740740742,
|
| 758 |
+
"loss": 0.0262,
|
| 759 |
+
"step": 1250
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"epoch": 4.67,
|
| 763 |
+
"learning_rate": 0.00010666666666666667,
|
| 764 |
+
"loss": 0.0448,
|
| 765 |
+
"step": 1260
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 4.7,
|
| 769 |
+
"learning_rate": 0.00010592592592592594,
|
| 770 |
+
"loss": 0.1568,
|
| 771 |
+
"step": 1270
|
| 772 |
+
},
|
| 773 |
+
{
|
| 774 |
+
"epoch": 4.74,
|
| 775 |
+
"learning_rate": 0.00010518518518518518,
|
| 776 |
+
"loss": 0.0858,
|
| 777 |
+
"step": 1280
|
| 778 |
+
},
|
| 779 |
+
{
|
| 780 |
+
"epoch": 4.78,
|
| 781 |
+
"learning_rate": 0.00010444444444444445,
|
| 782 |
+
"loss": 0.0615,
|
| 783 |
+
"step": 1290
|
| 784 |
+
},
|
| 785 |
+
{
|
| 786 |
+
"epoch": 4.81,
|
| 787 |
+
"learning_rate": 0.0001037037037037037,
|
| 788 |
+
"loss": 0.0791,
|
| 789 |
+
"step": 1300
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"epoch": 4.85,
|
| 793 |
+
"learning_rate": 0.00010296296296296297,
|
| 794 |
+
"loss": 0.1502,
|
| 795 |
+
"step": 1310
|
| 796 |
+
},
|
| 797 |
+
{
|
| 798 |
+
"epoch": 4.89,
|
| 799 |
+
"learning_rate": 0.00010222222222222222,
|
| 800 |
+
"loss": 0.0945,
|
| 801 |
+
"step": 1320
|
| 802 |
+
},
|
| 803 |
+
{
|
| 804 |
+
"epoch": 4.93,
|
| 805 |
+
"learning_rate": 0.00010148148148148149,
|
| 806 |
+
"loss": 0.0217,
|
| 807 |
+
"step": 1330
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 4.96,
|
| 811 |
+
"learning_rate": 0.00010074074074074073,
|
| 812 |
+
"loss": 0.0726,
|
| 813 |
+
"step": 1340
|
| 814 |
+
},
|
| 815 |
+
{
|
| 816 |
+
"epoch": 5.0,
|
| 817 |
+
"learning_rate": 0.0001,
|
| 818 |
+
"loss": 0.0872,
|
| 819 |
+
"step": 1350
|
| 820 |
+
},
|
| 821 |
+
{
|
| 822 |
+
"epoch": 5.04,
|
| 823 |
+
"learning_rate": 9.925925925925926e-05,
|
| 824 |
+
"loss": 0.0646,
|
| 825 |
+
"step": 1360
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 5.07,
|
| 829 |
+
"learning_rate": 9.851851851851852e-05,
|
| 830 |
+
"loss": 0.0929,
|
| 831 |
+
"step": 1370
|
| 832 |
+
},
|
| 833 |
+
{
|
| 834 |
+
"epoch": 5.11,
|
| 835 |
+
"learning_rate": 9.777777777777778e-05,
|
| 836 |
+
"loss": 0.0149,
|
| 837 |
+
"step": 1380
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 5.15,
|
| 841 |
+
"learning_rate": 9.703703703703704e-05,
|
| 842 |
+
"loss": 0.0757,
|
| 843 |
+
"step": 1390
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 5.19,
|
| 847 |
+
"learning_rate": 9.62962962962963e-05,
|
| 848 |
+
"loss": 0.021,
|
| 849 |
+
"step": 1400
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 5.22,
|
| 853 |
+
"learning_rate": 9.555555555555557e-05,
|
| 854 |
+
"loss": 0.0265,
|
| 855 |
+
"step": 1410
|
| 856 |
+
},
|
| 857 |
+
{
|
| 858 |
+
"epoch": 5.26,
|
| 859 |
+
"learning_rate": 9.481481481481483e-05,
|
| 860 |
+
"loss": 0.0923,
|
| 861 |
+
"step": 1420
|
| 862 |
+
},
|
| 863 |
+
{
|
| 864 |
+
"epoch": 5.3,
|
| 865 |
+
"learning_rate": 9.407407407407408e-05,
|
| 866 |
+
"loss": 0.0344,
|
| 867 |
+
"step": 1430
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"epoch": 5.33,
|
| 871 |
+
"learning_rate": 9.333333333333334e-05,
|
| 872 |
+
"loss": 0.0725,
|
| 873 |
+
"step": 1440
|
| 874 |
+
},
|
| 875 |
+
{
|
| 876 |
+
"epoch": 5.37,
|
| 877 |
+
"learning_rate": 9.25925925925926e-05,
|
| 878 |
+
"loss": 0.0452,
|
| 879 |
+
"step": 1450
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"epoch": 5.41,
|
| 883 |
+
"learning_rate": 9.185185185185186e-05,
|
| 884 |
+
"loss": 0.0904,
|
| 885 |
+
"step": 1460
|
| 886 |
+
},
|
| 887 |
+
{
|
| 888 |
+
"epoch": 5.44,
|
| 889 |
+
"learning_rate": 9.111111111111112e-05,
|
| 890 |
+
"loss": 0.0833,
|
| 891 |
+
"step": 1470
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 5.48,
|
| 895 |
+
"learning_rate": 9.037037037037038e-05,
|
| 896 |
+
"loss": 0.0235,
|
| 897 |
+
"step": 1480
|
| 898 |
+
},
|
| 899 |
+
{
|
| 900 |
+
"epoch": 5.52,
|
| 901 |
+
"learning_rate": 8.962962962962963e-05,
|
| 902 |
+
"loss": 0.0355,
|
| 903 |
+
"step": 1490
|
| 904 |
+
},
|
| 905 |
+
{
|
| 906 |
+
"epoch": 5.56,
|
| 907 |
+
"learning_rate": 8.888888888888889e-05,
|
| 908 |
+
"loss": 0.0139,
|
| 909 |
+
"step": 1500
|
| 910 |
+
},
|
| 911 |
+
{
|
| 912 |
+
"epoch": 5.59,
|
| 913 |
+
"learning_rate": 8.814814814814815e-05,
|
| 914 |
+
"loss": 0.0454,
|
| 915 |
+
"step": 1510
|
| 916 |
+
},
|
| 917 |
+
{
|
| 918 |
+
"epoch": 5.63,
|
| 919 |
+
"learning_rate": 8.740740740740741e-05,
|
| 920 |
+
"loss": 0.0556,
|
| 921 |
+
"step": 1520
|
| 922 |
+
},
|
| 923 |
+
{
|
| 924 |
+
"epoch": 5.67,
|
| 925 |
+
"learning_rate": 8.666666666666667e-05,
|
| 926 |
+
"loss": 0.0757,
|
| 927 |
+
"step": 1530
|
| 928 |
+
},
|
| 929 |
+
{
|
| 930 |
+
"epoch": 5.7,
|
| 931 |
+
"learning_rate": 8.592592592592593e-05,
|
| 932 |
+
"loss": 0.0382,
|
| 933 |
+
"step": 1540
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 5.74,
|
| 937 |
+
"learning_rate": 8.518518518518518e-05,
|
| 938 |
+
"loss": 0.0211,
|
| 939 |
+
"step": 1550
|
| 940 |
+
},
|
| 941 |
+
{
|
| 942 |
+
"epoch": 5.78,
|
| 943 |
+
"learning_rate": 8.444444444444444e-05,
|
| 944 |
+
"loss": 0.0542,
|
| 945 |
+
"step": 1560
|
| 946 |
+
},
|
| 947 |
+
{
|
| 948 |
+
"epoch": 5.81,
|
| 949 |
+
"learning_rate": 8.37037037037037e-05,
|
| 950 |
+
"loss": 0.0081,
|
| 951 |
+
"step": 1570
|
| 952 |
+
},
|
| 953 |
+
{
|
| 954 |
+
"epoch": 5.85,
|
| 955 |
+
"learning_rate": 8.296296296296296e-05,
|
| 956 |
+
"loss": 0.0543,
|
| 957 |
+
"step": 1580
|
| 958 |
+
},
|
| 959 |
+
{
|
| 960 |
+
"epoch": 5.89,
|
| 961 |
+
"learning_rate": 8.222222222222222e-05,
|
| 962 |
+
"loss": 0.0325,
|
| 963 |
+
"step": 1590
|
| 964 |
+
},
|
| 965 |
+
{
|
| 966 |
+
"epoch": 5.93,
|
| 967 |
+
"learning_rate": 8.148148148148148e-05,
|
| 968 |
+
"loss": 0.057,
|
| 969 |
+
"step": 1600
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 5.96,
|
| 973 |
+
"learning_rate": 8.074074074074075e-05,
|
| 974 |
+
"loss": 0.0534,
|
| 975 |
+
"step": 1610
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 6.0,
|
| 979 |
+
"learning_rate": 8e-05,
|
| 980 |
+
"loss": 0.057,
|
| 981 |
+
"step": 1620
|
| 982 |
+
},
|
| 983 |
+
{
|
| 984 |
+
"epoch": 6.04,
|
| 985 |
+
"learning_rate": 7.925925925925926e-05,
|
| 986 |
+
"loss": 0.0262,
|
| 987 |
+
"step": 1630
|
| 988 |
+
},
|
| 989 |
+
{
|
| 990 |
+
"epoch": 6.07,
|
| 991 |
+
"learning_rate": 7.851851851851852e-05,
|
| 992 |
+
"loss": 0.0527,
|
| 993 |
+
"step": 1640
|
| 994 |
+
},
|
| 995 |
+
{
|
| 996 |
+
"epoch": 6.11,
|
| 997 |
+
"learning_rate": 7.777777777777778e-05,
|
| 998 |
+
"loss": 0.0426,
|
| 999 |
+
"step": 1650
|
| 1000 |
+
},
|
| 1001 |
+
{
|
| 1002 |
+
"epoch": 6.15,
|
| 1003 |
+
"learning_rate": 7.703703703703704e-05,
|
| 1004 |
+
"loss": 0.017,
|
| 1005 |
+
"step": 1660
|
| 1006 |
+
},
|
| 1007 |
+
{
|
| 1008 |
+
"epoch": 6.19,
|
| 1009 |
+
"learning_rate": 7.62962962962963e-05,
|
| 1010 |
+
"loss": 0.0428,
|
| 1011 |
+
"step": 1670
|
| 1012 |
+
},
|
| 1013 |
+
{
|
| 1014 |
+
"epoch": 6.22,
|
| 1015 |
+
"learning_rate": 7.555555555555556e-05,
|
| 1016 |
+
"loss": 0.0371,
|
| 1017 |
+
"step": 1680
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 6.26,
|
| 1021 |
+
"learning_rate": 7.481481481481481e-05,
|
| 1022 |
+
"loss": 0.0119,
|
| 1023 |
+
"step": 1690
|
| 1024 |
+
},
|
| 1025 |
+
{
|
| 1026 |
+
"epoch": 6.3,
|
| 1027 |
+
"learning_rate": 7.407407407407407e-05,
|
| 1028 |
+
"loss": 0.0597,
|
| 1029 |
+
"step": 1700
|
| 1030 |
+
},
|
| 1031 |
+
{
|
| 1032 |
+
"epoch": 6.33,
|
| 1033 |
+
"learning_rate": 7.333333333333333e-05,
|
| 1034 |
+
"loss": 0.008,
|
| 1035 |
+
"step": 1710
|
| 1036 |
+
},
|
| 1037 |
+
{
|
| 1038 |
+
"epoch": 6.37,
|
| 1039 |
+
"learning_rate": 7.25925925925926e-05,
|
| 1040 |
+
"loss": 0.0173,
|
| 1041 |
+
"step": 1720
|
| 1042 |
+
},
|
| 1043 |
+
{
|
| 1044 |
+
"epoch": 6.41,
|
| 1045 |
+
"learning_rate": 7.185185185185186e-05,
|
| 1046 |
+
"loss": 0.0409,
|
| 1047 |
+
"step": 1730
|
| 1048 |
+
},
|
| 1049 |
+
{
|
| 1050 |
+
"epoch": 6.44,
|
| 1051 |
+
"learning_rate": 7.111111111111112e-05,
|
| 1052 |
+
"loss": 0.0346,
|
| 1053 |
+
"step": 1740
|
| 1054 |
+
},
|
| 1055 |
+
{
|
| 1056 |
+
"epoch": 6.48,
|
| 1057 |
+
"learning_rate": 7.037037037037038e-05,
|
| 1058 |
+
"loss": 0.0053,
|
| 1059 |
+
"step": 1750
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 6.52,
|
| 1063 |
+
"learning_rate": 6.962962962962964e-05,
|
| 1064 |
+
"loss": 0.0202,
|
| 1065 |
+
"step": 1760
|
| 1066 |
+
},
|
| 1067 |
+
{
|
| 1068 |
+
"epoch": 6.56,
|
| 1069 |
+
"learning_rate": 6.88888888888889e-05,
|
| 1070 |
+
"loss": 0.0149,
|
| 1071 |
+
"step": 1770
|
| 1072 |
+
},
|
| 1073 |
+
{
|
| 1074 |
+
"epoch": 6.59,
|
| 1075 |
+
"learning_rate": 6.814814814814815e-05,
|
| 1076 |
+
"loss": 0.011,
|
| 1077 |
+
"step": 1780
|
| 1078 |
+
},
|
| 1079 |
+
{
|
| 1080 |
+
"epoch": 6.63,
|
| 1081 |
+
"learning_rate": 6.740740740740741e-05,
|
| 1082 |
+
"loss": 0.0407,
|
| 1083 |
+
"step": 1790
|
| 1084 |
+
},
|
| 1085 |
+
{
|
| 1086 |
+
"epoch": 6.67,
|
| 1087 |
+
"learning_rate": 6.666666666666667e-05,
|
| 1088 |
+
"loss": 0.022,
|
| 1089 |
+
"step": 1800
|
| 1090 |
+
},
|
| 1091 |
+
{
|
| 1092 |
+
"epoch": 6.7,
|
| 1093 |
+
"learning_rate": 6.592592592592593e-05,
|
| 1094 |
+
"loss": 0.0375,
|
| 1095 |
+
"step": 1810
|
| 1096 |
+
},
|
| 1097 |
+
{
|
| 1098 |
+
"epoch": 6.74,
|
| 1099 |
+
"learning_rate": 6.51851851851852e-05,
|
| 1100 |
+
"loss": 0.0043,
|
| 1101 |
+
"step": 1820
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 6.78,
|
| 1105 |
+
"learning_rate": 6.444444444444446e-05,
|
| 1106 |
+
"loss": 0.0057,
|
| 1107 |
+
"step": 1830
|
| 1108 |
+
},
|
| 1109 |
+
{
|
| 1110 |
+
"epoch": 6.81,
|
| 1111 |
+
"learning_rate": 6.37037037037037e-05,
|
| 1112 |
+
"loss": 0.0176,
|
| 1113 |
+
"step": 1840
|
| 1114 |
+
},
|
| 1115 |
+
{
|
| 1116 |
+
"epoch": 6.85,
|
| 1117 |
+
"learning_rate": 6.296296296296296e-05,
|
| 1118 |
+
"loss": 0.0042,
|
| 1119 |
+
"step": 1850
|
| 1120 |
+
},
|
| 1121 |
+
{
|
| 1122 |
+
"epoch": 6.89,
|
| 1123 |
+
"learning_rate": 6.222222222222222e-05,
|
| 1124 |
+
"loss": 0.0059,
|
| 1125 |
+
"step": 1860
|
| 1126 |
+
},
|
| 1127 |
+
{
|
| 1128 |
+
"epoch": 6.93,
|
| 1129 |
+
"learning_rate": 6.148148148148148e-05,
|
| 1130 |
+
"loss": 0.0126,
|
| 1131 |
+
"step": 1870
|
| 1132 |
+
},
|
| 1133 |
+
{
|
| 1134 |
+
"epoch": 6.96,
|
| 1135 |
+
"learning_rate": 6.074074074074074e-05,
|
| 1136 |
+
"loss": 0.0049,
|
| 1137 |
+
"step": 1880
|
| 1138 |
+
},
|
| 1139 |
+
{
|
| 1140 |
+
"epoch": 7.0,
|
| 1141 |
+
"learning_rate": 6e-05,
|
| 1142 |
+
"loss": 0.0196,
|
| 1143 |
+
"step": 1890
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 7.04,
|
| 1147 |
+
"learning_rate": 5.925925925925926e-05,
|
| 1148 |
+
"loss": 0.0064,
|
| 1149 |
+
"step": 1900
|
| 1150 |
+
},
|
| 1151 |
+
{
|
| 1152 |
+
"epoch": 7.07,
|
| 1153 |
+
"learning_rate": 5.851851851851852e-05,
|
| 1154 |
+
"loss": 0.006,
|
| 1155 |
+
"step": 1910
|
| 1156 |
+
},
|
| 1157 |
+
{
|
| 1158 |
+
"epoch": 7.11,
|
| 1159 |
+
"learning_rate": 5.7777777777777776e-05,
|
| 1160 |
+
"loss": 0.004,
|
| 1161 |
+
"step": 1920
|
| 1162 |
+
},
|
| 1163 |
+
{
|
| 1164 |
+
"epoch": 7.15,
|
| 1165 |
+
"learning_rate": 5.703703703703704e-05,
|
| 1166 |
+
"loss": 0.0075,
|
| 1167 |
+
"step": 1930
|
| 1168 |
+
},
|
| 1169 |
+
{
|
| 1170 |
+
"epoch": 7.19,
|
| 1171 |
+
"learning_rate": 5.62962962962963e-05,
|
| 1172 |
+
"loss": 0.059,
|
| 1173 |
+
"step": 1940
|
| 1174 |
+
},
|
| 1175 |
+
{
|
| 1176 |
+
"epoch": 7.22,
|
| 1177 |
+
"learning_rate": 5.555555555555556e-05,
|
| 1178 |
+
"loss": 0.004,
|
| 1179 |
+
"step": 1950
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 7.26,
|
| 1183 |
+
"learning_rate": 5.4814814814814817e-05,
|
| 1184 |
+
"loss": 0.0039,
|
| 1185 |
+
"step": 1960
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 7.3,
|
| 1189 |
+
"learning_rate": 5.4074074074074075e-05,
|
| 1190 |
+
"loss": 0.0078,
|
| 1191 |
+
"step": 1970
|
| 1192 |
+
},
|
| 1193 |
+
{
|
| 1194 |
+
"epoch": 7.33,
|
| 1195 |
+
"learning_rate": 5.333333333333333e-05,
|
| 1196 |
+
"loss": 0.0046,
|
| 1197 |
+
"step": 1980
|
| 1198 |
+
},
|
| 1199 |
+
{
|
| 1200 |
+
"epoch": 7.37,
|
| 1201 |
+
"learning_rate": 5.259259259259259e-05,
|
| 1202 |
+
"loss": 0.0135,
|
| 1203 |
+
"step": 1990
|
| 1204 |
+
},
|
| 1205 |
+
{
|
| 1206 |
+
"epoch": 7.41,
|
| 1207 |
+
"learning_rate": 5.185185185185185e-05,
|
| 1208 |
+
"loss": 0.0208,
|
| 1209 |
+
"step": 2000
|
| 1210 |
+
},
|
| 1211 |
+
{
|
| 1212 |
+
"epoch": 7.44,
|
| 1213 |
+
"learning_rate": 5.111111111111111e-05,
|
| 1214 |
+
"loss": 0.0124,
|
| 1215 |
+
"step": 2010
|
| 1216 |
+
},
|
| 1217 |
+
{
|
| 1218 |
+
"epoch": 7.48,
|
| 1219 |
+
"learning_rate": 5.0370370370370366e-05,
|
| 1220 |
+
"loss": 0.0039,
|
| 1221 |
+
"step": 2020
|
| 1222 |
+
},
|
| 1223 |
+
{
|
| 1224 |
+
"epoch": 7.52,
|
| 1225 |
+
"learning_rate": 4.962962962962963e-05,
|
| 1226 |
+
"loss": 0.0037,
|
| 1227 |
+
"step": 2030
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 7.56,
|
| 1231 |
+
"learning_rate": 4.888888888888889e-05,
|
| 1232 |
+
"loss": 0.0042,
|
| 1233 |
+
"step": 2040
|
| 1234 |
+
},
|
| 1235 |
+
{
|
| 1236 |
+
"epoch": 7.59,
|
| 1237 |
+
"learning_rate": 4.814814814814815e-05,
|
| 1238 |
+
"loss": 0.0032,
|
| 1239 |
+
"step": 2050
|
| 1240 |
+
},
|
| 1241 |
+
{
|
| 1242 |
+
"epoch": 7.63,
|
| 1243 |
+
"learning_rate": 4.740740740740741e-05,
|
| 1244 |
+
"loss": 0.0037,
|
| 1245 |
+
"step": 2060
|
| 1246 |
+
},
|
| 1247 |
+
{
|
| 1248 |
+
"epoch": 7.67,
|
| 1249 |
+
"learning_rate": 4.666666666666667e-05,
|
| 1250 |
+
"loss": 0.0035,
|
| 1251 |
+
"step": 2070
|
| 1252 |
+
},
|
| 1253 |
+
{
|
| 1254 |
+
"epoch": 7.7,
|
| 1255 |
+
"learning_rate": 4.592592592592593e-05,
|
| 1256 |
+
"loss": 0.0033,
|
| 1257 |
+
"step": 2080
|
| 1258 |
+
},
|
| 1259 |
+
{
|
| 1260 |
+
"epoch": 7.74,
|
| 1261 |
+
"learning_rate": 4.518518518518519e-05,
|
| 1262 |
+
"loss": 0.0164,
|
| 1263 |
+
"step": 2090
|
| 1264 |
+
},
|
| 1265 |
+
{
|
| 1266 |
+
"epoch": 7.78,
|
| 1267 |
+
"learning_rate": 4.4444444444444447e-05,
|
| 1268 |
+
"loss": 0.0036,
|
| 1269 |
+
"step": 2100
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 7.81,
|
| 1273 |
+
"learning_rate": 4.3703703703703705e-05,
|
| 1274 |
+
"loss": 0.0033,
|
| 1275 |
+
"step": 2110
|
| 1276 |
+
},
|
| 1277 |
+
{
|
| 1278 |
+
"epoch": 7.85,
|
| 1279 |
+
"learning_rate": 4.296296296296296e-05,
|
| 1280 |
+
"loss": 0.0037,
|
| 1281 |
+
"step": 2120
|
| 1282 |
+
},
|
| 1283 |
+
{
|
| 1284 |
+
"epoch": 7.89,
|
| 1285 |
+
"learning_rate": 4.222222222222222e-05,
|
| 1286 |
+
"loss": 0.0321,
|
| 1287 |
+
"step": 2130
|
| 1288 |
+
},
|
| 1289 |
+
{
|
| 1290 |
+
"epoch": 7.93,
|
| 1291 |
+
"learning_rate": 4.148148148148148e-05,
|
| 1292 |
+
"loss": 0.0034,
|
| 1293 |
+
"step": 2140
|
| 1294 |
+
},
|
| 1295 |
+
{
|
| 1296 |
+
"epoch": 7.96,
|
| 1297 |
+
"learning_rate": 4.074074074074074e-05,
|
| 1298 |
+
"loss": 0.0034,
|
| 1299 |
+
"step": 2150
|
| 1300 |
+
},
|
| 1301 |
+
{
|
| 1302 |
+
"epoch": 8.0,
|
| 1303 |
+
"learning_rate": 4e-05,
|
| 1304 |
+
"loss": 0.013,
|
| 1305 |
+
"step": 2160
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 8.04,
|
| 1309 |
+
"learning_rate": 3.925925925925926e-05,
|
| 1310 |
+
"loss": 0.0088,
|
| 1311 |
+
"step": 2170
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 8.07,
|
| 1315 |
+
"learning_rate": 3.851851851851852e-05,
|
| 1316 |
+
"loss": 0.0033,
|
| 1317 |
+
"step": 2180
|
| 1318 |
+
},
|
| 1319 |
+
{
|
| 1320 |
+
"epoch": 8.11,
|
| 1321 |
+
"learning_rate": 3.777777777777778e-05,
|
| 1322 |
+
"loss": 0.0197,
|
| 1323 |
+
"step": 2190
|
| 1324 |
+
},
|
| 1325 |
+
{
|
| 1326 |
+
"epoch": 8.15,
|
| 1327 |
+
"learning_rate": 3.7037037037037037e-05,
|
| 1328 |
+
"loss": 0.0033,
|
| 1329 |
+
"step": 2200
|
| 1330 |
+
},
|
| 1331 |
+
{
|
| 1332 |
+
"epoch": 8.19,
|
| 1333 |
+
"learning_rate": 3.62962962962963e-05,
|
| 1334 |
+
"loss": 0.0031,
|
| 1335 |
+
"step": 2210
|
| 1336 |
+
},
|
| 1337 |
+
{
|
| 1338 |
+
"epoch": 8.22,
|
| 1339 |
+
"learning_rate": 3.555555555555556e-05,
|
| 1340 |
+
"loss": 0.0031,
|
| 1341 |
+
"step": 2220
|
| 1342 |
+
},
|
| 1343 |
+
{
|
| 1344 |
+
"epoch": 8.26,
|
| 1345 |
+
"learning_rate": 3.481481481481482e-05,
|
| 1346 |
+
"loss": 0.003,
|
| 1347 |
+
"step": 2230
|
| 1348 |
+
},
|
| 1349 |
+
{
|
| 1350 |
+
"epoch": 8.3,
|
| 1351 |
+
"learning_rate": 3.4074074074074077e-05,
|
| 1352 |
+
"loss": 0.003,
|
| 1353 |
+
"step": 2240
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 8.33,
|
| 1357 |
+
"learning_rate": 3.3333333333333335e-05,
|
| 1358 |
+
"loss": 0.0034,
|
| 1359 |
+
"step": 2250
|
| 1360 |
+
},
|
| 1361 |
+
{
|
| 1362 |
+
"epoch": 8.37,
|
| 1363 |
+
"learning_rate": 3.25925925925926e-05,
|
| 1364 |
+
"loss": 0.0135,
|
| 1365 |
+
"step": 2260
|
| 1366 |
+
},
|
| 1367 |
+
{
|
| 1368 |
+
"epoch": 8.41,
|
| 1369 |
+
"learning_rate": 3.185185185185185e-05,
|
| 1370 |
+
"loss": 0.0038,
|
| 1371 |
+
"step": 2270
|
| 1372 |
+
},
|
| 1373 |
+
{
|
| 1374 |
+
"epoch": 8.44,
|
| 1375 |
+
"learning_rate": 3.111111111111111e-05,
|
| 1376 |
+
"loss": 0.003,
|
| 1377 |
+
"step": 2280
|
| 1378 |
+
},
|
| 1379 |
+
{
|
| 1380 |
+
"epoch": 8.48,
|
| 1381 |
+
"learning_rate": 3.037037037037037e-05,
|
| 1382 |
+
"loss": 0.0028,
|
| 1383 |
+
"step": 2290
|
| 1384 |
+
},
|
| 1385 |
+
{
|
| 1386 |
+
"epoch": 8.52,
|
| 1387 |
+
"learning_rate": 2.962962962962963e-05,
|
| 1388 |
+
"loss": 0.0033,
|
| 1389 |
+
"step": 2300
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 8.56,
|
| 1393 |
+
"learning_rate": 2.8888888888888888e-05,
|
| 1394 |
+
"loss": 0.0028,
|
| 1395 |
+
"step": 2310
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 8.59,
|
| 1399 |
+
"learning_rate": 2.814814814814815e-05,
|
| 1400 |
+
"loss": 0.0085,
|
| 1401 |
+
"step": 2320
|
| 1402 |
+
},
|
| 1403 |
+
{
|
| 1404 |
+
"epoch": 8.63,
|
| 1405 |
+
"learning_rate": 2.7407407407407408e-05,
|
| 1406 |
+
"loss": 0.0028,
|
| 1407 |
+
"step": 2330
|
| 1408 |
+
},
|
| 1409 |
+
{
|
| 1410 |
+
"epoch": 8.67,
|
| 1411 |
+
"learning_rate": 2.6666666666666667e-05,
|
| 1412 |
+
"loss": 0.0134,
|
| 1413 |
+
"step": 2340
|
| 1414 |
+
},
|
| 1415 |
+
{
|
| 1416 |
+
"epoch": 8.7,
|
| 1417 |
+
"learning_rate": 2.5925925925925925e-05,
|
| 1418 |
+
"loss": 0.0029,
|
| 1419 |
+
"step": 2350
|
| 1420 |
+
},
|
| 1421 |
+
{
|
| 1422 |
+
"epoch": 8.74,
|
| 1423 |
+
"learning_rate": 2.5185185185185183e-05,
|
| 1424 |
+
"loss": 0.003,
|
| 1425 |
+
"step": 2360
|
| 1426 |
+
},
|
| 1427 |
+
{
|
| 1428 |
+
"epoch": 8.78,
|
| 1429 |
+
"learning_rate": 2.4444444444444445e-05,
|
| 1430 |
+
"loss": 0.0027,
|
| 1431 |
+
"step": 2370
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 8.81,
|
| 1435 |
+
"learning_rate": 2.3703703703703707e-05,
|
| 1436 |
+
"loss": 0.0027,
|
| 1437 |
+
"step": 2380
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 8.85,
|
| 1441 |
+
"learning_rate": 2.2962962962962965e-05,
|
| 1442 |
+
"loss": 0.0027,
|
| 1443 |
+
"step": 2390
|
| 1444 |
+
},
|
| 1445 |
+
{
|
| 1446 |
+
"epoch": 8.89,
|
| 1447 |
+
"learning_rate": 2.2222222222222223e-05,
|
| 1448 |
+
"loss": 0.0091,
|
| 1449 |
+
"step": 2400
|
| 1450 |
+
},
|
| 1451 |
+
{
|
| 1452 |
+
"epoch": 8.93,
|
| 1453 |
+
"learning_rate": 2.148148148148148e-05,
|
| 1454 |
+
"loss": 0.0028,
|
| 1455 |
+
"step": 2410
|
| 1456 |
+
},
|
| 1457 |
+
{
|
| 1458 |
+
"epoch": 8.96,
|
| 1459 |
+
"learning_rate": 2.074074074074074e-05,
|
| 1460 |
+
"loss": 0.0028,
|
| 1461 |
+
"step": 2420
|
| 1462 |
+
},
|
| 1463 |
+
{
|
| 1464 |
+
"epoch": 9.0,
|
| 1465 |
+
"learning_rate": 2e-05,
|
| 1466 |
+
"loss": 0.0028,
|
| 1467 |
+
"step": 2430
|
| 1468 |
+
},
|
| 1469 |
+
{
|
| 1470 |
+
"epoch": 9.04,
|
| 1471 |
+
"learning_rate": 1.925925925925926e-05,
|
| 1472 |
+
"loss": 0.003,
|
| 1473 |
+
"step": 2440
|
| 1474 |
+
},
|
| 1475 |
+
{
|
| 1476 |
+
"epoch": 9.07,
|
| 1477 |
+
"learning_rate": 1.8518518518518518e-05,
|
| 1478 |
+
"loss": 0.0086,
|
| 1479 |
+
"step": 2450
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 9.11,
|
| 1483 |
+
"learning_rate": 1.777777777777778e-05,
|
| 1484 |
+
"loss": 0.0096,
|
| 1485 |
+
"step": 2460
|
| 1486 |
+
},
|
| 1487 |
+
{
|
| 1488 |
+
"epoch": 9.15,
|
| 1489 |
+
"learning_rate": 1.7037037037037038e-05,
|
| 1490 |
+
"loss": 0.0027,
|
| 1491 |
+
"step": 2470
|
| 1492 |
+
},
|
| 1493 |
+
{
|
| 1494 |
+
"epoch": 9.19,
|
| 1495 |
+
"learning_rate": 1.62962962962963e-05,
|
| 1496 |
+
"loss": 0.0026,
|
| 1497 |
+
"step": 2480
|
| 1498 |
+
},
|
| 1499 |
+
{
|
| 1500 |
+
"epoch": 9.22,
|
| 1501 |
+
"learning_rate": 1.5555555555555555e-05,
|
| 1502 |
+
"loss": 0.0026,
|
| 1503 |
+
"step": 2490
|
| 1504 |
+
},
|
| 1505 |
+
{
|
| 1506 |
+
"epoch": 9.26,
|
| 1507 |
+
"learning_rate": 1.4814814814814815e-05,
|
| 1508 |
+
"loss": 0.0026,
|
| 1509 |
+
"step": 2500
|
| 1510 |
+
},
|
| 1511 |
+
{
|
| 1512 |
+
"epoch": 9.3,
|
| 1513 |
+
"learning_rate": 1.4074074074074075e-05,
|
| 1514 |
+
"loss": 0.0027,
|
| 1515 |
+
"step": 2510
|
| 1516 |
+
},
|
| 1517 |
+
{
|
| 1518 |
+
"epoch": 9.33,
|
| 1519 |
+
"learning_rate": 1.3333333333333333e-05,
|
| 1520 |
+
"loss": 0.0026,
|
| 1521 |
+
"step": 2520
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 9.37,
|
| 1525 |
+
"learning_rate": 1.2592592592592592e-05,
|
| 1526 |
+
"loss": 0.0025,
|
| 1527 |
+
"step": 2530
|
| 1528 |
+
},
|
| 1529 |
+
{
|
| 1530 |
+
"epoch": 9.41,
|
| 1531 |
+
"learning_rate": 1.1851851851851853e-05,
|
| 1532 |
+
"loss": 0.0027,
|
| 1533 |
+
"step": 2540
|
| 1534 |
+
},
|
| 1535 |
+
{
|
| 1536 |
+
"epoch": 9.44,
|
| 1537 |
+
"learning_rate": 1.1111111111111112e-05,
|
| 1538 |
+
"loss": 0.0024,
|
| 1539 |
+
"step": 2550
|
| 1540 |
+
},
|
| 1541 |
+
{
|
| 1542 |
+
"epoch": 9.48,
|
| 1543 |
+
"learning_rate": 1.037037037037037e-05,
|
| 1544 |
+
"loss": 0.0025,
|
| 1545 |
+
"step": 2560
|
| 1546 |
+
},
|
| 1547 |
+
{
|
| 1548 |
+
"epoch": 9.52,
|
| 1549 |
+
"learning_rate": 9.62962962962963e-06,
|
| 1550 |
+
"loss": 0.0077,
|
| 1551 |
+
"step": 2570
|
| 1552 |
+
},
|
| 1553 |
+
{
|
| 1554 |
+
"epoch": 9.56,
|
| 1555 |
+
"learning_rate": 8.88888888888889e-06,
|
| 1556 |
+
"loss": 0.0176,
|
| 1557 |
+
"step": 2580
|
| 1558 |
+
},
|
| 1559 |
+
{
|
| 1560 |
+
"epoch": 9.59,
|
| 1561 |
+
"learning_rate": 8.14814814814815e-06,
|
| 1562 |
+
"loss": 0.0026,
|
| 1563 |
+
"step": 2590
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 9.63,
|
| 1567 |
+
"learning_rate": 7.4074074074074075e-06,
|
| 1568 |
+
"loss": 0.0024,
|
| 1569 |
+
"step": 2600
|
| 1570 |
+
},
|
| 1571 |
+
{
|
| 1572 |
+
"epoch": 9.67,
|
| 1573 |
+
"learning_rate": 6.666666666666667e-06,
|
| 1574 |
+
"loss": 0.0025,
|
| 1575 |
+
"step": 2610
|
| 1576 |
+
},
|
| 1577 |
+
{
|
| 1578 |
+
"epoch": 9.7,
|
| 1579 |
+
"learning_rate": 5.925925925925927e-06,
|
| 1580 |
+
"loss": 0.0025,
|
| 1581 |
+
"step": 2620
|
| 1582 |
+
},
|
| 1583 |
+
{
|
| 1584 |
+
"epoch": 9.74,
|
| 1585 |
+
"learning_rate": 5.185185185185185e-06,
|
| 1586 |
+
"loss": 0.0025,
|
| 1587 |
+
"step": 2630
|
| 1588 |
+
},
|
| 1589 |
+
{
|
| 1590 |
+
"epoch": 9.78,
|
| 1591 |
+
"learning_rate": 4.444444444444445e-06,
|
| 1592 |
+
"loss": 0.0026,
|
| 1593 |
+
"step": 2640
|
| 1594 |
+
},
|
| 1595 |
+
{
|
| 1596 |
+
"epoch": 9.81,
|
| 1597 |
+
"learning_rate": 3.7037037037037037e-06,
|
| 1598 |
+
"loss": 0.0026,
|
| 1599 |
+
"step": 2650
|
| 1600 |
+
},
|
| 1601 |
+
{
|
| 1602 |
+
"epoch": 9.85,
|
| 1603 |
+
"learning_rate": 2.9629629629629633e-06,
|
| 1604 |
+
"loss": 0.0025,
|
| 1605 |
+
"step": 2660
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 9.89,
|
| 1609 |
+
"learning_rate": 2.2222222222222225e-06,
|
| 1610 |
+
"loss": 0.0101,
|
| 1611 |
+
"step": 2670
|
| 1612 |
+
},
|
| 1613 |
+
{
|
| 1614 |
+
"epoch": 9.93,
|
| 1615 |
+
"learning_rate": 1.4814814814814817e-06,
|
| 1616 |
+
"loss": 0.0026,
|
| 1617 |
+
"step": 2680
|
| 1618 |
+
},
|
| 1619 |
+
{
|
| 1620 |
+
"epoch": 9.96,
|
| 1621 |
+
"learning_rate": 7.407407407407408e-07,
|
| 1622 |
+
"loss": 0.0026,
|
| 1623 |
+
"step": 2690
|
| 1624 |
+
},
|
| 1625 |
+
{
|
| 1626 |
+
"epoch": 10.0,
|
| 1627 |
+
"learning_rate": 0.0,
|
| 1628 |
+
"loss": 0.0025,
|
| 1629 |
+
"step": 2700
|
| 1630 |
+
},
|
| 1631 |
+
{
|
| 1632 |
+
"epoch": 10.0,
|
| 1633 |
+
"step": 2700,
|
| 1634 |
+
"total_flos": 3.347863979094835e+18,
|
| 1635 |
+
"train_loss": 0.2644773021671507,
|
| 1636 |
+
"train_runtime": 949.4602,
|
| 1637 |
+
"train_samples_per_second": 45.5,
|
| 1638 |
+
"train_steps_per_second": 2.844
|
| 1639 |
}
|
| 1640 |
],
|
| 1641 |
"logging_steps": 10,
|
| 1642 |
+
"max_steps": 2700,
|
| 1643 |
+
"num_train_epochs": 10,
|
| 1644 |
"save_steps": 3000,
|
| 1645 |
+
"total_flos": 3.347863979094835e+18,
|
| 1646 |
"trial_name": null,
|
| 1647 |
"trial_params": null
|
| 1648 |
}
|