Commit
·
1efaabf
1
Parent(s):
eec3fea
End of training
Browse files- all_results.json +5 -5
- train_results.json +5 -5
- trainer_state.json +1226 -818
all_results.json
CHANGED
|
@@ -1,5 +1,5 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
"eval_accuracy": 0.5895981374153325,
|
| 4 |
"eval_loss": 1.842421054840088,
|
| 5 |
"eval_runtime": 6167.0931,
|
|
@@ -7,9 +7,9 @@
|
|
| 7 |
"eval_samples_per_second": 21.881,
|
| 8 |
"eval_steps_per_second": 10.94,
|
| 9 |
"perplexity": 6.311800993960819,
|
| 10 |
-
"train_loss":
|
| 11 |
-
"train_runtime":
|
| 12 |
"train_samples": 134942,
|
| 13 |
-
"train_samples_per_second":
|
| 14 |
-
"train_steps_per_second":
|
| 15 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 5.0,
|
| 3 |
"eval_accuracy": 0.5895981374153325,
|
| 4 |
"eval_loss": 1.842421054840088,
|
| 5 |
"eval_runtime": 6167.0931,
|
|
|
|
| 7 |
"eval_samples_per_second": 21.881,
|
| 8 |
"eval_steps_per_second": 10.94,
|
| 9 |
"perplexity": 6.311800993960819,
|
| 10 |
+
"train_loss": 1.883479554409491,
|
| 11 |
+
"train_runtime": 65223.4395,
|
| 12 |
"train_samples": 134942,
|
| 13 |
+
"train_samples_per_second": 10.345,
|
| 14 |
+
"train_steps_per_second": 2.586
|
| 15 |
}
|
train_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"train_loss":
|
| 4 |
-
"train_runtime":
|
| 5 |
"train_samples": 134942,
|
| 6 |
-
"train_samples_per_second":
|
| 7 |
-
"train_steps_per_second":
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 5.0,
|
| 3 |
+
"train_loss": 1.883479554409491,
|
| 4 |
+
"train_runtime": 65223.4395,
|
| 5 |
"train_samples": 134942,
|
| 6 |
+
"train_samples_per_second": 10.345,
|
| 7 |
+
"train_steps_per_second": 2.586
|
| 8 |
}
|
trainer_state.json
CHANGED
|
@@ -1,1639 +1,2047 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch":
|
| 5 |
-
"global_step":
|
| 6 |
"is_hyper_param_search": false,
|
| 7 |
"is_local_process_zero": true,
|
| 8 |
"is_world_process_zero": true,
|
| 9 |
"log_history": [
|
| 10 |
{
|
| 11 |
"epoch": 0.01,
|
| 12 |
-
"learning_rate": 4.
|
| 13 |
-
"loss":
|
| 14 |
"step": 500
|
| 15 |
},
|
| 16 |
{
|
| 17 |
-
"epoch": 0.
|
| 18 |
-
"learning_rate": 4.
|
| 19 |
-
"loss": 2.
|
| 20 |
"step": 1000
|
| 21 |
},
|
| 22 |
{
|
| 23 |
-
"epoch": 0.
|
| 24 |
-
"learning_rate": 4.
|
| 25 |
-
"loss": 2.
|
| 26 |
"step": 1500
|
| 27 |
},
|
| 28 |
{
|
| 29 |
-
"epoch": 0.
|
| 30 |
-
"learning_rate": 4.
|
| 31 |
-
"loss": 2.
|
| 32 |
"step": 2000
|
| 33 |
},
|
| 34 |
{
|
| 35 |
-
"epoch": 0.
|
| 36 |
-
"learning_rate": 4.
|
| 37 |
-
"loss": 2.
|
| 38 |
"step": 2500
|
| 39 |
},
|
| 40 |
{
|
| 41 |
-
"epoch": 0.
|
| 42 |
-
"learning_rate": 4.
|
| 43 |
-
"loss": 2.
|
| 44 |
"step": 3000
|
| 45 |
},
|
| 46 |
{
|
| 47 |
-
"epoch": 0.
|
| 48 |
-
"learning_rate": 4.
|
| 49 |
-
"loss": 2.
|
| 50 |
"step": 3500
|
| 51 |
},
|
| 52 |
{
|
| 53 |
-
"epoch": 0.
|
| 54 |
-
"learning_rate": 4.
|
| 55 |
-
"loss": 2.
|
| 56 |
"step": 4000
|
| 57 |
},
|
| 58 |
{
|
| 59 |
-
"epoch": 0.
|
| 60 |
-
"learning_rate": 4.
|
| 61 |
-
"loss": 2.
|
| 62 |
"step": 4500
|
| 63 |
},
|
| 64 |
{
|
| 65 |
-
"epoch": 0.
|
| 66 |
-
"learning_rate": 4.
|
| 67 |
-
"loss": 2.
|
| 68 |
"step": 5000
|
| 69 |
},
|
| 70 |
{
|
| 71 |
-
"epoch": 0.
|
| 72 |
-
"learning_rate": 4.
|
| 73 |
-
"loss": 2.
|
| 74 |
"step": 5500
|
| 75 |
},
|
| 76 |
{
|
| 77 |
-
"epoch": 0.
|
| 78 |
-
"learning_rate": 4.
|
| 79 |
-
"loss": 2.
|
| 80 |
"step": 6000
|
| 81 |
},
|
| 82 |
{
|
| 83 |
-
"epoch": 0.
|
| 84 |
-
"learning_rate": 4.
|
| 85 |
-
"loss": 2.
|
| 86 |
"step": 6500
|
| 87 |
},
|
| 88 |
{
|
| 89 |
-
"epoch": 0.
|
| 90 |
-
"learning_rate": 4.
|
| 91 |
-
"loss": 2.
|
| 92 |
"step": 7000
|
| 93 |
},
|
| 94 |
{
|
| 95 |
-
"epoch": 0.
|
| 96 |
-
"learning_rate": 4.
|
| 97 |
-
"loss":
|
| 98 |
"step": 7500
|
| 99 |
},
|
| 100 |
{
|
| 101 |
-
"epoch": 0.
|
| 102 |
-
"learning_rate": 4.
|
| 103 |
-
"loss": 2.
|
| 104 |
"step": 8000
|
| 105 |
},
|
| 106 |
{
|
| 107 |
-
"epoch": 0.
|
| 108 |
-
"learning_rate": 4.
|
| 109 |
-
"loss": 2.
|
| 110 |
"step": 8500
|
| 111 |
},
|
| 112 |
{
|
| 113 |
-
"epoch": 0.
|
| 114 |
-
"learning_rate": 4.
|
| 115 |
-
"loss": 2.
|
| 116 |
"step": 9000
|
| 117 |
},
|
| 118 |
{
|
| 119 |
-
"epoch": 0.
|
| 120 |
-
"learning_rate": 4.
|
| 121 |
-
"loss": 2.
|
| 122 |
"step": 9500
|
| 123 |
},
|
| 124 |
{
|
| 125 |
-
"epoch": 0.
|
| 126 |
-
"learning_rate": 4.
|
| 127 |
-
"loss": 2.
|
| 128 |
"step": 10000
|
| 129 |
},
|
| 130 |
{
|
| 131 |
-
"epoch": 0.
|
| 132 |
-
"learning_rate": 4.
|
| 133 |
-
"loss": 2.
|
| 134 |
"step": 10500
|
| 135 |
},
|
| 136 |
{
|
| 137 |
-
"epoch": 0.
|
| 138 |
-
"learning_rate": 4.
|
| 139 |
-
"loss": 2.
|
| 140 |
"step": 11000
|
| 141 |
},
|
| 142 |
{
|
| 143 |
-
"epoch": 0.
|
| 144 |
-
"learning_rate": 4.
|
| 145 |
-
"loss": 2.
|
| 146 |
"step": 11500
|
| 147 |
},
|
| 148 |
{
|
| 149 |
-
"epoch": 0.
|
| 150 |
-
"learning_rate": 4.
|
| 151 |
-
"loss": 2.
|
| 152 |
"step": 12000
|
| 153 |
},
|
| 154 |
{
|
| 155 |
-
"epoch": 0.
|
| 156 |
-
"learning_rate": 4.
|
| 157 |
-
"loss": 2.
|
| 158 |
"step": 12500
|
| 159 |
},
|
| 160 |
{
|
| 161 |
-
"epoch": 0.
|
| 162 |
-
"learning_rate": 4.
|
| 163 |
-
"loss": 2.
|
| 164 |
"step": 13000
|
| 165 |
},
|
| 166 |
{
|
| 167 |
-
"epoch": 0.
|
| 168 |
-
"learning_rate": 4.
|
| 169 |
-
"loss": 2.
|
| 170 |
"step": 13500
|
| 171 |
},
|
| 172 |
{
|
| 173 |
-
"epoch": 0.
|
| 174 |
-
"learning_rate": 4.
|
| 175 |
-
"loss": 2.
|
| 176 |
"step": 14000
|
| 177 |
},
|
| 178 |
{
|
| 179 |
-
"epoch": 0.
|
| 180 |
-
"learning_rate": 4.
|
| 181 |
-
"loss": 2.
|
| 182 |
"step": 14500
|
| 183 |
},
|
| 184 |
{
|
| 185 |
-
"epoch": 0.
|
| 186 |
-
"learning_rate": 4.
|
| 187 |
-
"loss": 2.
|
| 188 |
"step": 15000
|
| 189 |
},
|
| 190 |
{
|
| 191 |
-
"epoch": 0.
|
| 192 |
-
"learning_rate": 4.
|
| 193 |
-
"loss": 2.
|
| 194 |
"step": 15500
|
| 195 |
},
|
| 196 |
{
|
| 197 |
-
"epoch": 0.
|
| 198 |
-
"learning_rate": 4.
|
| 199 |
-
"loss": 2.
|
| 200 |
"step": 16000
|
| 201 |
},
|
| 202 |
{
|
| 203 |
-
"epoch": 0.
|
| 204 |
-
"learning_rate": 4.
|
| 205 |
-
"loss": 2.
|
| 206 |
"step": 16500
|
| 207 |
},
|
| 208 |
{
|
| 209 |
-
"epoch": 0.
|
| 210 |
-
"learning_rate": 4.
|
| 211 |
-
"loss": 2.
|
| 212 |
"step": 17000
|
| 213 |
},
|
| 214 |
{
|
| 215 |
-
"epoch": 0.
|
| 216 |
-
"learning_rate": 4.
|
| 217 |
-
"loss": 2.
|
| 218 |
"step": 17500
|
| 219 |
},
|
| 220 |
{
|
| 221 |
-
"epoch": 0.
|
| 222 |
-
"learning_rate": 4.
|
| 223 |
-
"loss": 2.
|
| 224 |
"step": 18000
|
| 225 |
},
|
| 226 |
{
|
| 227 |
-
"epoch": 0.
|
| 228 |
-
"learning_rate": 4.
|
| 229 |
-
"loss": 2.
|
| 230 |
"step": 18500
|
| 231 |
},
|
| 232 |
{
|
| 233 |
-
"epoch": 0.
|
| 234 |
-
"learning_rate": 4.
|
| 235 |
-
"loss": 2.
|
| 236 |
"step": 19000
|
| 237 |
},
|
| 238 |
{
|
| 239 |
-
"epoch": 0.
|
| 240 |
-
"learning_rate": 4.
|
| 241 |
-
"loss": 2.
|
| 242 |
"step": 19500
|
| 243 |
},
|
| 244 |
{
|
| 245 |
-
"epoch": 0.
|
| 246 |
-
"learning_rate": 4.
|
| 247 |
-
"loss": 2.
|
| 248 |
"step": 20000
|
| 249 |
},
|
| 250 |
{
|
| 251 |
-
"epoch": 0.
|
| 252 |
-
"learning_rate": 4.
|
| 253 |
-
"loss": 2.
|
| 254 |
"step": 20500
|
| 255 |
},
|
| 256 |
{
|
| 257 |
-
"epoch": 0.
|
| 258 |
-
"learning_rate": 4.
|
| 259 |
-
"loss": 2.
|
| 260 |
"step": 21000
|
| 261 |
},
|
| 262 |
{
|
| 263 |
-
"epoch": 0.
|
| 264 |
-
"learning_rate": 4.
|
| 265 |
-
"loss": 2.
|
| 266 |
"step": 21500
|
| 267 |
},
|
| 268 |
{
|
| 269 |
-
"epoch": 0.
|
| 270 |
-
"learning_rate": 4.
|
| 271 |
-
"loss": 2.
|
| 272 |
"step": 22000
|
| 273 |
},
|
| 274 |
{
|
| 275 |
-
"epoch": 0.
|
| 276 |
-
"learning_rate": 4.
|
| 277 |
-
"loss": 2.
|
| 278 |
"step": 22500
|
| 279 |
},
|
| 280 |
{
|
| 281 |
-
"epoch": 0.
|
| 282 |
-
"learning_rate": 4.
|
| 283 |
-
"loss": 2.
|
| 284 |
"step": 23000
|
| 285 |
},
|
| 286 |
{
|
| 287 |
-
"epoch": 0.
|
| 288 |
-
"learning_rate": 4.
|
| 289 |
-
"loss": 2.
|
| 290 |
"step": 23500
|
| 291 |
},
|
| 292 |
{
|
| 293 |
-
"epoch": 0.
|
| 294 |
-
"learning_rate": 4.
|
| 295 |
-
"loss": 2.
|
| 296 |
"step": 24000
|
| 297 |
},
|
| 298 |
{
|
| 299 |
-
"epoch": 0.
|
| 300 |
-
"learning_rate": 4.
|
| 301 |
-
"loss": 2.
|
| 302 |
"step": 24500
|
| 303 |
},
|
| 304 |
{
|
| 305 |
-
"epoch": 0.
|
| 306 |
-
"learning_rate": 4.
|
| 307 |
-
"loss": 2.
|
| 308 |
"step": 25000
|
| 309 |
},
|
| 310 |
{
|
| 311 |
-
"epoch": 0.
|
| 312 |
-
"learning_rate": 4.
|
| 313 |
-
"loss": 2.
|
| 314 |
"step": 25500
|
| 315 |
},
|
| 316 |
{
|
| 317 |
-
"epoch": 0.
|
| 318 |
-
"learning_rate": 4.
|
| 319 |
-
"loss": 2.
|
| 320 |
"step": 26000
|
| 321 |
},
|
| 322 |
{
|
| 323 |
-
"epoch": 0.
|
| 324 |
-
"learning_rate": 4.
|
| 325 |
-
"loss": 2.
|
| 326 |
"step": 26500
|
| 327 |
},
|
| 328 |
{
|
| 329 |
-
"epoch": 0.
|
| 330 |
-
"learning_rate":
|
| 331 |
-
"loss": 2.
|
| 332 |
"step": 27000
|
| 333 |
},
|
| 334 |
{
|
| 335 |
-
"epoch": 0.
|
| 336 |
-
"learning_rate":
|
| 337 |
-
"loss": 2.
|
| 338 |
"step": 27500
|
| 339 |
},
|
| 340 |
{
|
| 341 |
-
"epoch": 0.
|
| 342 |
-
"learning_rate":
|
| 343 |
-
"loss": 2.
|
| 344 |
"step": 28000
|
| 345 |
},
|
| 346 |
{
|
| 347 |
-
"epoch": 0.
|
| 348 |
-
"learning_rate":
|
| 349 |
-
"loss": 2.
|
| 350 |
"step": 28500
|
| 351 |
},
|
| 352 |
{
|
| 353 |
-
"epoch": 0.
|
| 354 |
-
"learning_rate":
|
| 355 |
-
"loss": 2.
|
| 356 |
"step": 29000
|
| 357 |
},
|
| 358 |
{
|
| 359 |
-
"epoch": 0.
|
| 360 |
-
"learning_rate":
|
| 361 |
-
"loss": 2.
|
| 362 |
"step": 29500
|
| 363 |
},
|
| 364 |
{
|
| 365 |
-
"epoch": 0.
|
| 366 |
-
"learning_rate":
|
| 367 |
-
"loss": 2.
|
| 368 |
"step": 30000
|
| 369 |
},
|
| 370 |
{
|
| 371 |
-
"epoch": 0.
|
| 372 |
-
"learning_rate":
|
| 373 |
-
"loss": 2.
|
| 374 |
"step": 30500
|
| 375 |
},
|
| 376 |
{
|
| 377 |
-
"epoch": 0.
|
| 378 |
-
"learning_rate":
|
| 379 |
-
"loss": 2.
|
| 380 |
"step": 31000
|
| 381 |
},
|
| 382 |
{
|
| 383 |
-
"epoch": 0.
|
| 384 |
-
"learning_rate":
|
| 385 |
-
"loss": 2.
|
| 386 |
"step": 31500
|
| 387 |
},
|
| 388 |
{
|
| 389 |
-
"epoch": 0.
|
| 390 |
-
"learning_rate":
|
| 391 |
-
"loss": 2.
|
| 392 |
"step": 32000
|
| 393 |
},
|
| 394 |
{
|
| 395 |
-
"epoch": 0.
|
| 396 |
-
"learning_rate":
|
| 397 |
-
"loss": 2.
|
| 398 |
"step": 32500
|
| 399 |
},
|
| 400 |
{
|
| 401 |
-
"epoch": 0.
|
| 402 |
-
"learning_rate":
|
| 403 |
-
"loss": 2.
|
| 404 |
"step": 33000
|
| 405 |
},
|
| 406 |
{
|
| 407 |
-
"epoch": 0.
|
| 408 |
-
"learning_rate":
|
| 409 |
-
"loss": 2.
|
| 410 |
"step": 33500
|
| 411 |
},
|
| 412 |
{
|
| 413 |
-
"epoch":
|
| 414 |
-
"learning_rate": 3.
|
| 415 |
-
"loss":
|
| 416 |
"step": 34000
|
| 417 |
},
|
| 418 |
{
|
| 419 |
-
"epoch":
|
| 420 |
-
"learning_rate": 3.
|
| 421 |
-
"loss":
|
| 422 |
"step": 34500
|
| 423 |
},
|
| 424 |
{
|
| 425 |
-
"epoch":
|
| 426 |
-
"learning_rate": 3.
|
| 427 |
-
"loss":
|
| 428 |
"step": 35000
|
| 429 |
},
|
| 430 |
{
|
| 431 |
-
"epoch":
|
| 432 |
-
"learning_rate": 3.
|
| 433 |
-
"loss":
|
| 434 |
"step": 35500
|
| 435 |
},
|
| 436 |
{
|
| 437 |
-
"epoch":
|
| 438 |
-
"learning_rate": 3.
|
| 439 |
-
"loss":
|
| 440 |
"step": 36000
|
| 441 |
},
|
| 442 |
{
|
| 443 |
-
"epoch":
|
| 444 |
-
"learning_rate": 3.
|
| 445 |
-
"loss":
|
| 446 |
"step": 36500
|
| 447 |
},
|
| 448 |
{
|
| 449 |
-
"epoch":
|
| 450 |
-
"learning_rate": 3.
|
| 451 |
-
"loss":
|
| 452 |
"step": 37000
|
| 453 |
},
|
| 454 |
{
|
| 455 |
-
"epoch":
|
| 456 |
-
"learning_rate": 3.
|
| 457 |
-
"loss":
|
| 458 |
"step": 37500
|
| 459 |
},
|
| 460 |
{
|
| 461 |
-
"epoch":
|
| 462 |
-
"learning_rate": 3.
|
| 463 |
-
"loss":
|
| 464 |
"step": 38000
|
| 465 |
},
|
| 466 |
{
|
| 467 |
-
"epoch":
|
| 468 |
-
"learning_rate": 3.
|
| 469 |
-
"loss":
|
| 470 |
"step": 38500
|
| 471 |
},
|
| 472 |
{
|
| 473 |
-
"epoch":
|
| 474 |
-
"learning_rate": 3.
|
| 475 |
-
"loss":
|
| 476 |
"step": 39000
|
| 477 |
},
|
| 478 |
{
|
| 479 |
-
"epoch":
|
| 480 |
-
"learning_rate": 3.
|
| 481 |
-
"loss":
|
| 482 |
"step": 39500
|
| 483 |
},
|
| 484 |
{
|
| 485 |
-
"epoch":
|
| 486 |
-
"learning_rate": 3.
|
| 487 |
-
"loss":
|
| 488 |
"step": 40000
|
| 489 |
},
|
| 490 |
{
|
| 491 |
-
"epoch":
|
| 492 |
-
"learning_rate": 3.
|
| 493 |
-
"loss":
|
| 494 |
"step": 40500
|
| 495 |
},
|
| 496 |
{
|
| 497 |
-
"epoch":
|
| 498 |
-
"learning_rate": 3.
|
| 499 |
-
"loss":
|
| 500 |
"step": 41000
|
| 501 |
},
|
| 502 |
{
|
| 503 |
-
"epoch":
|
| 504 |
-
"learning_rate": 3.
|
| 505 |
-
"loss":
|
| 506 |
"step": 41500
|
| 507 |
},
|
| 508 |
{
|
| 509 |
-
"epoch":
|
| 510 |
-
"learning_rate": 3.
|
| 511 |
-
"loss":
|
| 512 |
"step": 42000
|
| 513 |
},
|
| 514 |
{
|
| 515 |
-
"epoch":
|
| 516 |
-
"learning_rate": 3.
|
| 517 |
-
"loss":
|
| 518 |
"step": 42500
|
| 519 |
},
|
| 520 |
{
|
| 521 |
-
"epoch":
|
| 522 |
-
"learning_rate": 3.
|
| 523 |
-
"loss":
|
| 524 |
"step": 43000
|
| 525 |
},
|
| 526 |
{
|
| 527 |
-
"epoch":
|
| 528 |
-
"learning_rate": 3.
|
| 529 |
-
"loss":
|
| 530 |
"step": 43500
|
| 531 |
},
|
| 532 |
{
|
| 533 |
-
"epoch":
|
| 534 |
-
"learning_rate": 3.
|
| 535 |
-
"loss":
|
| 536 |
"step": 44000
|
| 537 |
},
|
| 538 |
{
|
| 539 |
-
"epoch":
|
| 540 |
-
"learning_rate": 3.
|
| 541 |
-
"loss":
|
| 542 |
"step": 44500
|
| 543 |
},
|
| 544 |
{
|
| 545 |
-
"epoch":
|
| 546 |
-
"learning_rate": 3.
|
| 547 |
-
"loss":
|
| 548 |
"step": 45000
|
| 549 |
},
|
| 550 |
{
|
| 551 |
-
"epoch":
|
| 552 |
-
"learning_rate": 3.
|
| 553 |
-
"loss":
|
| 554 |
"step": 45500
|
| 555 |
},
|
| 556 |
{
|
| 557 |
-
"epoch":
|
| 558 |
-
"learning_rate": 3.
|
| 559 |
-
"loss":
|
| 560 |
"step": 46000
|
| 561 |
},
|
| 562 |
{
|
| 563 |
-
"epoch":
|
| 564 |
-
"learning_rate": 3.
|
| 565 |
-
"loss":
|
| 566 |
"step": 46500
|
| 567 |
},
|
| 568 |
{
|
| 569 |
-
"epoch":
|
| 570 |
-
"learning_rate": 3.
|
| 571 |
-
"loss":
|
| 572 |
"step": 47000
|
| 573 |
},
|
| 574 |
{
|
| 575 |
-
"epoch":
|
| 576 |
-
"learning_rate": 3.
|
| 577 |
-
"loss":
|
| 578 |
"step": 47500
|
| 579 |
},
|
| 580 |
{
|
| 581 |
-
"epoch":
|
| 582 |
-
"learning_rate": 3.
|
| 583 |
-
"loss":
|
| 584 |
"step": 48000
|
| 585 |
},
|
| 586 |
{
|
| 587 |
-
"epoch":
|
| 588 |
-
"learning_rate": 3.
|
| 589 |
-
"loss":
|
| 590 |
"step": 48500
|
| 591 |
},
|
| 592 |
{
|
| 593 |
-
"epoch":
|
| 594 |
-
"learning_rate": 3.
|
| 595 |
-
"loss":
|
| 596 |
"step": 49000
|
| 597 |
},
|
| 598 |
{
|
| 599 |
-
"epoch":
|
| 600 |
-
"learning_rate": 3.
|
| 601 |
-
"loss":
|
| 602 |
"step": 49500
|
| 603 |
},
|
| 604 |
{
|
| 605 |
-
"epoch":
|
| 606 |
-
"learning_rate": 3.
|
| 607 |
-
"loss":
|
| 608 |
"step": 50000
|
| 609 |
},
|
| 610 |
{
|
| 611 |
-
"epoch":
|
| 612 |
-
"learning_rate": 3.
|
| 613 |
-
"loss":
|
| 614 |
"step": 50500
|
| 615 |
},
|
| 616 |
{
|
| 617 |
-
"epoch":
|
| 618 |
-
"learning_rate": 3.
|
| 619 |
-
"loss":
|
| 620 |
"step": 51000
|
| 621 |
},
|
| 622 |
{
|
| 623 |
-
"epoch":
|
| 624 |
-
"learning_rate": 3.
|
| 625 |
-
"loss":
|
| 626 |
"step": 51500
|
| 627 |
},
|
| 628 |
{
|
| 629 |
-
"epoch":
|
| 630 |
-
"learning_rate": 3.
|
| 631 |
-
"loss":
|
| 632 |
"step": 52000
|
| 633 |
},
|
| 634 |
{
|
| 635 |
-
"epoch":
|
| 636 |
-
"learning_rate": 3.
|
| 637 |
-
"loss":
|
| 638 |
"step": 52500
|
| 639 |
},
|
| 640 |
{
|
| 641 |
-
"epoch":
|
| 642 |
-
"learning_rate": 3.
|
| 643 |
-
"loss":
|
| 644 |
"step": 53000
|
| 645 |
},
|
| 646 |
{
|
| 647 |
-
"epoch":
|
| 648 |
-
"learning_rate": 3.
|
| 649 |
-
"loss":
|
| 650 |
"step": 53500
|
| 651 |
},
|
| 652 |
{
|
| 653 |
-
"epoch":
|
| 654 |
-
"learning_rate":
|
| 655 |
-
"loss":
|
| 656 |
"step": 54000
|
| 657 |
},
|
| 658 |
{
|
| 659 |
-
"epoch":
|
| 660 |
-
"learning_rate":
|
| 661 |
-
"loss":
|
| 662 |
"step": 54500
|
| 663 |
},
|
| 664 |
{
|
| 665 |
-
"epoch":
|
| 666 |
-
"learning_rate":
|
| 667 |
-
"loss":
|
| 668 |
"step": 55000
|
| 669 |
},
|
| 670 |
{
|
| 671 |
-
"epoch":
|
| 672 |
-
"learning_rate":
|
| 673 |
-
"loss":
|
| 674 |
"step": 55500
|
| 675 |
},
|
| 676 |
{
|
| 677 |
-
"epoch":
|
| 678 |
-
"learning_rate":
|
| 679 |
-
"loss":
|
| 680 |
"step": 56000
|
| 681 |
},
|
| 682 |
{
|
| 683 |
-
"epoch":
|
| 684 |
-
"learning_rate":
|
| 685 |
-
"loss":
|
| 686 |
"step": 56500
|
| 687 |
},
|
| 688 |
{
|
| 689 |
-
"epoch":
|
| 690 |
-
"learning_rate":
|
| 691 |
-
"loss":
|
| 692 |
"step": 57000
|
| 693 |
},
|
| 694 |
{
|
| 695 |
-
"epoch":
|
| 696 |
-
"learning_rate":
|
| 697 |
-
"loss":
|
| 698 |
"step": 57500
|
| 699 |
},
|
| 700 |
{
|
| 701 |
-
"epoch":
|
| 702 |
-
"learning_rate":
|
| 703 |
-
"loss":
|
| 704 |
"step": 58000
|
| 705 |
},
|
| 706 |
{
|
| 707 |
-
"epoch":
|
| 708 |
-
"learning_rate":
|
| 709 |
-
"loss":
|
| 710 |
"step": 58500
|
| 711 |
},
|
| 712 |
{
|
| 713 |
-
"epoch":
|
| 714 |
-
"learning_rate":
|
| 715 |
-
"loss":
|
| 716 |
"step": 59000
|
| 717 |
},
|
| 718 |
{
|
| 719 |
-
"epoch":
|
| 720 |
-
"learning_rate":
|
| 721 |
-
"loss":
|
| 722 |
"step": 59500
|
| 723 |
},
|
| 724 |
{
|
| 725 |
-
"epoch":
|
| 726 |
-
"learning_rate":
|
| 727 |
-
"loss":
|
| 728 |
"step": 60000
|
| 729 |
},
|
| 730 |
{
|
| 731 |
-
"epoch":
|
| 732 |
-
"learning_rate":
|
| 733 |
-
"loss":
|
| 734 |
"step": 60500
|
| 735 |
},
|
| 736 |
{
|
| 737 |
-
"epoch":
|
| 738 |
-
"learning_rate":
|
| 739 |
-
"loss":
|
| 740 |
"step": 61000
|
| 741 |
},
|
| 742 |
{
|
| 743 |
-
"epoch":
|
| 744 |
-
"learning_rate":
|
| 745 |
-
"loss":
|
| 746 |
"step": 61500
|
| 747 |
},
|
| 748 |
{
|
| 749 |
-
"epoch":
|
| 750 |
-
"learning_rate":
|
| 751 |
-
"loss":
|
| 752 |
"step": 62000
|
| 753 |
},
|
| 754 |
{
|
| 755 |
-
"epoch":
|
| 756 |
-
"learning_rate":
|
| 757 |
-
"loss":
|
| 758 |
"step": 62500
|
| 759 |
},
|
| 760 |
{
|
| 761 |
-
"epoch":
|
| 762 |
-
"learning_rate":
|
| 763 |
-
"loss":
|
| 764 |
"step": 63000
|
| 765 |
},
|
| 766 |
{
|
| 767 |
-
"epoch":
|
| 768 |
-
"learning_rate":
|
| 769 |
-
"loss": 2.
|
| 770 |
"step": 63500
|
| 771 |
},
|
| 772 |
{
|
| 773 |
-
"epoch":
|
| 774 |
-
"learning_rate":
|
| 775 |
-
"loss":
|
| 776 |
"step": 64000
|
| 777 |
},
|
| 778 |
{
|
| 779 |
-
"epoch":
|
| 780 |
-
"learning_rate":
|
| 781 |
-
"loss":
|
| 782 |
"step": 64500
|
| 783 |
},
|
| 784 |
{
|
| 785 |
-
"epoch":
|
| 786 |
-
"learning_rate":
|
| 787 |
-
"loss": 2.
|
| 788 |
"step": 65000
|
| 789 |
},
|
| 790 |
{
|
| 791 |
-
"epoch":
|
| 792 |
-
"learning_rate":
|
| 793 |
-
"loss":
|
| 794 |
"step": 65500
|
| 795 |
},
|
| 796 |
{
|
| 797 |
-
"epoch":
|
| 798 |
-
"learning_rate":
|
| 799 |
-
"loss": 2.
|
| 800 |
"step": 66000
|
| 801 |
},
|
| 802 |
{
|
| 803 |
-
"epoch":
|
| 804 |
-
"learning_rate":
|
| 805 |
-
"loss": 2.
|
| 806 |
"step": 66500
|
| 807 |
},
|
| 808 |
{
|
| 809 |
-
"epoch":
|
| 810 |
-
"learning_rate":
|
| 811 |
-
"loss": 2.
|
| 812 |
"step": 67000
|
| 813 |
},
|
| 814 |
{
|
| 815 |
-
"epoch":
|
| 816 |
-
"learning_rate": 2.
|
| 817 |
-
"loss": 2.
|
| 818 |
"step": 67500
|
| 819 |
},
|
| 820 |
{
|
| 821 |
-
"epoch":
|
| 822 |
-
"learning_rate": 2.
|
| 823 |
-
"loss":
|
| 824 |
"step": 68000
|
| 825 |
},
|
| 826 |
{
|
| 827 |
-
"epoch":
|
| 828 |
-
"learning_rate": 2.
|
| 829 |
-
"loss":
|
| 830 |
"step": 68500
|
| 831 |
},
|
| 832 |
{
|
| 833 |
-
"epoch":
|
| 834 |
-
"learning_rate": 2.
|
| 835 |
-
"loss":
|
| 836 |
"step": 69000
|
| 837 |
},
|
| 838 |
{
|
| 839 |
-
"epoch":
|
| 840 |
-
"learning_rate": 2.
|
| 841 |
-
"loss":
|
| 842 |
"step": 69500
|
| 843 |
},
|
| 844 |
{
|
| 845 |
-
"epoch":
|
| 846 |
-
"learning_rate": 2.
|
| 847 |
-
"loss":
|
| 848 |
"step": 70000
|
| 849 |
},
|
| 850 |
{
|
| 851 |
-
"epoch":
|
| 852 |
-
"learning_rate": 2.
|
| 853 |
-
"loss":
|
| 854 |
"step": 70500
|
| 855 |
},
|
| 856 |
{
|
| 857 |
-
"epoch": 1
|
| 858 |
-
"learning_rate": 2.
|
| 859 |
-
"loss":
|
| 860 |
"step": 71000
|
| 861 |
},
|
| 862 |
{
|
| 863 |
-
"epoch":
|
| 864 |
-
"learning_rate": 2.
|
| 865 |
-
"loss":
|
| 866 |
"step": 71500
|
| 867 |
},
|
| 868 |
{
|
| 869 |
-
"epoch":
|
| 870 |
-
"learning_rate": 2.
|
| 871 |
-
"loss":
|
| 872 |
"step": 72000
|
| 873 |
},
|
| 874 |
{
|
| 875 |
-
"epoch":
|
| 876 |
-
"learning_rate": 2.
|
| 877 |
-
"loss":
|
| 878 |
"step": 72500
|
| 879 |
},
|
| 880 |
{
|
| 881 |
-
"epoch":
|
| 882 |
-
"learning_rate": 2.
|
| 883 |
-
"loss":
|
| 884 |
"step": 73000
|
| 885 |
},
|
| 886 |
{
|
| 887 |
-
"epoch":
|
| 888 |
-
"learning_rate": 2.
|
| 889 |
-
"loss":
|
| 890 |
"step": 73500
|
| 891 |
},
|
| 892 |
{
|
| 893 |
-
"epoch":
|
| 894 |
-
"learning_rate": 2.
|
| 895 |
-
"loss":
|
| 896 |
"step": 74000
|
| 897 |
},
|
| 898 |
{
|
| 899 |
-
"epoch":
|
| 900 |
-
"learning_rate": 2.
|
| 901 |
-
"loss":
|
| 902 |
"step": 74500
|
| 903 |
},
|
| 904 |
{
|
| 905 |
-
"epoch":
|
| 906 |
-
"learning_rate": 2.
|
| 907 |
-
"loss":
|
| 908 |
"step": 75000
|
| 909 |
},
|
| 910 |
{
|
| 911 |
-
"epoch":
|
| 912 |
-
"learning_rate": 2.
|
| 913 |
-
"loss":
|
| 914 |
"step": 75500
|
| 915 |
},
|
| 916 |
{
|
| 917 |
-
"epoch":
|
| 918 |
-
"learning_rate": 2.
|
| 919 |
-
"loss":
|
| 920 |
"step": 76000
|
| 921 |
},
|
| 922 |
{
|
| 923 |
-
"epoch":
|
| 924 |
-
"learning_rate": 2.
|
| 925 |
-
"loss":
|
| 926 |
"step": 76500
|
| 927 |
},
|
| 928 |
{
|
| 929 |
-
"epoch":
|
| 930 |
-
"learning_rate": 2.
|
| 931 |
-
"loss":
|
| 932 |
"step": 77000
|
| 933 |
},
|
| 934 |
{
|
| 935 |
-
"epoch":
|
| 936 |
-
"learning_rate": 2.
|
| 937 |
-
"loss":
|
| 938 |
"step": 77500
|
| 939 |
},
|
| 940 |
{
|
| 941 |
-
"epoch":
|
| 942 |
-
"learning_rate": 2.
|
| 943 |
-
"loss":
|
| 944 |
"step": 78000
|
| 945 |
},
|
| 946 |
{
|
| 947 |
-
"epoch":
|
| 948 |
-
"learning_rate": 2.
|
| 949 |
-
"loss":
|
| 950 |
"step": 78500
|
| 951 |
},
|
| 952 |
{
|
| 953 |
-
"epoch":
|
| 954 |
-
"learning_rate": 2.
|
| 955 |
-
"loss":
|
| 956 |
"step": 79000
|
| 957 |
},
|
| 958 |
{
|
| 959 |
-
"epoch":
|
| 960 |
-
"learning_rate": 2.
|
| 961 |
-
"loss":
|
| 962 |
"step": 79500
|
| 963 |
},
|
| 964 |
{
|
| 965 |
-
"epoch":
|
| 966 |
-
"learning_rate": 2.
|
| 967 |
-
"loss":
|
| 968 |
"step": 80000
|
| 969 |
},
|
| 970 |
{
|
| 971 |
-
"epoch":
|
| 972 |
-
"learning_rate": 2.
|
| 973 |
-
"loss":
|
| 974 |
"step": 80500
|
| 975 |
},
|
| 976 |
{
|
| 977 |
-
"epoch":
|
| 978 |
-
"learning_rate":
|
| 979 |
-
"loss":
|
| 980 |
"step": 81000
|
| 981 |
},
|
| 982 |
{
|
| 983 |
-
"epoch":
|
| 984 |
-
"learning_rate":
|
| 985 |
-
"loss":
|
| 986 |
"step": 81500
|
| 987 |
},
|
| 988 |
{
|
| 989 |
-
"epoch":
|
| 990 |
-
"learning_rate":
|
| 991 |
-
"loss":
|
| 992 |
"step": 82000
|
| 993 |
},
|
| 994 |
{
|
| 995 |
-
"epoch":
|
| 996 |
-
"learning_rate":
|
| 997 |
-
"loss":
|
| 998 |
"step": 82500
|
| 999 |
},
|
| 1000 |
{
|
| 1001 |
-
"epoch":
|
| 1002 |
-
"learning_rate":
|
| 1003 |
-
"loss":
|
| 1004 |
"step": 83000
|
| 1005 |
},
|
| 1006 |
{
|
| 1007 |
-
"epoch":
|
| 1008 |
-
"learning_rate":
|
| 1009 |
-
"loss":
|
| 1010 |
"step": 83500
|
| 1011 |
},
|
| 1012 |
{
|
| 1013 |
-
"epoch":
|
| 1014 |
-
"learning_rate":
|
| 1015 |
-
"loss":
|
| 1016 |
"step": 84000
|
| 1017 |
},
|
| 1018 |
{
|
| 1019 |
-
"epoch":
|
| 1020 |
-
"learning_rate":
|
| 1021 |
-
"loss":
|
| 1022 |
"step": 84500
|
| 1023 |
},
|
| 1024 |
{
|
| 1025 |
-
"epoch":
|
| 1026 |
-
"learning_rate":
|
| 1027 |
-
"loss":
|
| 1028 |
"step": 85000
|
| 1029 |
},
|
| 1030 |
{
|
| 1031 |
-
"epoch":
|
| 1032 |
-
"learning_rate":
|
| 1033 |
-
"loss":
|
| 1034 |
"step": 85500
|
| 1035 |
},
|
| 1036 |
{
|
| 1037 |
-
"epoch":
|
| 1038 |
-
"learning_rate":
|
| 1039 |
-
"loss":
|
| 1040 |
"step": 86000
|
| 1041 |
},
|
| 1042 |
{
|
| 1043 |
-
"epoch":
|
| 1044 |
-
"learning_rate":
|
| 1045 |
-
"loss":
|
| 1046 |
"step": 86500
|
| 1047 |
},
|
| 1048 |
{
|
| 1049 |
-
"epoch":
|
| 1050 |
-
"learning_rate":
|
| 1051 |
-
"loss":
|
| 1052 |
"step": 87000
|
| 1053 |
},
|
| 1054 |
{
|
| 1055 |
-
"epoch":
|
| 1056 |
-
"learning_rate":
|
| 1057 |
-
"loss":
|
| 1058 |
"step": 87500
|
| 1059 |
},
|
| 1060 |
{
|
| 1061 |
-
"epoch":
|
| 1062 |
-
"learning_rate":
|
| 1063 |
-
"loss":
|
| 1064 |
"step": 88000
|
| 1065 |
},
|
| 1066 |
{
|
| 1067 |
-
"epoch":
|
| 1068 |
-
"learning_rate":
|
| 1069 |
-
"loss":
|
| 1070 |
"step": 88500
|
| 1071 |
},
|
| 1072 |
{
|
| 1073 |
-
"epoch":
|
| 1074 |
-
"learning_rate":
|
| 1075 |
-
"loss":
|
| 1076 |
"step": 89000
|
| 1077 |
},
|
| 1078 |
{
|
| 1079 |
-
"epoch":
|
| 1080 |
-
"learning_rate":
|
| 1081 |
-
"loss":
|
| 1082 |
"step": 89500
|
| 1083 |
},
|
| 1084 |
{
|
| 1085 |
-
"epoch":
|
| 1086 |
-
"learning_rate":
|
| 1087 |
-
"loss":
|
| 1088 |
"step": 90000
|
| 1089 |
},
|
| 1090 |
{
|
| 1091 |
-
"epoch":
|
| 1092 |
-
"learning_rate":
|
| 1093 |
-
"loss":
|
| 1094 |
"step": 90500
|
| 1095 |
},
|
| 1096 |
{
|
| 1097 |
-
"epoch":
|
| 1098 |
-
"learning_rate":
|
| 1099 |
-
"loss":
|
| 1100 |
"step": 91000
|
| 1101 |
},
|
| 1102 |
{
|
| 1103 |
-
"epoch":
|
| 1104 |
-
"learning_rate":
|
| 1105 |
-
"loss":
|
| 1106 |
"step": 91500
|
| 1107 |
},
|
| 1108 |
{
|
| 1109 |
-
"epoch":
|
| 1110 |
-
"learning_rate":
|
| 1111 |
-
"loss":
|
| 1112 |
"step": 92000
|
| 1113 |
},
|
| 1114 |
{
|
| 1115 |
-
"epoch":
|
| 1116 |
-
"learning_rate":
|
| 1117 |
-
"loss":
|
| 1118 |
"step": 92500
|
| 1119 |
},
|
| 1120 |
{
|
| 1121 |
-
"epoch":
|
| 1122 |
-
"learning_rate":
|
| 1123 |
-
"loss":
|
| 1124 |
"step": 93000
|
| 1125 |
},
|
| 1126 |
{
|
| 1127 |
-
"epoch":
|
| 1128 |
-
"learning_rate":
|
| 1129 |
-
"loss":
|
| 1130 |
"step": 93500
|
| 1131 |
},
|
| 1132 |
{
|
| 1133 |
-
"epoch":
|
| 1134 |
-
"learning_rate":
|
| 1135 |
-
"loss":
|
| 1136 |
"step": 94000
|
| 1137 |
},
|
| 1138 |
{
|
| 1139 |
-
"epoch":
|
| 1140 |
-
"learning_rate":
|
| 1141 |
-
"loss":
|
| 1142 |
"step": 94500
|
| 1143 |
},
|
| 1144 |
{
|
| 1145 |
-
"epoch":
|
| 1146 |
-
"learning_rate":
|
| 1147 |
-
"loss":
|
| 1148 |
"step": 95000
|
| 1149 |
},
|
| 1150 |
{
|
| 1151 |
-
"epoch":
|
| 1152 |
-
"learning_rate":
|
| 1153 |
-
"loss":
|
| 1154 |
"step": 95500
|
| 1155 |
},
|
| 1156 |
{
|
| 1157 |
-
"epoch":
|
| 1158 |
-
"learning_rate":
|
| 1159 |
-
"loss":
|
| 1160 |
"step": 96000
|
| 1161 |
},
|
| 1162 |
{
|
| 1163 |
-
"epoch":
|
| 1164 |
-
"learning_rate":
|
| 1165 |
-
"loss":
|
| 1166 |
"step": 96500
|
| 1167 |
},
|
| 1168 |
{
|
| 1169 |
-
"epoch":
|
| 1170 |
-
"learning_rate":
|
| 1171 |
-
"loss":
|
| 1172 |
"step": 97000
|
| 1173 |
},
|
| 1174 |
{
|
| 1175 |
-
"epoch":
|
| 1176 |
-
"learning_rate":
|
| 1177 |
-
"loss":
|
| 1178 |
"step": 97500
|
| 1179 |
},
|
| 1180 |
{
|
| 1181 |
-
"epoch":
|
| 1182 |
-
"learning_rate":
|
| 1183 |
-
"loss":
|
| 1184 |
"step": 98000
|
| 1185 |
},
|
| 1186 |
{
|
| 1187 |
-
"epoch":
|
| 1188 |
-
"learning_rate":
|
| 1189 |
-
"loss":
|
| 1190 |
"step": 98500
|
| 1191 |
},
|
| 1192 |
{
|
| 1193 |
-
"epoch":
|
| 1194 |
-
"learning_rate":
|
| 1195 |
-
"loss":
|
| 1196 |
"step": 99000
|
| 1197 |
},
|
| 1198 |
{
|
| 1199 |
-
"epoch":
|
| 1200 |
-
"learning_rate":
|
| 1201 |
-
"loss":
|
| 1202 |
"step": 99500
|
| 1203 |
},
|
| 1204 |
{
|
| 1205 |
-
"epoch":
|
| 1206 |
-
"learning_rate":
|
| 1207 |
-
"loss":
|
| 1208 |
"step": 100000
|
| 1209 |
},
|
| 1210 |
{
|
| 1211 |
-
"epoch":
|
| 1212 |
-
"learning_rate":
|
| 1213 |
-
"loss":
|
| 1214 |
"step": 100500
|
| 1215 |
},
|
| 1216 |
{
|
| 1217 |
-
"epoch":
|
| 1218 |
-
"learning_rate":
|
| 1219 |
-
"loss":
|
| 1220 |
"step": 101000
|
| 1221 |
},
|
| 1222 |
{
|
| 1223 |
-
"epoch":
|
| 1224 |
-
"learning_rate": 1.
|
| 1225 |
-
"loss":
|
| 1226 |
"step": 101500
|
| 1227 |
},
|
| 1228 |
{
|
| 1229 |
-
"epoch":
|
| 1230 |
-
"learning_rate": 1.
|
| 1231 |
-
"loss":
|
| 1232 |
"step": 102000
|
| 1233 |
},
|
| 1234 |
{
|
| 1235 |
-
"epoch":
|
| 1236 |
-
"learning_rate": 1.
|
| 1237 |
-
"loss":
|
| 1238 |
"step": 102500
|
| 1239 |
},
|
| 1240 |
{
|
| 1241 |
-
"epoch":
|
| 1242 |
-
"learning_rate": 1.
|
| 1243 |
-
"loss":
|
| 1244 |
"step": 103000
|
| 1245 |
},
|
| 1246 |
{
|
| 1247 |
-
"epoch":
|
| 1248 |
-
"learning_rate": 1.
|
| 1249 |
-
"loss":
|
| 1250 |
"step": 103500
|
| 1251 |
},
|
| 1252 |
{
|
| 1253 |
-
"epoch":
|
| 1254 |
-
"learning_rate": 1.
|
| 1255 |
-
"loss":
|
| 1256 |
"step": 104000
|
| 1257 |
},
|
| 1258 |
{
|
| 1259 |
-
"epoch": 1
|
| 1260 |
-
"learning_rate": 1.
|
| 1261 |
-
"loss":
|
| 1262 |
"step": 104500
|
| 1263 |
},
|
| 1264 |
{
|
| 1265 |
-
"epoch":
|
| 1266 |
-
"learning_rate": 1.
|
| 1267 |
-
"loss":
|
| 1268 |
"step": 105000
|
| 1269 |
},
|
| 1270 |
{
|
| 1271 |
-
"epoch":
|
| 1272 |
-
"learning_rate": 1.
|
| 1273 |
-
"loss":
|
| 1274 |
"step": 105500
|
| 1275 |
},
|
| 1276 |
{
|
| 1277 |
-
"epoch":
|
| 1278 |
-
"learning_rate": 1.
|
| 1279 |
-
"loss":
|
| 1280 |
"step": 106000
|
| 1281 |
},
|
| 1282 |
{
|
| 1283 |
-
"epoch":
|
| 1284 |
-
"learning_rate": 1.
|
| 1285 |
-
"loss":
|
| 1286 |
"step": 106500
|
| 1287 |
},
|
| 1288 |
{
|
| 1289 |
-
"epoch":
|
| 1290 |
-
"learning_rate": 1.
|
| 1291 |
-
"loss":
|
| 1292 |
"step": 107000
|
| 1293 |
},
|
| 1294 |
{
|
| 1295 |
-
"epoch":
|
| 1296 |
-
"learning_rate": 1.
|
| 1297 |
-
"loss":
|
| 1298 |
"step": 107500
|
| 1299 |
},
|
| 1300 |
{
|
| 1301 |
-
"epoch":
|
| 1302 |
-
"learning_rate":
|
| 1303 |
-
"loss":
|
| 1304 |
"step": 108000
|
| 1305 |
},
|
| 1306 |
{
|
| 1307 |
-
"epoch":
|
| 1308 |
-
"learning_rate":
|
| 1309 |
-
"loss":
|
| 1310 |
"step": 108500
|
| 1311 |
},
|
| 1312 |
{
|
| 1313 |
-
"epoch":
|
| 1314 |
-
"learning_rate":
|
| 1315 |
-
"loss":
|
| 1316 |
"step": 109000
|
| 1317 |
},
|
| 1318 |
{
|
| 1319 |
-
"epoch":
|
| 1320 |
-
"learning_rate":
|
| 1321 |
-
"loss":
|
| 1322 |
"step": 109500
|
| 1323 |
},
|
| 1324 |
{
|
| 1325 |
-
"epoch":
|
| 1326 |
-
"learning_rate":
|
| 1327 |
-
"loss":
|
| 1328 |
"step": 110000
|
| 1329 |
},
|
| 1330 |
{
|
| 1331 |
-
"epoch":
|
| 1332 |
-
"learning_rate":
|
| 1333 |
-
"loss":
|
| 1334 |
"step": 110500
|
| 1335 |
},
|
| 1336 |
{
|
| 1337 |
-
"epoch":
|
| 1338 |
-
"learning_rate":
|
| 1339 |
-
"loss":
|
| 1340 |
"step": 111000
|
| 1341 |
},
|
| 1342 |
{
|
| 1343 |
-
"epoch":
|
| 1344 |
-
"learning_rate":
|
| 1345 |
-
"loss":
|
| 1346 |
"step": 111500
|
| 1347 |
},
|
| 1348 |
{
|
| 1349 |
-
"epoch":
|
| 1350 |
-
"learning_rate":
|
| 1351 |
-
"loss":
|
| 1352 |
"step": 112000
|
| 1353 |
},
|
| 1354 |
{
|
| 1355 |
-
"epoch":
|
| 1356 |
-
"learning_rate":
|
| 1357 |
-
"loss":
|
| 1358 |
"step": 112500
|
| 1359 |
},
|
| 1360 |
{
|
| 1361 |
-
"epoch":
|
| 1362 |
-
"learning_rate":
|
| 1363 |
-
"loss":
|
| 1364 |
"step": 113000
|
| 1365 |
},
|
| 1366 |
{
|
| 1367 |
-
"epoch":
|
| 1368 |
-
"learning_rate":
|
| 1369 |
-
"loss":
|
| 1370 |
"step": 113500
|
| 1371 |
},
|
| 1372 |
{
|
| 1373 |
-
"epoch":
|
| 1374 |
-
"learning_rate":
|
| 1375 |
-
"loss":
|
| 1376 |
"step": 114000
|
| 1377 |
},
|
| 1378 |
{
|
| 1379 |
-
"epoch":
|
| 1380 |
-
"learning_rate":
|
| 1381 |
-
"loss":
|
| 1382 |
"step": 114500
|
| 1383 |
},
|
| 1384 |
{
|
| 1385 |
-
"epoch":
|
| 1386 |
-
"learning_rate":
|
| 1387 |
-
"loss":
|
| 1388 |
"step": 115000
|
| 1389 |
},
|
| 1390 |
{
|
| 1391 |
-
"epoch":
|
| 1392 |
-
"learning_rate":
|
| 1393 |
-
"loss":
|
| 1394 |
"step": 115500
|
| 1395 |
},
|
| 1396 |
{
|
| 1397 |
-
"epoch":
|
| 1398 |
-
"learning_rate":
|
| 1399 |
-
"loss":
|
| 1400 |
"step": 116000
|
| 1401 |
},
|
| 1402 |
{
|
| 1403 |
-
"epoch":
|
| 1404 |
-
"learning_rate":
|
| 1405 |
-
"loss":
|
| 1406 |
"step": 116500
|
| 1407 |
},
|
| 1408 |
{
|
| 1409 |
-
"epoch":
|
| 1410 |
-
"learning_rate":
|
| 1411 |
-
"loss":
|
| 1412 |
"step": 117000
|
| 1413 |
},
|
| 1414 |
{
|
| 1415 |
-
"epoch":
|
| 1416 |
-
"learning_rate":
|
| 1417 |
-
"loss":
|
| 1418 |
"step": 117500
|
| 1419 |
},
|
| 1420 |
{
|
| 1421 |
-
"epoch":
|
| 1422 |
-
"learning_rate":
|
| 1423 |
-
"loss":
|
| 1424 |
"step": 118000
|
| 1425 |
},
|
| 1426 |
{
|
| 1427 |
-
"epoch":
|
| 1428 |
-
"learning_rate":
|
| 1429 |
-
"loss":
|
| 1430 |
"step": 118500
|
| 1431 |
},
|
| 1432 |
{
|
| 1433 |
-
"epoch":
|
| 1434 |
-
"learning_rate":
|
| 1435 |
-
"loss":
|
| 1436 |
"step": 119000
|
| 1437 |
},
|
| 1438 |
{
|
| 1439 |
-
"epoch":
|
| 1440 |
-
"learning_rate":
|
| 1441 |
-
"loss":
|
| 1442 |
"step": 119500
|
| 1443 |
},
|
| 1444 |
{
|
| 1445 |
-
"epoch":
|
| 1446 |
-
"learning_rate":
|
| 1447 |
-
"loss":
|
| 1448 |
"step": 120000
|
| 1449 |
},
|
| 1450 |
{
|
| 1451 |
-
"epoch":
|
| 1452 |
-
"learning_rate":
|
| 1453 |
-
"loss":
|
| 1454 |
"step": 120500
|
| 1455 |
},
|
| 1456 |
{
|
| 1457 |
-
"epoch":
|
| 1458 |
-
"learning_rate":
|
| 1459 |
-
"loss":
|
| 1460 |
"step": 121000
|
| 1461 |
},
|
| 1462 |
{
|
| 1463 |
-
"epoch":
|
| 1464 |
-
"learning_rate":
|
| 1465 |
-
"loss":
|
| 1466 |
"step": 121500
|
| 1467 |
},
|
| 1468 |
{
|
| 1469 |
-
"epoch":
|
| 1470 |
-
"learning_rate":
|
| 1471 |
-
"loss":
|
| 1472 |
"step": 122000
|
| 1473 |
},
|
| 1474 |
{
|
| 1475 |
-
"epoch":
|
| 1476 |
-
"learning_rate":
|
| 1477 |
-
"loss":
|
| 1478 |
"step": 122500
|
| 1479 |
},
|
| 1480 |
{
|
| 1481 |
-
"epoch":
|
| 1482 |
-
"learning_rate":
|
| 1483 |
-
"loss":
|
| 1484 |
"step": 123000
|
| 1485 |
},
|
| 1486 |
{
|
| 1487 |
-
"epoch":
|
| 1488 |
-
"learning_rate":
|
| 1489 |
-
"loss":
|
| 1490 |
"step": 123500
|
| 1491 |
},
|
| 1492 |
{
|
| 1493 |
-
"epoch":
|
| 1494 |
-
"learning_rate":
|
| 1495 |
-
"loss":
|
| 1496 |
"step": 124000
|
| 1497 |
},
|
| 1498 |
{
|
| 1499 |
-
"epoch":
|
| 1500 |
-
"learning_rate":
|
| 1501 |
-
"loss":
|
| 1502 |
"step": 124500
|
| 1503 |
},
|
| 1504 |
{
|
| 1505 |
-
"epoch":
|
| 1506 |
-
"learning_rate":
|
| 1507 |
-
"loss":
|
| 1508 |
"step": 125000
|
| 1509 |
},
|
| 1510 |
{
|
| 1511 |
-
"epoch":
|
| 1512 |
-
"learning_rate":
|
| 1513 |
-
"loss":
|
| 1514 |
"step": 125500
|
| 1515 |
},
|
| 1516 |
{
|
| 1517 |
-
"epoch":
|
| 1518 |
-
"learning_rate":
|
| 1519 |
-
"loss":
|
| 1520 |
"step": 126000
|
| 1521 |
},
|
| 1522 |
{
|
| 1523 |
-
"epoch":
|
| 1524 |
-
"learning_rate":
|
| 1525 |
-
"loss":
|
| 1526 |
"step": 126500
|
| 1527 |
},
|
| 1528 |
{
|
| 1529 |
-
"epoch":
|
| 1530 |
-
"learning_rate":
|
| 1531 |
-
"loss":
|
| 1532 |
"step": 127000
|
| 1533 |
},
|
| 1534 |
{
|
| 1535 |
-
"epoch":
|
| 1536 |
-
"learning_rate":
|
| 1537 |
-
"loss":
|
| 1538 |
"step": 127500
|
| 1539 |
},
|
| 1540 |
{
|
| 1541 |
-
"epoch":
|
| 1542 |
-
"learning_rate":
|
| 1543 |
-
"loss":
|
| 1544 |
"step": 128000
|
| 1545 |
},
|
| 1546 |
{
|
| 1547 |
-
"epoch":
|
| 1548 |
-
"learning_rate":
|
| 1549 |
-
"loss":
|
| 1550 |
"step": 128500
|
| 1551 |
},
|
| 1552 |
{
|
| 1553 |
-
"epoch":
|
| 1554 |
-
"learning_rate":
|
| 1555 |
-
"loss":
|
| 1556 |
"step": 129000
|
| 1557 |
},
|
| 1558 |
{
|
| 1559 |
-
"epoch":
|
| 1560 |
-
"learning_rate":
|
| 1561 |
-
"loss":
|
| 1562 |
"step": 129500
|
| 1563 |
},
|
| 1564 |
{
|
| 1565 |
-
"epoch":
|
| 1566 |
-
"learning_rate": 1.
|
| 1567 |
-
"loss":
|
| 1568 |
"step": 130000
|
| 1569 |
},
|
| 1570 |
{
|
| 1571 |
-
"epoch":
|
| 1572 |
-
"learning_rate": 1.
|
| 1573 |
-
"loss":
|
| 1574 |
"step": 130500
|
| 1575 |
},
|
| 1576 |
{
|
| 1577 |
-
"epoch":
|
| 1578 |
-
"learning_rate": 1.
|
| 1579 |
-
"loss":
|
| 1580 |
"step": 131000
|
| 1581 |
},
|
| 1582 |
{
|
| 1583 |
-
"epoch":
|
| 1584 |
-
"learning_rate": 1.
|
| 1585 |
-
"loss":
|
| 1586 |
"step": 131500
|
| 1587 |
},
|
| 1588 |
{
|
| 1589 |
-
"epoch":
|
| 1590 |
-
"learning_rate": 1.
|
| 1591 |
-
"loss":
|
| 1592 |
"step": 132000
|
| 1593 |
},
|
| 1594 |
{
|
| 1595 |
-
"epoch":
|
| 1596 |
-
"learning_rate":
|
| 1597 |
-
"loss":
|
| 1598 |
"step": 132500
|
| 1599 |
},
|
| 1600 |
{
|
| 1601 |
-
"epoch":
|
| 1602 |
-
"learning_rate":
|
| 1603 |
-
"loss":
|
| 1604 |
"step": 133000
|
| 1605 |
},
|
| 1606 |
{
|
| 1607 |
-
"epoch":
|
| 1608 |
-
"learning_rate":
|
| 1609 |
-
"loss":
|
| 1610 |
"step": 133500
|
| 1611 |
},
|
| 1612 |
{
|
| 1613 |
-
"epoch":
|
| 1614 |
-
"learning_rate":
|
| 1615 |
-
"loss":
|
| 1616 |
"step": 134000
|
| 1617 |
},
|
| 1618 |
{
|
| 1619 |
-
"epoch":
|
| 1620 |
-
"learning_rate": 1.
|
| 1621 |
-
"loss":
|
| 1622 |
"step": 134500
|
| 1623 |
},
|
| 1624 |
{
|
| 1625 |
-
"epoch":
|
| 1626 |
-
"
|
| 1627 |
-
"
|
| 1628 |
-
"
|
| 1629 |
-
|
| 1630 |
-
|
| 1631 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1632 |
}
|
| 1633 |
],
|
| 1634 |
-
"max_steps":
|
| 1635 |
-
"num_train_epochs":
|
| 1636 |
-
"total_flos":
|
| 1637 |
"trial_name": null,
|
| 1638 |
"trial_params": null
|
| 1639 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 5.0,
|
| 5 |
+
"global_step": 168680,
|
| 6 |
"is_hyper_param_search": false,
|
| 7 |
"is_local_process_zero": true,
|
| 8 |
"is_world_process_zero": true,
|
| 9 |
"log_history": [
|
| 10 |
{
|
| 11 |
"epoch": 0.01,
|
| 12 |
+
"learning_rate": 4.985179037230259e-05,
|
| 13 |
+
"loss": 1.9823,
|
| 14 |
"step": 500
|
| 15 |
},
|
| 16 |
{
|
| 17 |
+
"epoch": 0.03,
|
| 18 |
+
"learning_rate": 4.970358074460517e-05,
|
| 19 |
+
"loss": 2.0137,
|
| 20 |
"step": 1000
|
| 21 |
},
|
| 22 |
{
|
| 23 |
+
"epoch": 0.04,
|
| 24 |
+
"learning_rate": 4.9555371116907756e-05,
|
| 25 |
+
"loss": 2.0097,
|
| 26 |
"step": 1500
|
| 27 |
},
|
| 28 |
{
|
| 29 |
+
"epoch": 0.06,
|
| 30 |
+
"learning_rate": 4.940716148921034e-05,
|
| 31 |
+
"loss": 2.0074,
|
| 32 |
"step": 2000
|
| 33 |
},
|
| 34 |
{
|
| 35 |
+
"epoch": 0.07,
|
| 36 |
+
"learning_rate": 4.9258951861512924e-05,
|
| 37 |
+
"loss": 2.0169,
|
| 38 |
"step": 2500
|
| 39 |
},
|
| 40 |
{
|
| 41 |
+
"epoch": 0.09,
|
| 42 |
+
"learning_rate": 4.911074223381551e-05,
|
| 43 |
+
"loss": 2.0187,
|
| 44 |
"step": 3000
|
| 45 |
},
|
| 46 |
{
|
| 47 |
+
"epoch": 0.1,
|
| 48 |
+
"learning_rate": 4.896253260611809e-05,
|
| 49 |
+
"loss": 2.0182,
|
| 50 |
"step": 3500
|
| 51 |
},
|
| 52 |
{
|
| 53 |
+
"epoch": 0.12,
|
| 54 |
+
"learning_rate": 4.881432297842068e-05,
|
| 55 |
+
"loss": 2.035,
|
| 56 |
"step": 4000
|
| 57 |
},
|
| 58 |
{
|
| 59 |
+
"epoch": 0.13,
|
| 60 |
+
"learning_rate": 4.866611335072326e-05,
|
| 61 |
+
"loss": 2.0271,
|
| 62 |
"step": 4500
|
| 63 |
},
|
| 64 |
{
|
| 65 |
+
"epoch": 0.15,
|
| 66 |
+
"learning_rate": 4.8517903723025847e-05,
|
| 67 |
+
"loss": 2.0089,
|
| 68 |
"step": 5000
|
| 69 |
},
|
| 70 |
{
|
| 71 |
+
"epoch": 0.16,
|
| 72 |
+
"learning_rate": 4.836969409532843e-05,
|
| 73 |
+
"loss": 2.0152,
|
| 74 |
"step": 5500
|
| 75 |
},
|
| 76 |
{
|
| 77 |
+
"epoch": 0.18,
|
| 78 |
+
"learning_rate": 4.8221484467631015e-05,
|
| 79 |
+
"loss": 2.0181,
|
| 80 |
"step": 6000
|
| 81 |
},
|
| 82 |
{
|
| 83 |
+
"epoch": 0.19,
|
| 84 |
+
"learning_rate": 4.80732748399336e-05,
|
| 85 |
+
"loss": 2.0154,
|
| 86 |
"step": 6500
|
| 87 |
},
|
| 88 |
{
|
| 89 |
+
"epoch": 0.21,
|
| 90 |
+
"learning_rate": 4.7925065212236184e-05,
|
| 91 |
+
"loss": 2.026,
|
| 92 |
"step": 7000
|
| 93 |
},
|
| 94 |
{
|
| 95 |
+
"epoch": 0.22,
|
| 96 |
+
"learning_rate": 4.7776855584538775e-05,
|
| 97 |
+
"loss": 1.9978,
|
| 98 |
"step": 7500
|
| 99 |
},
|
| 100 |
{
|
| 101 |
+
"epoch": 0.24,
|
| 102 |
+
"learning_rate": 4.762864595684136e-05,
|
| 103 |
+
"loss": 2.0187,
|
| 104 |
"step": 8000
|
| 105 |
},
|
| 106 |
{
|
| 107 |
+
"epoch": 0.25,
|
| 108 |
+
"learning_rate": 4.7480436329143944e-05,
|
| 109 |
+
"loss": 2.0338,
|
| 110 |
"step": 8500
|
| 111 |
},
|
| 112 |
{
|
| 113 |
+
"epoch": 0.27,
|
| 114 |
+
"learning_rate": 4.733222670144653e-05,
|
| 115 |
+
"loss": 2.0258,
|
| 116 |
"step": 9000
|
| 117 |
},
|
| 118 |
{
|
| 119 |
+
"epoch": 0.28,
|
| 120 |
+
"learning_rate": 4.718401707374911e-05,
|
| 121 |
+
"loss": 2.0205,
|
| 122 |
"step": 9500
|
| 123 |
},
|
| 124 |
{
|
| 125 |
+
"epoch": 0.3,
|
| 126 |
+
"learning_rate": 4.70358074460517e-05,
|
| 127 |
+
"loss": 2.0218,
|
| 128 |
"step": 10000
|
| 129 |
},
|
| 130 |
{
|
| 131 |
+
"epoch": 0.31,
|
| 132 |
+
"learning_rate": 4.688759781835428e-05,
|
| 133 |
+
"loss": 2.0151,
|
| 134 |
"step": 10500
|
| 135 |
},
|
| 136 |
{
|
| 137 |
+
"epoch": 0.33,
|
| 138 |
+
"learning_rate": 4.6739388190656866e-05,
|
| 139 |
+
"loss": 2.0203,
|
| 140 |
"step": 11000
|
| 141 |
},
|
| 142 |
{
|
| 143 |
+
"epoch": 0.34,
|
| 144 |
+
"learning_rate": 4.659117856295945e-05,
|
| 145 |
+
"loss": 2.0439,
|
| 146 |
"step": 11500
|
| 147 |
},
|
| 148 |
{
|
| 149 |
+
"epoch": 0.36,
|
| 150 |
+
"learning_rate": 4.6442968935262035e-05,
|
| 151 |
+
"loss": 2.0389,
|
| 152 |
"step": 12000
|
| 153 |
},
|
| 154 |
{
|
| 155 |
+
"epoch": 0.37,
|
| 156 |
+
"learning_rate": 4.629475930756462e-05,
|
| 157 |
+
"loss": 2.019,
|
| 158 |
"step": 12500
|
| 159 |
},
|
| 160 |
{
|
| 161 |
+
"epoch": 0.39,
|
| 162 |
+
"learning_rate": 4.6146549679867204e-05,
|
| 163 |
+
"loss": 2.0375,
|
| 164 |
"step": 13000
|
| 165 |
},
|
| 166 |
{
|
| 167 |
+
"epoch": 0.4,
|
| 168 |
+
"learning_rate": 4.599834005216979e-05,
|
| 169 |
+
"loss": 2.0539,
|
| 170 |
"step": 13500
|
| 171 |
},
|
| 172 |
{
|
| 173 |
+
"epoch": 0.41,
|
| 174 |
+
"learning_rate": 4.585013042447238e-05,
|
| 175 |
+
"loss": 2.0415,
|
| 176 |
"step": 14000
|
| 177 |
},
|
| 178 |
{
|
| 179 |
+
"epoch": 0.43,
|
| 180 |
+
"learning_rate": 4.5701920796774964e-05,
|
| 181 |
+
"loss": 2.0262,
|
| 182 |
"step": 14500
|
| 183 |
},
|
| 184 |
{
|
| 185 |
+
"epoch": 0.44,
|
| 186 |
+
"learning_rate": 4.555371116907755e-05,
|
| 187 |
+
"loss": 2.0349,
|
| 188 |
"step": 15000
|
| 189 |
},
|
| 190 |
{
|
| 191 |
+
"epoch": 0.46,
|
| 192 |
+
"learning_rate": 4.540550154138013e-05,
|
| 193 |
+
"loss": 2.0392,
|
| 194 |
"step": 15500
|
| 195 |
},
|
| 196 |
{
|
| 197 |
+
"epoch": 0.47,
|
| 198 |
+
"learning_rate": 4.525729191368272e-05,
|
| 199 |
+
"loss": 2.037,
|
| 200 |
"step": 16000
|
| 201 |
},
|
| 202 |
{
|
| 203 |
+
"epoch": 0.49,
|
| 204 |
+
"learning_rate": 4.51090822859853e-05,
|
| 205 |
+
"loss": 2.0583,
|
| 206 |
"step": 16500
|
| 207 |
},
|
| 208 |
{
|
| 209 |
+
"epoch": 0.5,
|
| 210 |
+
"learning_rate": 4.4960872658287886e-05,
|
| 211 |
+
"loss": 2.0516,
|
| 212 |
"step": 17000
|
| 213 |
},
|
| 214 |
{
|
| 215 |
+
"epoch": 0.52,
|
| 216 |
+
"learning_rate": 4.481266303059047e-05,
|
| 217 |
+
"loss": 2.0336,
|
| 218 |
"step": 17500
|
| 219 |
},
|
| 220 |
{
|
| 221 |
+
"epoch": 0.53,
|
| 222 |
+
"learning_rate": 4.4664453402893055e-05,
|
| 223 |
+
"loss": 2.0478,
|
| 224 |
"step": 18000
|
| 225 |
},
|
| 226 |
{
|
| 227 |
+
"epoch": 0.55,
|
| 228 |
+
"learning_rate": 4.451624377519564e-05,
|
| 229 |
+
"loss": 2.0305,
|
| 230 |
"step": 18500
|
| 231 |
},
|
| 232 |
{
|
| 233 |
+
"epoch": 0.56,
|
| 234 |
+
"learning_rate": 4.4368034147498224e-05,
|
| 235 |
+
"loss": 2.0388,
|
| 236 |
"step": 19000
|
| 237 |
},
|
| 238 |
{
|
| 239 |
+
"epoch": 0.58,
|
| 240 |
+
"learning_rate": 4.421982451980081e-05,
|
| 241 |
+
"loss": 2.0297,
|
| 242 |
"step": 19500
|
| 243 |
},
|
| 244 |
{
|
| 245 |
+
"epoch": 0.59,
|
| 246 |
+
"learning_rate": 4.407161489210339e-05,
|
| 247 |
+
"loss": 2.0473,
|
| 248 |
"step": 20000
|
| 249 |
},
|
| 250 |
{
|
| 251 |
+
"epoch": 0.61,
|
| 252 |
+
"learning_rate": 4.392340526440598e-05,
|
| 253 |
+
"loss": 2.0396,
|
| 254 |
"step": 20500
|
| 255 |
},
|
| 256 |
{
|
| 257 |
+
"epoch": 0.62,
|
| 258 |
+
"learning_rate": 4.377519563670856e-05,
|
| 259 |
+
"loss": 2.0398,
|
| 260 |
"step": 21000
|
| 261 |
},
|
| 262 |
{
|
| 263 |
+
"epoch": 0.64,
|
| 264 |
+
"learning_rate": 4.3626986009011146e-05,
|
| 265 |
+
"loss": 2.0358,
|
| 266 |
"step": 21500
|
| 267 |
},
|
| 268 |
{
|
| 269 |
+
"epoch": 0.65,
|
| 270 |
+
"learning_rate": 4.347877638131374e-05,
|
| 271 |
+
"loss": 2.0565,
|
| 272 |
"step": 22000
|
| 273 |
},
|
| 274 |
{
|
| 275 |
+
"epoch": 0.67,
|
| 276 |
+
"learning_rate": 4.333056675361632e-05,
|
| 277 |
+
"loss": 2.0321,
|
| 278 |
"step": 22500
|
| 279 |
},
|
| 280 |
{
|
| 281 |
+
"epoch": 0.68,
|
| 282 |
+
"learning_rate": 4.3182357125918906e-05,
|
| 283 |
+
"loss": 2.0502,
|
| 284 |
"step": 23000
|
| 285 |
},
|
| 286 |
{
|
| 287 |
+
"epoch": 0.7,
|
| 288 |
+
"learning_rate": 4.303414749822149e-05,
|
| 289 |
+
"loss": 2.064,
|
| 290 |
"step": 23500
|
| 291 |
},
|
| 292 |
{
|
| 293 |
+
"epoch": 0.71,
|
| 294 |
+
"learning_rate": 4.2885937870524075e-05,
|
| 295 |
+
"loss": 2.0351,
|
| 296 |
"step": 24000
|
| 297 |
},
|
| 298 |
{
|
| 299 |
+
"epoch": 0.73,
|
| 300 |
+
"learning_rate": 4.273772824282666e-05,
|
| 301 |
+
"loss": 2.0362,
|
| 302 |
"step": 24500
|
| 303 |
},
|
| 304 |
{
|
| 305 |
+
"epoch": 0.74,
|
| 306 |
+
"learning_rate": 4.258951861512924e-05,
|
| 307 |
+
"loss": 2.0603,
|
| 308 |
"step": 25000
|
| 309 |
},
|
| 310 |
{
|
| 311 |
+
"epoch": 0.76,
|
| 312 |
+
"learning_rate": 4.244130898743183e-05,
|
| 313 |
+
"loss": 2.0452,
|
| 314 |
"step": 25500
|
| 315 |
},
|
| 316 |
{
|
| 317 |
+
"epoch": 0.77,
|
| 318 |
+
"learning_rate": 4.229309935973441e-05,
|
| 319 |
+
"loss": 2.0752,
|
| 320 |
"step": 26000
|
| 321 |
},
|
| 322 |
{
|
| 323 |
+
"epoch": 0.79,
|
| 324 |
+
"learning_rate": 4.2144889732036997e-05,
|
| 325 |
+
"loss": 2.062,
|
| 326 |
"step": 26500
|
| 327 |
},
|
| 328 |
{
|
| 329 |
+
"epoch": 0.8,
|
| 330 |
+
"learning_rate": 4.199668010433958e-05,
|
| 331 |
+
"loss": 2.0593,
|
| 332 |
"step": 27000
|
| 333 |
},
|
| 334 |
{
|
| 335 |
+
"epoch": 0.82,
|
| 336 |
+
"learning_rate": 4.1848470476642165e-05,
|
| 337 |
+
"loss": 2.0548,
|
| 338 |
"step": 27500
|
| 339 |
},
|
| 340 |
{
|
| 341 |
+
"epoch": 0.83,
|
| 342 |
+
"learning_rate": 4.170026084894475e-05,
|
| 343 |
+
"loss": 2.0516,
|
| 344 |
"step": 28000
|
| 345 |
},
|
| 346 |
{
|
| 347 |
+
"epoch": 0.84,
|
| 348 |
+
"learning_rate": 4.1552051221247334e-05,
|
| 349 |
+
"loss": 2.0609,
|
| 350 |
"step": 28500
|
| 351 |
},
|
| 352 |
{
|
| 353 |
+
"epoch": 0.86,
|
| 354 |
+
"learning_rate": 4.140384159354992e-05,
|
| 355 |
+
"loss": 2.0655,
|
| 356 |
"step": 29000
|
| 357 |
},
|
| 358 |
{
|
| 359 |
+
"epoch": 0.87,
|
| 360 |
+
"learning_rate": 4.12556319658525e-05,
|
| 361 |
+
"loss": 2.0562,
|
| 362 |
"step": 29500
|
| 363 |
},
|
| 364 |
{
|
| 365 |
+
"epoch": 0.89,
|
| 366 |
+
"learning_rate": 4.110742233815509e-05,
|
| 367 |
+
"loss": 2.0557,
|
| 368 |
"step": 30000
|
| 369 |
},
|
| 370 |
{
|
| 371 |
+
"epoch": 0.9,
|
| 372 |
+
"learning_rate": 4.095921271045767e-05,
|
| 373 |
+
"loss": 2.0608,
|
| 374 |
"step": 30500
|
| 375 |
},
|
| 376 |
{
|
| 377 |
+
"epoch": 0.92,
|
| 378 |
+
"learning_rate": 4.0811003082760256e-05,
|
| 379 |
+
"loss": 2.0641,
|
| 380 |
"step": 31000
|
| 381 |
},
|
| 382 |
{
|
| 383 |
+
"epoch": 0.93,
|
| 384 |
+
"learning_rate": 4.066279345506284e-05,
|
| 385 |
+
"loss": 2.0649,
|
| 386 |
"step": 31500
|
| 387 |
},
|
| 388 |
{
|
| 389 |
+
"epoch": 0.95,
|
| 390 |
+
"learning_rate": 4.0514583827365425e-05,
|
| 391 |
+
"loss": 2.0603,
|
| 392 |
"step": 32000
|
| 393 |
},
|
| 394 |
{
|
| 395 |
+
"epoch": 0.96,
|
| 396 |
+
"learning_rate": 4.036637419966801e-05,
|
| 397 |
+
"loss": 2.0696,
|
| 398 |
"step": 32500
|
| 399 |
},
|
| 400 |
{
|
| 401 |
+
"epoch": 0.98,
|
| 402 |
+
"learning_rate": 4.0218164571970594e-05,
|
| 403 |
+
"loss": 2.0619,
|
| 404 |
"step": 33000
|
| 405 |
},
|
| 406 |
{
|
| 407 |
+
"epoch": 0.99,
|
| 408 |
+
"learning_rate": 4.006995494427318e-05,
|
| 409 |
+
"loss": 2.0762,
|
| 410 |
"step": 33500
|
| 411 |
},
|
| 412 |
{
|
| 413 |
+
"epoch": 1.01,
|
| 414 |
+
"learning_rate": 3.992174531657576e-05,
|
| 415 |
+
"loss": 1.9612,
|
| 416 |
"step": 34000
|
| 417 |
},
|
| 418 |
{
|
| 419 |
+
"epoch": 1.02,
|
| 420 |
+
"learning_rate": 3.977353568887835e-05,
|
| 421 |
+
"loss": 1.8788,
|
| 422 |
"step": 34500
|
| 423 |
},
|
| 424 |
{
|
| 425 |
+
"epoch": 1.04,
|
| 426 |
+
"learning_rate": 3.962532606118093e-05,
|
| 427 |
+
"loss": 1.8705,
|
| 428 |
"step": 35000
|
| 429 |
},
|
| 430 |
{
|
| 431 |
+
"epoch": 1.05,
|
| 432 |
+
"learning_rate": 3.9477116433483516e-05,
|
| 433 |
+
"loss": 1.8726,
|
| 434 |
"step": 35500
|
| 435 |
},
|
| 436 |
{
|
| 437 |
+
"epoch": 1.07,
|
| 438 |
+
"learning_rate": 3.93289068057861e-05,
|
| 439 |
+
"loss": 1.887,
|
| 440 |
"step": 36000
|
| 441 |
},
|
| 442 |
{
|
| 443 |
+
"epoch": 1.08,
|
| 444 |
+
"learning_rate": 3.9180697178088685e-05,
|
| 445 |
+
"loss": 1.8795,
|
| 446 |
"step": 36500
|
| 447 |
},
|
| 448 |
{
|
| 449 |
+
"epoch": 1.1,
|
| 450 |
+
"learning_rate": 3.9032487550391276e-05,
|
| 451 |
+
"loss": 1.8778,
|
| 452 |
"step": 37000
|
| 453 |
},
|
| 454 |
{
|
| 455 |
+
"epoch": 1.11,
|
| 456 |
+
"learning_rate": 3.888427792269386e-05,
|
| 457 |
+
"loss": 1.8933,
|
| 458 |
"step": 37500
|
| 459 |
},
|
| 460 |
{
|
| 461 |
+
"epoch": 1.13,
|
| 462 |
+
"learning_rate": 3.8736068294996445e-05,
|
| 463 |
+
"loss": 1.8897,
|
| 464 |
"step": 38000
|
| 465 |
},
|
| 466 |
{
|
| 467 |
+
"epoch": 1.14,
|
| 468 |
+
"learning_rate": 3.858785866729903e-05,
|
| 469 |
+
"loss": 1.9019,
|
| 470 |
"step": 38500
|
| 471 |
},
|
| 472 |
{
|
| 473 |
+
"epoch": 1.16,
|
| 474 |
+
"learning_rate": 3.8439649039601614e-05,
|
| 475 |
+
"loss": 1.9074,
|
| 476 |
"step": 39000
|
| 477 |
},
|
| 478 |
{
|
| 479 |
+
"epoch": 1.17,
|
| 480 |
+
"learning_rate": 3.82914394119042e-05,
|
| 481 |
+
"loss": 1.916,
|
| 482 |
"step": 39500
|
| 483 |
},
|
| 484 |
{
|
| 485 |
+
"epoch": 1.19,
|
| 486 |
+
"learning_rate": 3.814322978420678e-05,
|
| 487 |
+
"loss": 1.9083,
|
| 488 |
"step": 40000
|
| 489 |
},
|
| 490 |
{
|
| 491 |
+
"epoch": 1.2,
|
| 492 |
+
"learning_rate": 3.799502015650937e-05,
|
| 493 |
+
"loss": 1.8971,
|
| 494 |
"step": 40500
|
| 495 |
},
|
| 496 |
{
|
| 497 |
+
"epoch": 1.22,
|
| 498 |
+
"learning_rate": 3.784681052881195e-05,
|
| 499 |
+
"loss": 1.9045,
|
| 500 |
"step": 41000
|
| 501 |
},
|
| 502 |
{
|
| 503 |
+
"epoch": 1.23,
|
| 504 |
+
"learning_rate": 3.7698600901114536e-05,
|
| 505 |
+
"loss": 1.9176,
|
| 506 |
"step": 41500
|
| 507 |
},
|
| 508 |
{
|
| 509 |
+
"epoch": 1.24,
|
| 510 |
+
"learning_rate": 3.755039127341713e-05,
|
| 511 |
+
"loss": 1.9223,
|
| 512 |
"step": 42000
|
| 513 |
},
|
| 514 |
{
|
| 515 |
+
"epoch": 1.26,
|
| 516 |
+
"learning_rate": 3.740218164571971e-05,
|
| 517 |
+
"loss": 1.9123,
|
| 518 |
"step": 42500
|
| 519 |
},
|
| 520 |
{
|
| 521 |
+
"epoch": 1.27,
|
| 522 |
+
"learning_rate": 3.7253972018022296e-05,
|
| 523 |
+
"loss": 1.9281,
|
| 524 |
"step": 43000
|
| 525 |
},
|
| 526 |
{
|
| 527 |
+
"epoch": 1.29,
|
| 528 |
+
"learning_rate": 3.710576239032488e-05,
|
| 529 |
+
"loss": 1.926,
|
| 530 |
"step": 43500
|
| 531 |
},
|
| 532 |
{
|
| 533 |
+
"epoch": 1.3,
|
| 534 |
+
"learning_rate": 3.6957552762627465e-05,
|
| 535 |
+
"loss": 1.9125,
|
| 536 |
"step": 44000
|
| 537 |
},
|
| 538 |
{
|
| 539 |
+
"epoch": 1.32,
|
| 540 |
+
"learning_rate": 3.680934313493005e-05,
|
| 541 |
+
"loss": 1.9354,
|
| 542 |
"step": 44500
|
| 543 |
},
|
| 544 |
{
|
| 545 |
+
"epoch": 1.33,
|
| 546 |
+
"learning_rate": 3.6661133507232633e-05,
|
| 547 |
+
"loss": 1.9339,
|
| 548 |
"step": 45000
|
| 549 |
},
|
| 550 |
{
|
| 551 |
+
"epoch": 1.35,
|
| 552 |
+
"learning_rate": 3.651292387953522e-05,
|
| 553 |
+
"loss": 1.9329,
|
| 554 |
"step": 45500
|
| 555 |
},
|
| 556 |
{
|
| 557 |
+
"epoch": 1.36,
|
| 558 |
+
"learning_rate": 3.63647142518378e-05,
|
| 559 |
+
"loss": 1.9223,
|
| 560 |
"step": 46000
|
| 561 |
},
|
| 562 |
{
|
| 563 |
+
"epoch": 1.38,
|
| 564 |
+
"learning_rate": 3.621650462414039e-05,
|
| 565 |
+
"loss": 1.9371,
|
| 566 |
"step": 46500
|
| 567 |
},
|
| 568 |
{
|
| 569 |
+
"epoch": 1.39,
|
| 570 |
+
"learning_rate": 3.606829499644297e-05,
|
| 571 |
+
"loss": 1.9348,
|
| 572 |
"step": 47000
|
| 573 |
},
|
| 574 |
{
|
| 575 |
+
"epoch": 1.41,
|
| 576 |
+
"learning_rate": 3.5920085368745556e-05,
|
| 577 |
+
"loss": 1.9424,
|
| 578 |
"step": 47500
|
| 579 |
},
|
| 580 |
{
|
| 581 |
+
"epoch": 1.42,
|
| 582 |
+
"learning_rate": 3.577187574104814e-05,
|
| 583 |
+
"loss": 1.9366,
|
| 584 |
"step": 48000
|
| 585 |
},
|
| 586 |
{
|
| 587 |
+
"epoch": 1.44,
|
| 588 |
+
"learning_rate": 3.5623666113350724e-05,
|
| 589 |
+
"loss": 1.9425,
|
| 590 |
"step": 48500
|
| 591 |
},
|
| 592 |
{
|
| 593 |
+
"epoch": 1.45,
|
| 594 |
+
"learning_rate": 3.547545648565331e-05,
|
| 595 |
+
"loss": 1.9292,
|
| 596 |
"step": 49000
|
| 597 |
},
|
| 598 |
{
|
| 599 |
+
"epoch": 1.47,
|
| 600 |
+
"learning_rate": 3.532724685795589e-05,
|
| 601 |
+
"loss": 1.9534,
|
| 602 |
"step": 49500
|
| 603 |
},
|
| 604 |
{
|
| 605 |
+
"epoch": 1.48,
|
| 606 |
+
"learning_rate": 3.517903723025848e-05,
|
| 607 |
+
"loss": 1.9499,
|
| 608 |
"step": 50000
|
| 609 |
},
|
| 610 |
{
|
| 611 |
+
"epoch": 1.5,
|
| 612 |
+
"learning_rate": 3.503082760256106e-05,
|
| 613 |
+
"loss": 1.9541,
|
| 614 |
"step": 50500
|
| 615 |
},
|
| 616 |
{
|
| 617 |
+
"epoch": 1.51,
|
| 618 |
+
"learning_rate": 3.4882617974863646e-05,
|
| 619 |
+
"loss": 1.9433,
|
| 620 |
"step": 51000
|
| 621 |
},
|
| 622 |
{
|
| 623 |
+
"epoch": 1.53,
|
| 624 |
+
"learning_rate": 3.473440834716624e-05,
|
| 625 |
+
"loss": 1.9481,
|
| 626 |
"step": 51500
|
| 627 |
},
|
| 628 |
{
|
| 629 |
+
"epoch": 1.54,
|
| 630 |
+
"learning_rate": 3.458619871946882e-05,
|
| 631 |
+
"loss": 1.9647,
|
| 632 |
"step": 52000
|
| 633 |
},
|
| 634 |
{
|
| 635 |
+
"epoch": 1.56,
|
| 636 |
+
"learning_rate": 3.4437989091771406e-05,
|
| 637 |
+
"loss": 1.9608,
|
| 638 |
"step": 52500
|
| 639 |
},
|
| 640 |
{
|
| 641 |
+
"epoch": 1.57,
|
| 642 |
+
"learning_rate": 3.428977946407399e-05,
|
| 643 |
+
"loss": 1.9775,
|
| 644 |
"step": 53000
|
| 645 |
},
|
| 646 |
{
|
| 647 |
+
"epoch": 1.59,
|
| 648 |
+
"learning_rate": 3.4141569836376575e-05,
|
| 649 |
+
"loss": 1.9676,
|
| 650 |
"step": 53500
|
| 651 |
},
|
| 652 |
{
|
| 653 |
+
"epoch": 1.6,
|
| 654 |
+
"learning_rate": 3.399336020867916e-05,
|
| 655 |
+
"loss": 1.978,
|
| 656 |
"step": 54000
|
| 657 |
},
|
| 658 |
{
|
| 659 |
+
"epoch": 1.62,
|
| 660 |
+
"learning_rate": 3.3845150580981744e-05,
|
| 661 |
+
"loss": 1.9705,
|
| 662 |
"step": 54500
|
| 663 |
},
|
| 664 |
{
|
| 665 |
+
"epoch": 1.63,
|
| 666 |
+
"learning_rate": 3.369694095328433e-05,
|
| 667 |
+
"loss": 1.9706,
|
| 668 |
"step": 55000
|
| 669 |
},
|
| 670 |
{
|
| 671 |
+
"epoch": 1.65,
|
| 672 |
+
"learning_rate": 3.354873132558691e-05,
|
| 673 |
+
"loss": 1.9605,
|
| 674 |
"step": 55500
|
| 675 |
},
|
| 676 |
{
|
| 677 |
+
"epoch": 1.66,
|
| 678 |
+
"learning_rate": 3.34005216978895e-05,
|
| 679 |
+
"loss": 1.9515,
|
| 680 |
"step": 56000
|
| 681 |
},
|
| 682 |
{
|
| 683 |
+
"epoch": 1.67,
|
| 684 |
+
"learning_rate": 3.325231207019208e-05,
|
| 685 |
+
"loss": 1.9836,
|
| 686 |
"step": 56500
|
| 687 |
},
|
| 688 |
{
|
| 689 |
+
"epoch": 1.69,
|
| 690 |
+
"learning_rate": 3.3104102442494666e-05,
|
| 691 |
+
"loss": 1.9752,
|
| 692 |
"step": 57000
|
| 693 |
},
|
| 694 |
{
|
| 695 |
+
"epoch": 1.7,
|
| 696 |
+
"learning_rate": 3.295589281479725e-05,
|
| 697 |
+
"loss": 1.9636,
|
| 698 |
"step": 57500
|
| 699 |
},
|
| 700 |
{
|
| 701 |
+
"epoch": 1.72,
|
| 702 |
+
"learning_rate": 3.2807683187099835e-05,
|
| 703 |
+
"loss": 1.9643,
|
| 704 |
"step": 58000
|
| 705 |
},
|
| 706 |
{
|
| 707 |
+
"epoch": 1.73,
|
| 708 |
+
"learning_rate": 3.265947355940242e-05,
|
| 709 |
+
"loss": 1.9778,
|
| 710 |
"step": 58500
|
| 711 |
},
|
| 712 |
{
|
| 713 |
+
"epoch": 1.75,
|
| 714 |
+
"learning_rate": 3.2511263931705004e-05,
|
| 715 |
+
"loss": 1.9894,
|
| 716 |
"step": 59000
|
| 717 |
},
|
| 718 |
{
|
| 719 |
+
"epoch": 1.76,
|
| 720 |
+
"learning_rate": 3.236305430400759e-05,
|
| 721 |
+
"loss": 1.9787,
|
| 722 |
"step": 59500
|
| 723 |
},
|
| 724 |
{
|
| 725 |
+
"epoch": 1.78,
|
| 726 |
+
"learning_rate": 3.221484467631017e-05,
|
| 727 |
+
"loss": 1.9945,
|
| 728 |
"step": 60000
|
| 729 |
},
|
| 730 |
{
|
| 731 |
+
"epoch": 1.79,
|
| 732 |
+
"learning_rate": 3.206663504861276e-05,
|
| 733 |
+
"loss": 1.9872,
|
| 734 |
"step": 60500
|
| 735 |
},
|
| 736 |
{
|
| 737 |
+
"epoch": 1.81,
|
| 738 |
+
"learning_rate": 3.191842542091534e-05,
|
| 739 |
+
"loss": 1.9684,
|
| 740 |
"step": 61000
|
| 741 |
},
|
| 742 |
{
|
| 743 |
+
"epoch": 1.82,
|
| 744 |
+
"learning_rate": 3.1770215793217926e-05,
|
| 745 |
+
"loss": 1.9923,
|
| 746 |
"step": 61500
|
| 747 |
},
|
| 748 |
{
|
| 749 |
+
"epoch": 1.84,
|
| 750 |
+
"learning_rate": 3.162200616552051e-05,
|
| 751 |
+
"loss": 1.9902,
|
| 752 |
"step": 62000
|
| 753 |
},
|
| 754 |
{
|
| 755 |
+
"epoch": 1.85,
|
| 756 |
+
"learning_rate": 3.1473796537823095e-05,
|
| 757 |
+
"loss": 1.991,
|
| 758 |
"step": 62500
|
| 759 |
},
|
| 760 |
{
|
| 761 |
+
"epoch": 1.87,
|
| 762 |
+
"learning_rate": 3.132558691012568e-05,
|
| 763 |
+
"loss": 1.9865,
|
| 764 |
"step": 63000
|
| 765 |
},
|
| 766 |
{
|
| 767 |
+
"epoch": 1.88,
|
| 768 |
+
"learning_rate": 3.1177377282428263e-05,
|
| 769 |
+
"loss": 2.0036,
|
| 770 |
"step": 63500
|
| 771 |
},
|
| 772 |
{
|
| 773 |
+
"epoch": 1.9,
|
| 774 |
+
"learning_rate": 3.102916765473085e-05,
|
| 775 |
+
"loss": 1.9841,
|
| 776 |
"step": 64000
|
| 777 |
},
|
| 778 |
{
|
| 779 |
+
"epoch": 1.91,
|
| 780 |
+
"learning_rate": 3.088095802703343e-05,
|
| 781 |
+
"loss": 1.9879,
|
| 782 |
"step": 64500
|
| 783 |
},
|
| 784 |
{
|
| 785 |
+
"epoch": 1.93,
|
| 786 |
+
"learning_rate": 3.073274839933602e-05,
|
| 787 |
+
"loss": 2.0156,
|
| 788 |
"step": 65000
|
| 789 |
},
|
| 790 |
{
|
| 791 |
+
"epoch": 1.94,
|
| 792 |
+
"learning_rate": 3.05845387716386e-05,
|
| 793 |
+
"loss": 1.9924,
|
| 794 |
"step": 65500
|
| 795 |
},
|
| 796 |
{
|
| 797 |
+
"epoch": 1.96,
|
| 798 |
+
"learning_rate": 3.0436329143941196e-05,
|
| 799 |
+
"loss": 2.0209,
|
| 800 |
"step": 66000
|
| 801 |
},
|
| 802 |
{
|
| 803 |
+
"epoch": 1.97,
|
| 804 |
+
"learning_rate": 3.028811951624378e-05,
|
| 805 |
+
"loss": 2.0109,
|
| 806 |
"step": 66500
|
| 807 |
},
|
| 808 |
{
|
| 809 |
+
"epoch": 1.99,
|
| 810 |
+
"learning_rate": 3.0139909888546365e-05,
|
| 811 |
+
"loss": 2.0246,
|
| 812 |
"step": 67000
|
| 813 |
},
|
| 814 |
{
|
| 815 |
+
"epoch": 2.0,
|
| 816 |
+
"learning_rate": 2.999170026084895e-05,
|
| 817 |
+
"loss": 2.0058,
|
| 818 |
"step": 67500
|
| 819 |
},
|
| 820 |
{
|
| 821 |
+
"epoch": 2.02,
|
| 822 |
+
"learning_rate": 2.9843490633151533e-05,
|
| 823 |
+
"loss": 1.8351,
|
| 824 |
"step": 68000
|
| 825 |
},
|
| 826 |
{
|
| 827 |
+
"epoch": 2.03,
|
| 828 |
+
"learning_rate": 2.9695281005454118e-05,
|
| 829 |
+
"loss": 1.8244,
|
| 830 |
"step": 68500
|
| 831 |
},
|
| 832 |
{
|
| 833 |
+
"epoch": 2.05,
|
| 834 |
+
"learning_rate": 2.9547071377756702e-05,
|
| 835 |
+
"loss": 1.8274,
|
| 836 |
"step": 69000
|
| 837 |
},
|
| 838 |
{
|
| 839 |
+
"epoch": 2.06,
|
| 840 |
+
"learning_rate": 2.9398861750059287e-05,
|
| 841 |
+
"loss": 1.8208,
|
| 842 |
"step": 69500
|
| 843 |
},
|
| 844 |
{
|
| 845 |
+
"epoch": 2.07,
|
| 846 |
+
"learning_rate": 2.925065212236187e-05,
|
| 847 |
+
"loss": 1.8353,
|
| 848 |
"step": 70000
|
| 849 |
},
|
| 850 |
{
|
| 851 |
+
"epoch": 2.09,
|
| 852 |
+
"learning_rate": 2.9102442494664455e-05,
|
| 853 |
+
"loss": 1.8316,
|
| 854 |
"step": 70500
|
| 855 |
},
|
| 856 |
{
|
| 857 |
+
"epoch": 2.1,
|
| 858 |
+
"learning_rate": 2.895423286696704e-05,
|
| 859 |
+
"loss": 1.8313,
|
| 860 |
"step": 71000
|
| 861 |
},
|
| 862 |
{
|
| 863 |
+
"epoch": 2.12,
|
| 864 |
+
"learning_rate": 2.8806023239269624e-05,
|
| 865 |
+
"loss": 1.8318,
|
| 866 |
"step": 71500
|
| 867 |
},
|
| 868 |
{
|
| 869 |
+
"epoch": 2.13,
|
| 870 |
+
"learning_rate": 2.865781361157221e-05,
|
| 871 |
+
"loss": 1.8448,
|
| 872 |
"step": 72000
|
| 873 |
},
|
| 874 |
{
|
| 875 |
+
"epoch": 2.15,
|
| 876 |
+
"learning_rate": 2.8509603983874793e-05,
|
| 877 |
+
"loss": 1.8367,
|
| 878 |
"step": 72500
|
| 879 |
},
|
| 880 |
{
|
| 881 |
+
"epoch": 2.16,
|
| 882 |
+
"learning_rate": 2.8361394356177377e-05,
|
| 883 |
+
"loss": 1.8535,
|
| 884 |
"step": 73000
|
| 885 |
},
|
| 886 |
{
|
| 887 |
+
"epoch": 2.18,
|
| 888 |
+
"learning_rate": 2.8213184728479962e-05,
|
| 889 |
+
"loss": 1.8474,
|
| 890 |
"step": 73500
|
| 891 |
},
|
| 892 |
{
|
| 893 |
+
"epoch": 2.19,
|
| 894 |
+
"learning_rate": 2.8064975100782546e-05,
|
| 895 |
+
"loss": 1.8525,
|
| 896 |
"step": 74000
|
| 897 |
},
|
| 898 |
{
|
| 899 |
+
"epoch": 2.21,
|
| 900 |
+
"learning_rate": 2.7916765473085134e-05,
|
| 901 |
+
"loss": 1.8612,
|
| 902 |
"step": 74500
|
| 903 |
},
|
| 904 |
{
|
| 905 |
+
"epoch": 2.22,
|
| 906 |
+
"learning_rate": 2.776855584538772e-05,
|
| 907 |
+
"loss": 1.8486,
|
| 908 |
"step": 75000
|
| 909 |
},
|
| 910 |
{
|
| 911 |
+
"epoch": 2.24,
|
| 912 |
+
"learning_rate": 2.7620346217690303e-05,
|
| 913 |
+
"loss": 1.8339,
|
| 914 |
"step": 75500
|
| 915 |
},
|
| 916 |
{
|
| 917 |
+
"epoch": 2.25,
|
| 918 |
+
"learning_rate": 2.7472136589992887e-05,
|
| 919 |
+
"loss": 1.85,
|
| 920 |
"step": 76000
|
| 921 |
},
|
| 922 |
{
|
| 923 |
+
"epoch": 2.27,
|
| 924 |
+
"learning_rate": 2.7323926962295472e-05,
|
| 925 |
+
"loss": 1.8645,
|
| 926 |
"step": 76500
|
| 927 |
},
|
| 928 |
{
|
| 929 |
+
"epoch": 2.28,
|
| 930 |
+
"learning_rate": 2.7175717334598056e-05,
|
| 931 |
+
"loss": 1.849,
|
| 932 |
"step": 77000
|
| 933 |
},
|
| 934 |
{
|
| 935 |
+
"epoch": 2.3,
|
| 936 |
+
"learning_rate": 2.702750770690064e-05,
|
| 937 |
+
"loss": 1.8608,
|
| 938 |
"step": 77500
|
| 939 |
},
|
| 940 |
{
|
| 941 |
+
"epoch": 2.31,
|
| 942 |
+
"learning_rate": 2.6879298079203225e-05,
|
| 943 |
+
"loss": 1.8568,
|
| 944 |
"step": 78000
|
| 945 |
},
|
| 946 |
{
|
| 947 |
+
"epoch": 2.33,
|
| 948 |
+
"learning_rate": 2.673108845150581e-05,
|
| 949 |
+
"loss": 1.8607,
|
| 950 |
"step": 78500
|
| 951 |
},
|
| 952 |
{
|
| 953 |
+
"epoch": 2.34,
|
| 954 |
+
"learning_rate": 2.6582878823808394e-05,
|
| 955 |
+
"loss": 1.8607,
|
| 956 |
"step": 79000
|
| 957 |
},
|
| 958 |
{
|
| 959 |
+
"epoch": 2.36,
|
| 960 |
+
"learning_rate": 2.6434669196110978e-05,
|
| 961 |
+
"loss": 1.8478,
|
| 962 |
"step": 79500
|
| 963 |
},
|
| 964 |
{
|
| 965 |
+
"epoch": 2.37,
|
| 966 |
+
"learning_rate": 2.6286459568413563e-05,
|
| 967 |
+
"loss": 1.8785,
|
| 968 |
"step": 80000
|
| 969 |
},
|
| 970 |
{
|
| 971 |
+
"epoch": 2.39,
|
| 972 |
+
"learning_rate": 2.6138249940716147e-05,
|
| 973 |
+
"loss": 1.8729,
|
| 974 |
"step": 80500
|
| 975 |
},
|
| 976 |
{
|
| 977 |
+
"epoch": 2.4,
|
| 978 |
+
"learning_rate": 2.5990040313018738e-05,
|
| 979 |
+
"loss": 1.8511,
|
| 980 |
"step": 81000
|
| 981 |
},
|
| 982 |
{
|
| 983 |
+
"epoch": 2.42,
|
| 984 |
+
"learning_rate": 2.5841830685321323e-05,
|
| 985 |
+
"loss": 1.8701,
|
| 986 |
"step": 81500
|
| 987 |
},
|
| 988 |
{
|
| 989 |
+
"epoch": 2.43,
|
| 990 |
+
"learning_rate": 2.5693621057623907e-05,
|
| 991 |
+
"loss": 1.8792,
|
| 992 |
"step": 82000
|
| 993 |
},
|
| 994 |
{
|
| 995 |
+
"epoch": 2.45,
|
| 996 |
+
"learning_rate": 2.554541142992649e-05,
|
| 997 |
+
"loss": 1.8703,
|
| 998 |
"step": 82500
|
| 999 |
},
|
| 1000 |
{
|
| 1001 |
+
"epoch": 2.46,
|
| 1002 |
+
"learning_rate": 2.5397201802229076e-05,
|
| 1003 |
+
"loss": 1.8479,
|
| 1004 |
"step": 83000
|
| 1005 |
},
|
| 1006 |
{
|
| 1007 |
+
"epoch": 2.48,
|
| 1008 |
+
"learning_rate": 2.524899217453166e-05,
|
| 1009 |
+
"loss": 1.8603,
|
| 1010 |
"step": 83500
|
| 1011 |
},
|
| 1012 |
{
|
| 1013 |
+
"epoch": 2.49,
|
| 1014 |
+
"learning_rate": 2.5100782546834245e-05,
|
| 1015 |
+
"loss": 1.8752,
|
| 1016 |
"step": 84000
|
| 1017 |
},
|
| 1018 |
{
|
| 1019 |
+
"epoch": 2.5,
|
| 1020 |
+
"learning_rate": 2.495257291913683e-05,
|
| 1021 |
+
"loss": 1.8685,
|
| 1022 |
"step": 84500
|
| 1023 |
},
|
| 1024 |
{
|
| 1025 |
+
"epoch": 2.52,
|
| 1026 |
+
"learning_rate": 2.4804363291439414e-05,
|
| 1027 |
+
"loss": 1.8814,
|
| 1028 |
"step": 85000
|
| 1029 |
},
|
| 1030 |
{
|
| 1031 |
+
"epoch": 2.53,
|
| 1032 |
+
"learning_rate": 2.4656153663741998e-05,
|
| 1033 |
+
"loss": 1.8903,
|
| 1034 |
"step": 85500
|
| 1035 |
},
|
| 1036 |
{
|
| 1037 |
+
"epoch": 2.55,
|
| 1038 |
+
"learning_rate": 2.4507944036044582e-05,
|
| 1039 |
+
"loss": 1.8863,
|
| 1040 |
"step": 86000
|
| 1041 |
},
|
| 1042 |
{
|
| 1043 |
+
"epoch": 2.56,
|
| 1044 |
+
"learning_rate": 2.4359734408347167e-05,
|
| 1045 |
+
"loss": 1.8704,
|
| 1046 |
"step": 86500
|
| 1047 |
},
|
| 1048 |
{
|
| 1049 |
+
"epoch": 2.58,
|
| 1050 |
+
"learning_rate": 2.421152478064975e-05,
|
| 1051 |
+
"loss": 1.8822,
|
| 1052 |
"step": 87000
|
| 1053 |
},
|
| 1054 |
{
|
| 1055 |
+
"epoch": 2.59,
|
| 1056 |
+
"learning_rate": 2.4063315152952336e-05,
|
| 1057 |
+
"loss": 1.8917,
|
| 1058 |
"step": 87500
|
| 1059 |
},
|
| 1060 |
{
|
| 1061 |
+
"epoch": 2.61,
|
| 1062 |
+
"learning_rate": 2.391510552525492e-05,
|
| 1063 |
+
"loss": 1.8734,
|
| 1064 |
"step": 88000
|
| 1065 |
},
|
| 1066 |
{
|
| 1067 |
+
"epoch": 2.62,
|
| 1068 |
+
"learning_rate": 2.3766895897557508e-05,
|
| 1069 |
+
"loss": 1.8797,
|
| 1070 |
"step": 88500
|
| 1071 |
},
|
| 1072 |
{
|
| 1073 |
+
"epoch": 2.64,
|
| 1074 |
+
"learning_rate": 2.3618686269860092e-05,
|
| 1075 |
+
"loss": 1.8659,
|
| 1076 |
"step": 89000
|
| 1077 |
},
|
| 1078 |
{
|
| 1079 |
+
"epoch": 2.65,
|
| 1080 |
+
"learning_rate": 2.3470476642162677e-05,
|
| 1081 |
+
"loss": 1.8844,
|
| 1082 |
"step": 89500
|
| 1083 |
},
|
| 1084 |
{
|
| 1085 |
+
"epoch": 2.67,
|
| 1086 |
+
"learning_rate": 2.332226701446526e-05,
|
| 1087 |
+
"loss": 1.8956,
|
| 1088 |
"step": 90000
|
| 1089 |
},
|
| 1090 |
{
|
| 1091 |
+
"epoch": 2.68,
|
| 1092 |
+
"learning_rate": 2.3174057386767846e-05,
|
| 1093 |
+
"loss": 1.89,
|
| 1094 |
"step": 90500
|
| 1095 |
},
|
| 1096 |
{
|
| 1097 |
+
"epoch": 2.7,
|
| 1098 |
+
"learning_rate": 2.302584775907043e-05,
|
| 1099 |
+
"loss": 1.8722,
|
| 1100 |
"step": 91000
|
| 1101 |
},
|
| 1102 |
{
|
| 1103 |
+
"epoch": 2.71,
|
| 1104 |
+
"learning_rate": 2.2877638131373014e-05,
|
| 1105 |
+
"loss": 1.8978,
|
| 1106 |
"step": 91500
|
| 1107 |
},
|
| 1108 |
{
|
| 1109 |
+
"epoch": 2.73,
|
| 1110 |
+
"learning_rate": 2.2729428503675602e-05,
|
| 1111 |
+
"loss": 1.8858,
|
| 1112 |
"step": 92000
|
| 1113 |
},
|
| 1114 |
{
|
| 1115 |
+
"epoch": 2.74,
|
| 1116 |
+
"learning_rate": 2.2581218875978187e-05,
|
| 1117 |
+
"loss": 1.8845,
|
| 1118 |
"step": 92500
|
| 1119 |
},
|
| 1120 |
{
|
| 1121 |
+
"epoch": 2.76,
|
| 1122 |
+
"learning_rate": 2.243300924828077e-05,
|
| 1123 |
+
"loss": 1.879,
|
| 1124 |
"step": 93000
|
| 1125 |
},
|
| 1126 |
{
|
| 1127 |
+
"epoch": 2.77,
|
| 1128 |
+
"learning_rate": 2.2284799620583355e-05,
|
| 1129 |
+
"loss": 1.8963,
|
| 1130 |
"step": 93500
|
| 1131 |
},
|
| 1132 |
{
|
| 1133 |
+
"epoch": 2.79,
|
| 1134 |
+
"learning_rate": 2.213658999288594e-05,
|
| 1135 |
+
"loss": 1.9058,
|
| 1136 |
"step": 94000
|
| 1137 |
},
|
| 1138 |
{
|
| 1139 |
+
"epoch": 2.8,
|
| 1140 |
+
"learning_rate": 2.1988380365188524e-05,
|
| 1141 |
+
"loss": 1.8961,
|
| 1142 |
"step": 94500
|
| 1143 |
},
|
| 1144 |
{
|
| 1145 |
+
"epoch": 2.82,
|
| 1146 |
+
"learning_rate": 2.184017073749111e-05,
|
| 1147 |
+
"loss": 1.8718,
|
| 1148 |
"step": 95000
|
| 1149 |
},
|
| 1150 |
{
|
| 1151 |
+
"epoch": 2.83,
|
| 1152 |
+
"learning_rate": 2.1691961109793693e-05,
|
| 1153 |
+
"loss": 1.8841,
|
| 1154 |
"step": 95500
|
| 1155 |
},
|
| 1156 |
{
|
| 1157 |
+
"epoch": 2.85,
|
| 1158 |
+
"learning_rate": 2.1543751482096277e-05,
|
| 1159 |
+
"loss": 1.8928,
|
| 1160 |
"step": 96000
|
| 1161 |
},
|
| 1162 |
{
|
| 1163 |
+
"epoch": 2.86,
|
| 1164 |
+
"learning_rate": 2.1395541854398862e-05,
|
| 1165 |
+
"loss": 1.8924,
|
| 1166 |
"step": 96500
|
| 1167 |
},
|
| 1168 |
{
|
| 1169 |
+
"epoch": 2.88,
|
| 1170 |
+
"learning_rate": 2.1247332226701446e-05,
|
| 1171 |
+
"loss": 1.912,
|
| 1172 |
"step": 97000
|
| 1173 |
},
|
| 1174 |
{
|
| 1175 |
+
"epoch": 2.89,
|
| 1176 |
+
"learning_rate": 2.109912259900403e-05,
|
| 1177 |
+
"loss": 1.886,
|
| 1178 |
"step": 97500
|
| 1179 |
},
|
| 1180 |
{
|
| 1181 |
+
"epoch": 2.9,
|
| 1182 |
+
"learning_rate": 2.0950912971306615e-05,
|
| 1183 |
+
"loss": 1.9036,
|
| 1184 |
"step": 98000
|
| 1185 |
},
|
| 1186 |
{
|
| 1187 |
+
"epoch": 2.92,
|
| 1188 |
+
"learning_rate": 2.08027033436092e-05,
|
| 1189 |
+
"loss": 1.9081,
|
| 1190 |
"step": 98500
|
| 1191 |
},
|
| 1192 |
{
|
| 1193 |
+
"epoch": 2.93,
|
| 1194 |
+
"learning_rate": 2.0654493715911787e-05,
|
| 1195 |
+
"loss": 1.8979,
|
| 1196 |
"step": 99000
|
| 1197 |
},
|
| 1198 |
{
|
| 1199 |
+
"epoch": 2.95,
|
| 1200 |
+
"learning_rate": 2.0506284088214372e-05,
|
| 1201 |
+
"loss": 1.9121,
|
| 1202 |
"step": 99500
|
| 1203 |
},
|
| 1204 |
{
|
| 1205 |
+
"epoch": 2.96,
|
| 1206 |
+
"learning_rate": 2.0358074460516956e-05,
|
| 1207 |
+
"loss": 1.9092,
|
| 1208 |
"step": 100000
|
| 1209 |
},
|
| 1210 |
{
|
| 1211 |
+
"epoch": 2.98,
|
| 1212 |
+
"learning_rate": 2.020986483281954e-05,
|
| 1213 |
+
"loss": 1.8997,
|
| 1214 |
"step": 100500
|
| 1215 |
},
|
| 1216 |
{
|
| 1217 |
+
"epoch": 2.99,
|
| 1218 |
+
"learning_rate": 2.0061655205122125e-05,
|
| 1219 |
+
"loss": 1.9143,
|
| 1220 |
"step": 101000
|
| 1221 |
},
|
| 1222 |
{
|
| 1223 |
+
"epoch": 3.01,
|
| 1224 |
+
"learning_rate": 1.991344557742471e-05,
|
| 1225 |
+
"loss": 1.8386,
|
| 1226 |
"step": 101500
|
| 1227 |
},
|
| 1228 |
{
|
| 1229 |
+
"epoch": 3.02,
|
| 1230 |
+
"learning_rate": 1.9765235949727294e-05,
|
| 1231 |
+
"loss": 1.7796,
|
| 1232 |
"step": 102000
|
| 1233 |
},
|
| 1234 |
{
|
| 1235 |
+
"epoch": 3.04,
|
| 1236 |
+
"learning_rate": 1.961702632202988e-05,
|
| 1237 |
+
"loss": 1.7844,
|
| 1238 |
"step": 102500
|
| 1239 |
},
|
| 1240 |
{
|
| 1241 |
+
"epoch": 3.05,
|
| 1242 |
+
"learning_rate": 1.9468816694332466e-05,
|
| 1243 |
+
"loss": 1.7941,
|
| 1244 |
"step": 103000
|
| 1245 |
},
|
| 1246 |
{
|
| 1247 |
+
"epoch": 3.07,
|
| 1248 |
+
"learning_rate": 1.932060706663505e-05,
|
| 1249 |
+
"loss": 1.8003,
|
| 1250 |
"step": 103500
|
| 1251 |
},
|
| 1252 |
{
|
| 1253 |
+
"epoch": 3.08,
|
| 1254 |
+
"learning_rate": 1.9172397438937635e-05,
|
| 1255 |
+
"loss": 1.7926,
|
| 1256 |
"step": 104000
|
| 1257 |
},
|
| 1258 |
{
|
| 1259 |
+
"epoch": 3.1,
|
| 1260 |
+
"learning_rate": 1.902418781124022e-05,
|
| 1261 |
+
"loss": 1.7893,
|
| 1262 |
"step": 104500
|
| 1263 |
},
|
| 1264 |
{
|
| 1265 |
+
"epoch": 3.11,
|
| 1266 |
+
"learning_rate": 1.8875978183542804e-05,
|
| 1267 |
+
"loss": 1.8033,
|
| 1268 |
"step": 105000
|
| 1269 |
},
|
| 1270 |
{
|
| 1271 |
+
"epoch": 3.13,
|
| 1272 |
+
"learning_rate": 1.8727768555845388e-05,
|
| 1273 |
+
"loss": 1.8031,
|
| 1274 |
"step": 105500
|
| 1275 |
},
|
| 1276 |
{
|
| 1277 |
+
"epoch": 3.14,
|
| 1278 |
+
"learning_rate": 1.8579558928147972e-05,
|
| 1279 |
+
"loss": 1.7973,
|
| 1280 |
"step": 106000
|
| 1281 |
},
|
| 1282 |
{
|
| 1283 |
+
"epoch": 3.16,
|
| 1284 |
+
"learning_rate": 1.843134930045056e-05,
|
| 1285 |
+
"loss": 1.7955,
|
| 1286 |
"step": 106500
|
| 1287 |
},
|
| 1288 |
{
|
| 1289 |
+
"epoch": 3.17,
|
| 1290 |
+
"learning_rate": 1.8283139672753145e-05,
|
| 1291 |
+
"loss": 1.7894,
|
| 1292 |
"step": 107000
|
| 1293 |
},
|
| 1294 |
{
|
| 1295 |
+
"epoch": 3.19,
|
| 1296 |
+
"learning_rate": 1.813493004505573e-05,
|
| 1297 |
+
"loss": 1.8005,
|
| 1298 |
"step": 107500
|
| 1299 |
},
|
| 1300 |
{
|
| 1301 |
+
"epoch": 3.2,
|
| 1302 |
+
"learning_rate": 1.7986720417358314e-05,
|
| 1303 |
+
"loss": 1.8061,
|
| 1304 |
"step": 108000
|
| 1305 |
},
|
| 1306 |
{
|
| 1307 |
+
"epoch": 3.22,
|
| 1308 |
+
"learning_rate": 1.7838510789660898e-05,
|
| 1309 |
+
"loss": 1.8076,
|
| 1310 |
"step": 108500
|
| 1311 |
},
|
| 1312 |
{
|
| 1313 |
+
"epoch": 3.23,
|
| 1314 |
+
"learning_rate": 1.7690301161963482e-05,
|
| 1315 |
+
"loss": 1.7835,
|
| 1316 |
"step": 109000
|
| 1317 |
},
|
| 1318 |
{
|
| 1319 |
+
"epoch": 3.25,
|
| 1320 |
+
"learning_rate": 1.7542091534266067e-05,
|
| 1321 |
+
"loss": 1.805,
|
| 1322 |
"step": 109500
|
| 1323 |
},
|
| 1324 |
{
|
| 1325 |
+
"epoch": 3.26,
|
| 1326 |
+
"learning_rate": 1.739388190656865e-05,
|
| 1327 |
+
"loss": 1.8052,
|
| 1328 |
"step": 110000
|
| 1329 |
},
|
| 1330 |
{
|
| 1331 |
+
"epoch": 3.28,
|
| 1332 |
+
"learning_rate": 1.7245672278871236e-05,
|
| 1333 |
+
"loss": 1.8022,
|
| 1334 |
"step": 110500
|
| 1335 |
},
|
| 1336 |
{
|
| 1337 |
+
"epoch": 3.29,
|
| 1338 |
+
"learning_rate": 1.709746265117382e-05,
|
| 1339 |
+
"loss": 1.7789,
|
| 1340 |
"step": 111000
|
| 1341 |
},
|
| 1342 |
{
|
| 1343 |
+
"epoch": 3.31,
|
| 1344 |
+
"learning_rate": 1.6949253023476404e-05,
|
| 1345 |
+
"loss": 1.8049,
|
| 1346 |
"step": 111500
|
| 1347 |
},
|
| 1348 |
{
|
| 1349 |
+
"epoch": 3.32,
|
| 1350 |
+
"learning_rate": 1.680104339577899e-05,
|
| 1351 |
+
"loss": 1.8087,
|
| 1352 |
"step": 112000
|
| 1353 |
},
|
| 1354 |
{
|
| 1355 |
+
"epoch": 3.33,
|
| 1356 |
+
"learning_rate": 1.6652833768081573e-05,
|
| 1357 |
+
"loss": 1.8048,
|
| 1358 |
"step": 112500
|
| 1359 |
},
|
| 1360 |
{
|
| 1361 |
+
"epoch": 3.35,
|
| 1362 |
+
"learning_rate": 1.6504624140384158e-05,
|
| 1363 |
+
"loss": 1.7941,
|
| 1364 |
"step": 113000
|
| 1365 |
},
|
| 1366 |
{
|
| 1367 |
+
"epoch": 3.36,
|
| 1368 |
+
"learning_rate": 1.6356414512686745e-05,
|
| 1369 |
+
"loss": 1.8048,
|
| 1370 |
"step": 113500
|
| 1371 |
},
|
| 1372 |
{
|
| 1373 |
+
"epoch": 3.38,
|
| 1374 |
+
"learning_rate": 1.620820488498933e-05,
|
| 1375 |
+
"loss": 1.7987,
|
| 1376 |
"step": 114000
|
| 1377 |
},
|
| 1378 |
{
|
| 1379 |
+
"epoch": 3.39,
|
| 1380 |
+
"learning_rate": 1.6059995257291914e-05,
|
| 1381 |
+
"loss": 1.7982,
|
| 1382 |
"step": 114500
|
| 1383 |
},
|
| 1384 |
{
|
| 1385 |
+
"epoch": 3.41,
|
| 1386 |
+
"learning_rate": 1.59117856295945e-05,
|
| 1387 |
+
"loss": 1.8063,
|
| 1388 |
"step": 115000
|
| 1389 |
},
|
| 1390 |
{
|
| 1391 |
+
"epoch": 3.42,
|
| 1392 |
+
"learning_rate": 1.5763576001897083e-05,
|
| 1393 |
+
"loss": 1.8015,
|
| 1394 |
"step": 115500
|
| 1395 |
},
|
| 1396 |
{
|
| 1397 |
+
"epoch": 3.44,
|
| 1398 |
+
"learning_rate": 1.5615366374199667e-05,
|
| 1399 |
+
"loss": 1.8037,
|
| 1400 |
"step": 116000
|
| 1401 |
},
|
| 1402 |
{
|
| 1403 |
+
"epoch": 3.45,
|
| 1404 |
+
"learning_rate": 1.5467156746502255e-05,
|
| 1405 |
+
"loss": 1.8172,
|
| 1406 |
"step": 116500
|
| 1407 |
},
|
| 1408 |
{
|
| 1409 |
+
"epoch": 3.47,
|
| 1410 |
+
"learning_rate": 1.531894711880484e-05,
|
| 1411 |
+
"loss": 1.806,
|
| 1412 |
"step": 117000
|
| 1413 |
},
|
| 1414 |
{
|
| 1415 |
+
"epoch": 3.48,
|
| 1416 |
+
"learning_rate": 1.5170737491107422e-05,
|
| 1417 |
+
"loss": 1.8122,
|
| 1418 |
"step": 117500
|
| 1419 |
},
|
| 1420 |
{
|
| 1421 |
+
"epoch": 3.5,
|
| 1422 |
+
"learning_rate": 1.5022527863410007e-05,
|
| 1423 |
+
"loss": 1.7992,
|
| 1424 |
"step": 118000
|
| 1425 |
},
|
| 1426 |
{
|
| 1427 |
+
"epoch": 3.51,
|
| 1428 |
+
"learning_rate": 1.4874318235712591e-05,
|
| 1429 |
+
"loss": 1.7984,
|
| 1430 |
"step": 118500
|
| 1431 |
},
|
| 1432 |
{
|
| 1433 |
+
"epoch": 3.53,
|
| 1434 |
+
"learning_rate": 1.4726108608015177e-05,
|
| 1435 |
+
"loss": 1.7979,
|
| 1436 |
"step": 119000
|
| 1437 |
},
|
| 1438 |
{
|
| 1439 |
+
"epoch": 3.54,
|
| 1440 |
+
"learning_rate": 1.4577898980317762e-05,
|
| 1441 |
+
"loss": 1.8054,
|
| 1442 |
"step": 119500
|
| 1443 |
},
|
| 1444 |
{
|
| 1445 |
+
"epoch": 3.56,
|
| 1446 |
+
"learning_rate": 1.4429689352620346e-05,
|
| 1447 |
+
"loss": 1.7976,
|
| 1448 |
"step": 120000
|
| 1449 |
},
|
| 1450 |
{
|
| 1451 |
+
"epoch": 3.57,
|
| 1452 |
+
"learning_rate": 1.428147972492293e-05,
|
| 1453 |
+
"loss": 1.8096,
|
| 1454 |
"step": 120500
|
| 1455 |
},
|
| 1456 |
{
|
| 1457 |
+
"epoch": 3.59,
|
| 1458 |
+
"learning_rate": 1.4133270097225518e-05,
|
| 1459 |
+
"loss": 1.8067,
|
| 1460 |
"step": 121000
|
| 1461 |
},
|
| 1462 |
{
|
| 1463 |
+
"epoch": 3.6,
|
| 1464 |
+
"learning_rate": 1.3985060469528103e-05,
|
| 1465 |
+
"loss": 1.8065,
|
| 1466 |
"step": 121500
|
| 1467 |
},
|
| 1468 |
{
|
| 1469 |
+
"epoch": 3.62,
|
| 1470 |
+
"learning_rate": 1.3836850841830687e-05,
|
| 1471 |
+
"loss": 1.8052,
|
| 1472 |
"step": 122000
|
| 1473 |
},
|
| 1474 |
{
|
| 1475 |
+
"epoch": 3.63,
|
| 1476 |
+
"learning_rate": 1.3688641214133272e-05,
|
| 1477 |
+
"loss": 1.8016,
|
| 1478 |
"step": 122500
|
| 1479 |
},
|
| 1480 |
{
|
| 1481 |
+
"epoch": 3.65,
|
| 1482 |
+
"learning_rate": 1.3540431586435856e-05,
|
| 1483 |
+
"loss": 1.8101,
|
| 1484 |
"step": 123000
|
| 1485 |
},
|
| 1486 |
{
|
| 1487 |
+
"epoch": 3.66,
|
| 1488 |
+
"learning_rate": 1.339222195873844e-05,
|
| 1489 |
+
"loss": 1.8089,
|
| 1490 |
"step": 123500
|
| 1491 |
},
|
| 1492 |
{
|
| 1493 |
+
"epoch": 3.68,
|
| 1494 |
+
"learning_rate": 1.3244012331041025e-05,
|
| 1495 |
+
"loss": 1.8038,
|
| 1496 |
"step": 124000
|
| 1497 |
},
|
| 1498 |
{
|
| 1499 |
+
"epoch": 3.69,
|
| 1500 |
+
"learning_rate": 1.309580270334361e-05,
|
| 1501 |
+
"loss": 1.8165,
|
| 1502 |
"step": 124500
|
| 1503 |
},
|
| 1504 |
{
|
| 1505 |
+
"epoch": 3.71,
|
| 1506 |
+
"learning_rate": 1.2947593075646194e-05,
|
| 1507 |
+
"loss": 1.807,
|
| 1508 |
"step": 125000
|
| 1509 |
},
|
| 1510 |
{
|
| 1511 |
+
"epoch": 3.72,
|
| 1512 |
+
"learning_rate": 1.279938344794878e-05,
|
| 1513 |
+
"loss": 1.8113,
|
| 1514 |
"step": 125500
|
| 1515 |
},
|
| 1516 |
{
|
| 1517 |
+
"epoch": 3.73,
|
| 1518 |
+
"learning_rate": 1.2651173820251364e-05,
|
| 1519 |
+
"loss": 1.8073,
|
| 1520 |
"step": 126000
|
| 1521 |
},
|
| 1522 |
{
|
| 1523 |
+
"epoch": 3.75,
|
| 1524 |
+
"learning_rate": 1.2502964192553949e-05,
|
| 1525 |
+
"loss": 1.8129,
|
| 1526 |
"step": 126500
|
| 1527 |
},
|
| 1528 |
{
|
| 1529 |
+
"epoch": 3.76,
|
| 1530 |
+
"learning_rate": 1.2354754564856535e-05,
|
| 1531 |
+
"loss": 1.8475,
|
| 1532 |
"step": 127000
|
| 1533 |
},
|
| 1534 |
{
|
| 1535 |
+
"epoch": 3.78,
|
| 1536 |
+
"learning_rate": 1.220654493715912e-05,
|
| 1537 |
+
"loss": 1.8003,
|
| 1538 |
"step": 127500
|
| 1539 |
},
|
| 1540 |
{
|
| 1541 |
+
"epoch": 3.79,
|
| 1542 |
+
"learning_rate": 1.2058335309461704e-05,
|
| 1543 |
+
"loss": 1.816,
|
| 1544 |
"step": 128000
|
| 1545 |
},
|
| 1546 |
{
|
| 1547 |
+
"epoch": 3.81,
|
| 1548 |
+
"learning_rate": 1.1910125681764288e-05,
|
| 1549 |
+
"loss": 1.8198,
|
| 1550 |
"step": 128500
|
| 1551 |
},
|
| 1552 |
{
|
| 1553 |
+
"epoch": 3.82,
|
| 1554 |
+
"learning_rate": 1.1761916054066872e-05,
|
| 1555 |
+
"loss": 1.8339,
|
| 1556 |
"step": 129000
|
| 1557 |
},
|
| 1558 |
{
|
| 1559 |
+
"epoch": 3.84,
|
| 1560 |
+
"learning_rate": 1.1613706426369457e-05,
|
| 1561 |
+
"loss": 1.8053,
|
| 1562 |
"step": 129500
|
| 1563 |
},
|
| 1564 |
{
|
| 1565 |
+
"epoch": 3.85,
|
| 1566 |
+
"learning_rate": 1.1465496798672043e-05,
|
| 1567 |
+
"loss": 1.7948,
|
| 1568 |
"step": 130000
|
| 1569 |
},
|
| 1570 |
{
|
| 1571 |
+
"epoch": 3.87,
|
| 1572 |
+
"learning_rate": 1.1317287170974627e-05,
|
| 1573 |
+
"loss": 1.8261,
|
| 1574 |
"step": 130500
|
| 1575 |
},
|
| 1576 |
{
|
| 1577 |
+
"epoch": 3.88,
|
| 1578 |
+
"learning_rate": 1.1169077543277212e-05,
|
| 1579 |
+
"loss": 1.8097,
|
| 1580 |
"step": 131000
|
| 1581 |
},
|
| 1582 |
{
|
| 1583 |
+
"epoch": 3.9,
|
| 1584 |
+
"learning_rate": 1.1020867915579796e-05,
|
| 1585 |
+
"loss": 1.8101,
|
| 1586 |
"step": 131500
|
| 1587 |
},
|
| 1588 |
{
|
| 1589 |
+
"epoch": 3.91,
|
| 1590 |
+
"learning_rate": 1.087265828788238e-05,
|
| 1591 |
+
"loss": 1.8135,
|
| 1592 |
"step": 132000
|
| 1593 |
},
|
| 1594 |
{
|
| 1595 |
+
"epoch": 3.93,
|
| 1596 |
+
"learning_rate": 1.0724448660184967e-05,
|
| 1597 |
+
"loss": 1.8245,
|
| 1598 |
"step": 132500
|
| 1599 |
},
|
| 1600 |
{
|
| 1601 |
+
"epoch": 3.94,
|
| 1602 |
+
"learning_rate": 1.0576239032487551e-05,
|
| 1603 |
+
"loss": 1.8059,
|
| 1604 |
"step": 133000
|
| 1605 |
},
|
| 1606 |
{
|
| 1607 |
+
"epoch": 3.96,
|
| 1608 |
+
"learning_rate": 1.0428029404790136e-05,
|
| 1609 |
+
"loss": 1.8147,
|
| 1610 |
"step": 133500
|
| 1611 |
},
|
| 1612 |
{
|
| 1613 |
+
"epoch": 3.97,
|
| 1614 |
+
"learning_rate": 1.0279819777092722e-05,
|
| 1615 |
+
"loss": 1.8108,
|
| 1616 |
"step": 134000
|
| 1617 |
},
|
| 1618 |
{
|
| 1619 |
+
"epoch": 3.99,
|
| 1620 |
+
"learning_rate": 1.0131610149395306e-05,
|
| 1621 |
+
"loss": 1.8207,
|
| 1622 |
"step": 134500
|
| 1623 |
},
|
| 1624 |
{
|
| 1625 |
+
"epoch": 4.0,
|
| 1626 |
+
"learning_rate": 9.98340052169789e-06,
|
| 1627 |
+
"loss": 1.8035,
|
| 1628 |
+
"step": 135000
|
| 1629 |
+
},
|
| 1630 |
+
{
|
| 1631 |
+
"epoch": 4.02,
|
| 1632 |
+
"learning_rate": 9.835190894000475e-06,
|
| 1633 |
+
"loss": 1.7493,
|
| 1634 |
+
"step": 135500
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 4.03,
|
| 1638 |
+
"learning_rate": 9.68698126630306e-06,
|
| 1639 |
+
"loss": 1.7577,
|
| 1640 |
+
"step": 136000
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 4.05,
|
| 1644 |
+
"learning_rate": 9.538771638605644e-06,
|
| 1645 |
+
"loss": 1.7447,
|
| 1646 |
+
"step": 136500
|
| 1647 |
+
},
|
| 1648 |
+
{
|
| 1649 |
+
"epoch": 4.06,
|
| 1650 |
+
"learning_rate": 9.390562010908228e-06,
|
| 1651 |
+
"loss": 1.7555,
|
| 1652 |
+
"step": 137000
|
| 1653 |
+
},
|
| 1654 |
+
{
|
| 1655 |
+
"epoch": 4.08,
|
| 1656 |
+
"learning_rate": 9.242352383210814e-06,
|
| 1657 |
+
"loss": 1.7644,
|
| 1658 |
+
"step": 137500
|
| 1659 |
+
},
|
| 1660 |
+
{
|
| 1661 |
+
"epoch": 4.09,
|
| 1662 |
+
"learning_rate": 9.094142755513399e-06,
|
| 1663 |
+
"loss": 1.7464,
|
| 1664 |
+
"step": 138000
|
| 1665 |
+
},
|
| 1666 |
+
{
|
| 1667 |
+
"epoch": 4.11,
|
| 1668 |
+
"learning_rate": 8.945933127815983e-06,
|
| 1669 |
+
"loss": 1.751,
|
| 1670 |
+
"step": 138500
|
| 1671 |
+
},
|
| 1672 |
+
{
|
| 1673 |
+
"epoch": 4.12,
|
| 1674 |
+
"learning_rate": 8.797723500118567e-06,
|
| 1675 |
+
"loss": 1.7587,
|
| 1676 |
+
"step": 139000
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"epoch": 4.14,
|
| 1680 |
+
"learning_rate": 8.649513872421154e-06,
|
| 1681 |
+
"loss": 1.7638,
|
| 1682 |
+
"step": 139500
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 4.15,
|
| 1686 |
+
"learning_rate": 8.501304244723738e-06,
|
| 1687 |
+
"loss": 1.763,
|
| 1688 |
+
"step": 140000
|
| 1689 |
+
},
|
| 1690 |
+
{
|
| 1691 |
+
"epoch": 4.16,
|
| 1692 |
+
"learning_rate": 8.353094617026322e-06,
|
| 1693 |
+
"loss": 1.7529,
|
| 1694 |
+
"step": 140500
|
| 1695 |
+
},
|
| 1696 |
+
{
|
| 1697 |
+
"epoch": 4.18,
|
| 1698 |
+
"learning_rate": 8.204884989328908e-06,
|
| 1699 |
+
"loss": 1.745,
|
| 1700 |
+
"step": 141000
|
| 1701 |
+
},
|
| 1702 |
+
{
|
| 1703 |
+
"epoch": 4.19,
|
| 1704 |
+
"learning_rate": 8.056675361631493e-06,
|
| 1705 |
+
"loss": 1.7497,
|
| 1706 |
+
"step": 141500
|
| 1707 |
+
},
|
| 1708 |
+
{
|
| 1709 |
+
"epoch": 4.21,
|
| 1710 |
+
"learning_rate": 7.908465733934077e-06,
|
| 1711 |
+
"loss": 1.7582,
|
| 1712 |
+
"step": 142000
|
| 1713 |
+
},
|
| 1714 |
+
{
|
| 1715 |
+
"epoch": 4.22,
|
| 1716 |
+
"learning_rate": 7.760256106236662e-06,
|
| 1717 |
+
"loss": 1.7425,
|
| 1718 |
+
"step": 142500
|
| 1719 |
+
},
|
| 1720 |
+
{
|
| 1721 |
+
"epoch": 4.24,
|
| 1722 |
+
"learning_rate": 7.612046478539246e-06,
|
| 1723 |
+
"loss": 1.7576,
|
| 1724 |
+
"step": 143000
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 4.25,
|
| 1728 |
+
"learning_rate": 7.4638368508418305e-06,
|
| 1729 |
+
"loss": 1.7482,
|
| 1730 |
+
"step": 143500
|
| 1731 |
+
},
|
| 1732 |
+
{
|
| 1733 |
+
"epoch": 4.27,
|
| 1734 |
+
"learning_rate": 7.315627223144415e-06,
|
| 1735 |
+
"loss": 1.7449,
|
| 1736 |
+
"step": 144000
|
| 1737 |
+
},
|
| 1738 |
+
{
|
| 1739 |
+
"epoch": 4.28,
|
| 1740 |
+
"learning_rate": 7.167417595447e-06,
|
| 1741 |
+
"loss": 1.7451,
|
| 1742 |
+
"step": 144500
|
| 1743 |
+
},
|
| 1744 |
+
{
|
| 1745 |
+
"epoch": 4.3,
|
| 1746 |
+
"learning_rate": 7.0192079677495855e-06,
|
| 1747 |
+
"loss": 1.765,
|
| 1748 |
+
"step": 145000
|
| 1749 |
+
},
|
| 1750 |
+
{
|
| 1751 |
+
"epoch": 4.31,
|
| 1752 |
+
"learning_rate": 6.870998340052171e-06,
|
| 1753 |
+
"loss": 1.7539,
|
| 1754 |
+
"step": 145500
|
| 1755 |
+
},
|
| 1756 |
+
{
|
| 1757 |
+
"epoch": 4.33,
|
| 1758 |
+
"learning_rate": 6.722788712354755e-06,
|
| 1759 |
+
"loss": 1.7476,
|
| 1760 |
+
"step": 146000
|
| 1761 |
+
},
|
| 1762 |
+
{
|
| 1763 |
+
"epoch": 4.34,
|
| 1764 |
+
"learning_rate": 6.5745790846573396e-06,
|
| 1765 |
+
"loss": 1.7635,
|
| 1766 |
+
"step": 146500
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 4.36,
|
| 1770 |
+
"learning_rate": 6.426369456959924e-06,
|
| 1771 |
+
"loss": 1.7672,
|
| 1772 |
+
"step": 147000
|
| 1773 |
+
},
|
| 1774 |
+
{
|
| 1775 |
+
"epoch": 4.37,
|
| 1776 |
+
"learning_rate": 6.278159829262508e-06,
|
| 1777 |
+
"loss": 1.7673,
|
| 1778 |
+
"step": 147500
|
| 1779 |
+
},
|
| 1780 |
+
{
|
| 1781 |
+
"epoch": 4.39,
|
| 1782 |
+
"learning_rate": 6.129950201565094e-06,
|
| 1783 |
+
"loss": 1.7783,
|
| 1784 |
+
"step": 148000
|
| 1785 |
+
},
|
| 1786 |
+
{
|
| 1787 |
+
"epoch": 4.4,
|
| 1788 |
+
"learning_rate": 5.981740573867679e-06,
|
| 1789 |
+
"loss": 1.7514,
|
| 1790 |
+
"step": 148500
|
| 1791 |
+
},
|
| 1792 |
+
{
|
| 1793 |
+
"epoch": 4.42,
|
| 1794 |
+
"learning_rate": 5.833530946170263e-06,
|
| 1795 |
+
"loss": 1.7546,
|
| 1796 |
+
"step": 149000
|
| 1797 |
+
},
|
| 1798 |
+
{
|
| 1799 |
+
"epoch": 4.43,
|
| 1800 |
+
"learning_rate": 5.6853213184728486e-06,
|
| 1801 |
+
"loss": 1.7502,
|
| 1802 |
+
"step": 149500
|
| 1803 |
+
},
|
| 1804 |
+
{
|
| 1805 |
+
"epoch": 4.45,
|
| 1806 |
+
"learning_rate": 5.537111690775433e-06,
|
| 1807 |
+
"loss": 1.7542,
|
| 1808 |
+
"step": 150000
|
| 1809 |
+
},
|
| 1810 |
+
{
|
| 1811 |
+
"epoch": 4.46,
|
| 1812 |
+
"learning_rate": 5.388902063078017e-06,
|
| 1813 |
+
"loss": 1.772,
|
| 1814 |
+
"step": 150500
|
| 1815 |
+
},
|
| 1816 |
+
{
|
| 1817 |
+
"epoch": 4.48,
|
| 1818 |
+
"learning_rate": 5.240692435380603e-06,
|
| 1819 |
+
"loss": 1.7575,
|
| 1820 |
+
"step": 151000
|
| 1821 |
+
},
|
| 1822 |
+
{
|
| 1823 |
+
"epoch": 4.49,
|
| 1824 |
+
"learning_rate": 5.092482807683188e-06,
|
| 1825 |
+
"loss": 1.755,
|
| 1826 |
+
"step": 151500
|
| 1827 |
+
},
|
| 1828 |
+
{
|
| 1829 |
+
"epoch": 4.51,
|
| 1830 |
+
"learning_rate": 4.944273179985772e-06,
|
| 1831 |
+
"loss": 1.7393,
|
| 1832 |
+
"step": 152000
|
| 1833 |
+
},
|
| 1834 |
+
{
|
| 1835 |
+
"epoch": 4.52,
|
| 1836 |
+
"learning_rate": 4.796063552288357e-06,
|
| 1837 |
+
"loss": 1.7534,
|
| 1838 |
+
"step": 152500
|
| 1839 |
+
},
|
| 1840 |
+
{
|
| 1841 |
+
"epoch": 4.54,
|
| 1842 |
+
"learning_rate": 4.647853924590942e-06,
|
| 1843 |
+
"loss": 1.75,
|
| 1844 |
+
"step": 153000
|
| 1845 |
+
},
|
| 1846 |
+
{
|
| 1847 |
+
"epoch": 4.55,
|
| 1848 |
+
"learning_rate": 4.499644296893526e-06,
|
| 1849 |
+
"loss": 1.7493,
|
| 1850 |
+
"step": 153500
|
| 1851 |
+
},
|
| 1852 |
+
{
|
| 1853 |
+
"epoch": 4.56,
|
| 1854 |
+
"learning_rate": 4.351434669196111e-06,
|
| 1855 |
+
"loss": 1.7508,
|
| 1856 |
+
"step": 154000
|
| 1857 |
+
},
|
| 1858 |
+
{
|
| 1859 |
+
"epoch": 4.58,
|
| 1860 |
+
"learning_rate": 4.203225041498696e-06,
|
| 1861 |
+
"loss": 1.7482,
|
| 1862 |
+
"step": 154500
|
| 1863 |
+
},
|
| 1864 |
+
{
|
| 1865 |
+
"epoch": 4.59,
|
| 1866 |
+
"learning_rate": 4.055015413801281e-06,
|
| 1867 |
+
"loss": 1.7632,
|
| 1868 |
+
"step": 155000
|
| 1869 |
+
},
|
| 1870 |
+
{
|
| 1871 |
+
"epoch": 4.61,
|
| 1872 |
+
"learning_rate": 3.906805786103866e-06,
|
| 1873 |
+
"loss": 1.7621,
|
| 1874 |
+
"step": 155500
|
| 1875 |
+
},
|
| 1876 |
+
{
|
| 1877 |
+
"epoch": 4.62,
|
| 1878 |
+
"learning_rate": 3.75859615840645e-06,
|
| 1879 |
+
"loss": 1.7632,
|
| 1880 |
+
"step": 156000
|
| 1881 |
+
},
|
| 1882 |
+
{
|
| 1883 |
+
"epoch": 4.64,
|
| 1884 |
+
"learning_rate": 3.6103865307090354e-06,
|
| 1885 |
+
"loss": 1.7635,
|
| 1886 |
+
"step": 156500
|
| 1887 |
+
},
|
| 1888 |
+
{
|
| 1889 |
+
"epoch": 4.65,
|
| 1890 |
+
"learning_rate": 3.46217690301162e-06,
|
| 1891 |
+
"loss": 1.7587,
|
| 1892 |
+
"step": 157000
|
| 1893 |
+
},
|
| 1894 |
+
{
|
| 1895 |
+
"epoch": 4.67,
|
| 1896 |
+
"learning_rate": 3.3139672753142047e-06,
|
| 1897 |
+
"loss": 1.7453,
|
| 1898 |
+
"step": 157500
|
| 1899 |
+
},
|
| 1900 |
+
{
|
| 1901 |
+
"epoch": 4.68,
|
| 1902 |
+
"learning_rate": 3.165757647616789e-06,
|
| 1903 |
+
"loss": 1.7471,
|
| 1904 |
+
"step": 158000
|
| 1905 |
+
},
|
| 1906 |
+
{
|
| 1907 |
+
"epoch": 4.7,
|
| 1908 |
+
"learning_rate": 3.017548019919374e-06,
|
| 1909 |
+
"loss": 1.7476,
|
| 1910 |
+
"step": 158500
|
| 1911 |
+
},
|
| 1912 |
+
{
|
| 1913 |
+
"epoch": 4.71,
|
| 1914 |
+
"learning_rate": 2.8693383922219588e-06,
|
| 1915 |
+
"loss": 1.7567,
|
| 1916 |
+
"step": 159000
|
| 1917 |
+
},
|
| 1918 |
+
{
|
| 1919 |
+
"epoch": 4.73,
|
| 1920 |
+
"learning_rate": 2.7211287645245436e-06,
|
| 1921 |
+
"loss": 1.7796,
|
| 1922 |
+
"step": 159500
|
| 1923 |
+
},
|
| 1924 |
+
{
|
| 1925 |
+
"epoch": 4.74,
|
| 1926 |
+
"learning_rate": 2.5729191368271284e-06,
|
| 1927 |
+
"loss": 1.7471,
|
| 1928 |
+
"step": 160000
|
| 1929 |
+
},
|
| 1930 |
+
{
|
| 1931 |
+
"epoch": 4.76,
|
| 1932 |
+
"learning_rate": 2.424709509129713e-06,
|
| 1933 |
+
"loss": 1.7699,
|
| 1934 |
+
"step": 160500
|
| 1935 |
+
},
|
| 1936 |
+
{
|
| 1937 |
+
"epoch": 4.77,
|
| 1938 |
+
"learning_rate": 2.276499881432298e-06,
|
| 1939 |
+
"loss": 1.739,
|
| 1940 |
+
"step": 161000
|
| 1941 |
+
},
|
| 1942 |
+
{
|
| 1943 |
+
"epoch": 4.79,
|
| 1944 |
+
"learning_rate": 2.1282902537348825e-06,
|
| 1945 |
+
"loss": 1.7413,
|
| 1946 |
+
"step": 161500
|
| 1947 |
+
},
|
| 1948 |
+
{
|
| 1949 |
+
"epoch": 4.8,
|
| 1950 |
+
"learning_rate": 1.9800806260374674e-06,
|
| 1951 |
+
"loss": 1.7553,
|
| 1952 |
+
"step": 162000
|
| 1953 |
+
},
|
| 1954 |
+
{
|
| 1955 |
+
"epoch": 4.82,
|
| 1956 |
+
"learning_rate": 1.8318709983400524e-06,
|
| 1957 |
+
"loss": 1.7391,
|
| 1958 |
+
"step": 162500
|
| 1959 |
+
},
|
| 1960 |
+
{
|
| 1961 |
+
"epoch": 4.83,
|
| 1962 |
+
"learning_rate": 1.683661370642637e-06,
|
| 1963 |
+
"loss": 1.7446,
|
| 1964 |
+
"step": 163000
|
| 1965 |
+
},
|
| 1966 |
+
{
|
| 1967 |
+
"epoch": 4.85,
|
| 1968 |
+
"learning_rate": 1.5354517429452217e-06,
|
| 1969 |
+
"loss": 1.7626,
|
| 1970 |
+
"step": 163500
|
| 1971 |
+
},
|
| 1972 |
+
{
|
| 1973 |
+
"epoch": 4.86,
|
| 1974 |
+
"learning_rate": 1.3872421152478065e-06,
|
| 1975 |
+
"loss": 1.7454,
|
| 1976 |
+
"step": 164000
|
| 1977 |
+
},
|
| 1978 |
+
{
|
| 1979 |
+
"epoch": 4.88,
|
| 1980 |
+
"learning_rate": 1.2390324875503913e-06,
|
| 1981 |
+
"loss": 1.7657,
|
| 1982 |
+
"step": 164500
|
| 1983 |
+
},
|
| 1984 |
+
{
|
| 1985 |
+
"epoch": 4.89,
|
| 1986 |
+
"learning_rate": 1.0908228598529762e-06,
|
| 1987 |
+
"loss": 1.758,
|
| 1988 |
+
"step": 165000
|
| 1989 |
+
},
|
| 1990 |
+
{
|
| 1991 |
+
"epoch": 4.91,
|
| 1992 |
+
"learning_rate": 9.426132321555609e-07,
|
| 1993 |
+
"loss": 1.7517,
|
| 1994 |
+
"step": 165500
|
| 1995 |
+
},
|
| 1996 |
+
{
|
| 1997 |
+
"epoch": 4.92,
|
| 1998 |
+
"learning_rate": 7.944036044581456e-07,
|
| 1999 |
+
"loss": 1.7506,
|
| 2000 |
+
"step": 166000
|
| 2001 |
+
},
|
| 2002 |
+
{
|
| 2003 |
+
"epoch": 4.94,
|
| 2004 |
+
"learning_rate": 6.461939767607305e-07,
|
| 2005 |
+
"loss": 1.7592,
|
| 2006 |
+
"step": 166500
|
| 2007 |
+
},
|
| 2008 |
+
{
|
| 2009 |
+
"epoch": 4.95,
|
| 2010 |
+
"learning_rate": 4.979843490633152e-07,
|
| 2011 |
+
"loss": 1.7486,
|
| 2012 |
+
"step": 167000
|
| 2013 |
+
},
|
| 2014 |
+
{
|
| 2015 |
+
"epoch": 4.97,
|
| 2016 |
+
"learning_rate": 3.4977472136589993e-07,
|
| 2017 |
+
"loss": 1.7369,
|
| 2018 |
+
"step": 167500
|
| 2019 |
+
},
|
| 2020 |
+
{
|
| 2021 |
+
"epoch": 4.98,
|
| 2022 |
+
"learning_rate": 2.0156509366848474e-07,
|
| 2023 |
+
"loss": 1.7659,
|
| 2024 |
+
"step": 168000
|
| 2025 |
+
},
|
| 2026 |
+
{
|
| 2027 |
+
"epoch": 4.99,
|
| 2028 |
+
"learning_rate": 5.3355465971069484e-08,
|
| 2029 |
+
"loss": 1.747,
|
| 2030 |
+
"step": 168500
|
| 2031 |
+
},
|
| 2032 |
+
{
|
| 2033 |
+
"epoch": 5.0,
|
| 2034 |
+
"step": 168680,
|
| 2035 |
+
"total_flos": 6.266036456049869e+17,
|
| 2036 |
+
"train_loss": 1.883479554409491,
|
| 2037 |
+
"train_runtime": 65223.4395,
|
| 2038 |
+
"train_samples_per_second": 10.345,
|
| 2039 |
+
"train_steps_per_second": 2.586
|
| 2040 |
}
|
| 2041 |
],
|
| 2042 |
+
"max_steps": 168680,
|
| 2043 |
+
"num_train_epochs": 5,
|
| 2044 |
+
"total_flos": 6.266036456049869e+17,
|
| 2045 |
"trial_name": null,
|
| 2046 |
"trial_params": null
|
| 2047 |
}
|