End of training
Browse files- all_results.json +10 -10
- eval_results.json +6 -7
- runs/Aug07_12-22-23_dsai-server/events.out.tfevents.1691388807.dsai-server +3 -0
- train_results.json +5 -6
- trainer_state.json +640 -640
all_results.json
CHANGED
|
@@ -1,14 +1,14 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"eval_loss":
|
| 4 |
-
"eval_runtime":
|
| 5 |
"eval_samples": 5043,
|
| 6 |
-
"eval_samples_per_second":
|
| 7 |
-
"eval_steps_per_second":
|
| 8 |
-
"eval_wer":
|
| 9 |
-
"train_loss": 1.
|
| 10 |
-
"train_runtime":
|
| 11 |
"train_samples": 5043,
|
| 12 |
-
"train_samples_per_second":
|
| 13 |
-
"train_steps_per_second":
|
| 14 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 0.86,
|
| 3 |
+
"eval_loss": 1.1651121377944946,
|
| 4 |
+
"eval_runtime": 161.2265,
|
| 5 |
"eval_samples": 5043,
|
| 6 |
+
"eval_samples_per_second": 15.202,
|
| 7 |
+
"eval_steps_per_second": 3.802,
|
| 8 |
+
"eval_wer": 76.37636988522145,
|
| 9 |
+
"train_loss": 1.343574110031128,
|
| 10 |
+
"train_runtime": 2849.8517,
|
| 11 |
"train_samples": 5043,
|
| 12 |
+
"train_samples_per_second": 14.036,
|
| 13 |
+
"train_steps_per_second": 1.754
|
| 14 |
}
|
eval_results.json
CHANGED
|
@@ -1,9 +1,8 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"eval_loss":
|
| 4 |
-
"eval_runtime":
|
| 5 |
-
"
|
| 6 |
-
"
|
| 7 |
-
"
|
| 8 |
-
"eval_wer": 5.397983265393693
|
| 9 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 0.86,
|
| 3 |
+
"eval_loss": 1.1651121377944946,
|
| 4 |
+
"eval_runtime": 161.2265,
|
| 5 |
+
"eval_samples_per_second": 15.202,
|
| 6 |
+
"eval_steps_per_second": 3.802,
|
| 7 |
+
"eval_wer": 76.37636988522145
|
|
|
|
| 8 |
}
|
runs/Aug07_12-22-23_dsai-server/events.out.tfevents.1691388807.dsai-server
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d903968d5903c2120d02a88630de9967c7ee038a0d951a0c37105fa0c55a97ee
|
| 3 |
+
size 40
|
train_results.json
CHANGED
|
@@ -1,8 +1,7 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"train_loss": 1.
|
| 4 |
-
"train_runtime":
|
| 5 |
-
"
|
| 6 |
-
"
|
| 7 |
-
"train_steps_per_second": 0.501
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 0.86,
|
| 3 |
+
"train_loss": 1.343574110031128,
|
| 4 |
+
"train_runtime": 2849.8517,
|
| 5 |
+
"train_samples_per_second": 14.036,
|
| 6 |
+
"train_steps_per_second": 1.754
|
|
|
|
| 7 |
}
|
trainer_state.json
CHANGED
|
@@ -1,1270 +1,1270 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric":
|
| 3 |
"best_model_checkpoint": "./output/small/yt-special-batch8-tiny/checkpoint-5000",
|
| 4 |
-
"epoch":
|
| 5 |
"global_step": 5000,
|
| 6 |
"is_hyper_param_search": false,
|
| 7 |
"is_local_process_zero": true,
|
| 8 |
"is_world_process_zero": true,
|
| 9 |
"log_history": [
|
| 10 |
{
|
| 11 |
-
"epoch": 0.
|
| 12 |
-
"learning_rate": 4.
|
| 13 |
-
"loss":
|
| 14 |
"step": 25
|
| 15 |
},
|
| 16 |
{
|
| 17 |
-
"epoch": 0.
|
| 18 |
-
"learning_rate": 9.
|
| 19 |
-
"loss":
|
| 20 |
"step": 50
|
| 21 |
},
|
| 22 |
{
|
| 23 |
-
"epoch": 0.
|
| 24 |
-
"learning_rate": 1.
|
| 25 |
-
"loss":
|
| 26 |
"step": 75
|
| 27 |
},
|
| 28 |
{
|
| 29 |
-
"epoch": 0.
|
| 30 |
-
"learning_rate": 1.
|
| 31 |
-
"loss":
|
| 32 |
"step": 100
|
| 33 |
},
|
| 34 |
{
|
| 35 |
-
"epoch": 0.
|
| 36 |
-
"learning_rate": 2.
|
| 37 |
-
"loss":
|
| 38 |
"step": 125
|
| 39 |
},
|
| 40 |
{
|
| 41 |
-
"epoch": 0.
|
| 42 |
-
"learning_rate": 2.
|
| 43 |
-
"loss": 2.
|
| 44 |
"step": 150
|
| 45 |
},
|
| 46 |
{
|
| 47 |
-
"epoch": 0.
|
| 48 |
-
"learning_rate": 3.
|
| 49 |
-
"loss":
|
| 50 |
"step": 175
|
| 51 |
},
|
| 52 |
{
|
| 53 |
-
"epoch": 0.
|
| 54 |
-
"learning_rate": 3.
|
| 55 |
-
"loss":
|
| 56 |
"step": 200
|
| 57 |
},
|
| 58 |
{
|
| 59 |
-
"epoch": 0.
|
| 60 |
-
"learning_rate": 4.
|
| 61 |
-
"loss":
|
| 62 |
"step": 225
|
| 63 |
},
|
| 64 |
{
|
| 65 |
-
"epoch": 0.
|
| 66 |
-
"learning_rate": 4.
|
| 67 |
-
"loss":
|
| 68 |
"step": 250
|
| 69 |
},
|
| 70 |
{
|
| 71 |
-
"epoch": 0.
|
| 72 |
-
"learning_rate": 5.
|
| 73 |
-
"loss":
|
| 74 |
"step": 275
|
| 75 |
},
|
| 76 |
{
|
| 77 |
-
"epoch": 0.
|
| 78 |
-
"learning_rate": 5.
|
| 79 |
-
"loss":
|
| 80 |
"step": 300
|
| 81 |
},
|
| 82 |
{
|
| 83 |
-
"epoch": 0.
|
| 84 |
-
"learning_rate": 6.
|
| 85 |
-
"loss":
|
| 86 |
"step": 325
|
| 87 |
},
|
| 88 |
{
|
| 89 |
-
"epoch": 0.
|
| 90 |
-
"learning_rate": 6.
|
| 91 |
-
"loss":
|
| 92 |
"step": 350
|
| 93 |
},
|
| 94 |
{
|
| 95 |
-
"epoch": 0.
|
| 96 |
-
"learning_rate": 7.
|
| 97 |
-
"loss":
|
| 98 |
"step": 375
|
| 99 |
},
|
| 100 |
{
|
| 101 |
-
"epoch": 0.
|
| 102 |
-
"learning_rate": 7.
|
| 103 |
-
"loss":
|
| 104 |
"step": 400
|
| 105 |
},
|
| 106 |
{
|
| 107 |
-
"epoch": 0.
|
| 108 |
-
"learning_rate": 8.
|
| 109 |
-
"loss":
|
| 110 |
"step": 425
|
| 111 |
},
|
| 112 |
{
|
| 113 |
-
"epoch": 0.
|
| 114 |
-
"learning_rate": 8.
|
| 115 |
-
"loss":
|
| 116 |
"step": 450
|
| 117 |
},
|
| 118 |
{
|
| 119 |
-
"epoch": 0.
|
| 120 |
-
"learning_rate": 9.
|
| 121 |
-
"loss":
|
| 122 |
"step": 475
|
| 123 |
},
|
| 124 |
{
|
| 125 |
-
"epoch": 0.
|
| 126 |
-
"learning_rate": 9.
|
| 127 |
-
"loss":
|
| 128 |
"step": 500
|
| 129 |
},
|
| 130 |
{
|
| 131 |
-
"epoch": 0.
|
| 132 |
-
"learning_rate": 9.
|
| 133 |
-
"loss":
|
| 134 |
"step": 525
|
| 135 |
},
|
| 136 |
{
|
| 137 |
-
"epoch": 0.
|
| 138 |
-
"learning_rate": 9.
|
| 139 |
-
"loss":
|
| 140 |
"step": 550
|
| 141 |
},
|
| 142 |
{
|
| 143 |
-
"epoch": 0.
|
| 144 |
-
"learning_rate": 9.
|
| 145 |
-
"loss":
|
| 146 |
"step": 575
|
| 147 |
},
|
| 148 |
{
|
| 149 |
-
"epoch": 0.
|
| 150 |
-
"learning_rate": 9.
|
| 151 |
-
"loss":
|
| 152 |
"step": 600
|
| 153 |
},
|
| 154 |
{
|
| 155 |
-
"epoch": 0.
|
| 156 |
-
"learning_rate": 9.
|
| 157 |
-
"loss":
|
| 158 |
"step": 625
|
| 159 |
},
|
| 160 |
{
|
| 161 |
-
"epoch":
|
| 162 |
-
"learning_rate": 9.
|
| 163 |
-
"loss":
|
| 164 |
"step": 650
|
| 165 |
},
|
| 166 |
{
|
| 167 |
-
"epoch":
|
| 168 |
-
"learning_rate": 9.
|
| 169 |
-
"loss":
|
| 170 |
"step": 675
|
| 171 |
},
|
| 172 |
{
|
| 173 |
-
"epoch":
|
| 174 |
-
"learning_rate": 9.
|
| 175 |
-
"loss":
|
| 176 |
"step": 700
|
| 177 |
},
|
| 178 |
{
|
| 179 |
-
"epoch":
|
| 180 |
-
"learning_rate": 9.
|
| 181 |
-
"loss":
|
| 182 |
"step": 725
|
| 183 |
},
|
| 184 |
{
|
| 185 |
-
"epoch":
|
| 186 |
-
"learning_rate": 9.
|
| 187 |
-
"loss":
|
| 188 |
"step": 750
|
| 189 |
},
|
| 190 |
{
|
| 191 |
-
"epoch":
|
| 192 |
-
"learning_rate": 9.
|
| 193 |
-
"loss":
|
| 194 |
"step": 775
|
| 195 |
},
|
| 196 |
{
|
| 197 |
-
"epoch":
|
| 198 |
-
"learning_rate": 9.
|
| 199 |
-
"loss":
|
| 200 |
"step": 800
|
| 201 |
},
|
| 202 |
{
|
| 203 |
-
"epoch":
|
| 204 |
-
"learning_rate": 9.
|
| 205 |
-
"loss":
|
| 206 |
"step": 825
|
| 207 |
},
|
| 208 |
{
|
| 209 |
-
"epoch":
|
| 210 |
-
"learning_rate": 9.
|
| 211 |
-
"loss":
|
| 212 |
"step": 850
|
| 213 |
},
|
| 214 |
{
|
| 215 |
-
"epoch":
|
| 216 |
-
"learning_rate": 9.
|
| 217 |
-
"loss":
|
| 218 |
"step": 875
|
| 219 |
},
|
| 220 |
{
|
| 221 |
-
"epoch":
|
| 222 |
-
"learning_rate": 9.
|
| 223 |
-
"loss":
|
| 224 |
"step": 900
|
| 225 |
},
|
| 226 |
{
|
| 227 |
-
"epoch":
|
| 228 |
-
"learning_rate": 9.
|
| 229 |
-
"loss":
|
| 230 |
"step": 925
|
| 231 |
},
|
| 232 |
{
|
| 233 |
-
"epoch":
|
| 234 |
-
"learning_rate": 9.
|
| 235 |
-
"loss":
|
| 236 |
"step": 950
|
| 237 |
},
|
| 238 |
{
|
| 239 |
-
"epoch":
|
| 240 |
-
"learning_rate": 8.
|
| 241 |
-
"loss":
|
| 242 |
"step": 975
|
| 243 |
},
|
| 244 |
{
|
| 245 |
-
"epoch":
|
| 246 |
-
"learning_rate": 8.
|
| 247 |
-
"loss":
|
| 248 |
"step": 1000
|
| 249 |
},
|
| 250 |
{
|
| 251 |
-
"epoch":
|
| 252 |
-
"eval_loss":
|
| 253 |
-
"eval_runtime":
|
| 254 |
-
"eval_samples_per_second":
|
| 255 |
-
"eval_steps_per_second":
|
| 256 |
-
"eval_wer":
|
| 257 |
"step": 1000
|
| 258 |
},
|
| 259 |
{
|
| 260 |
-
"epoch":
|
| 261 |
-
"learning_rate": 8.
|
| 262 |
-
"loss":
|
| 263 |
"step": 1025
|
| 264 |
},
|
| 265 |
{
|
| 266 |
-
"epoch":
|
| 267 |
-
"learning_rate": 8.
|
| 268 |
-
"loss":
|
| 269 |
"step": 1050
|
| 270 |
},
|
| 271 |
{
|
| 272 |
-
"epoch":
|
| 273 |
-
"learning_rate": 8.
|
| 274 |
-
"loss":
|
| 275 |
"step": 1075
|
| 276 |
},
|
| 277 |
{
|
| 278 |
-
"epoch":
|
| 279 |
-
"learning_rate": 8.
|
| 280 |
-
"loss":
|
| 281 |
"step": 1100
|
| 282 |
},
|
| 283 |
{
|
| 284 |
-
"epoch":
|
| 285 |
-
"learning_rate": 8.
|
| 286 |
-
"loss":
|
| 287 |
"step": 1125
|
| 288 |
},
|
| 289 |
{
|
| 290 |
-
"epoch":
|
| 291 |
-
"learning_rate": 8.
|
| 292 |
-
"loss":
|
| 293 |
"step": 1150
|
| 294 |
},
|
| 295 |
{
|
| 296 |
-
"epoch":
|
| 297 |
-
"learning_rate": 8.
|
| 298 |
-
"loss":
|
| 299 |
"step": 1175
|
| 300 |
},
|
| 301 |
{
|
| 302 |
-
"epoch":
|
| 303 |
-
"learning_rate": 8.
|
| 304 |
-
"loss":
|
| 305 |
"step": 1200
|
| 306 |
},
|
| 307 |
{
|
| 308 |
-
"epoch":
|
| 309 |
-
"learning_rate": 8.
|
| 310 |
-
"loss":
|
| 311 |
"step": 1225
|
| 312 |
},
|
| 313 |
{
|
| 314 |
-
"epoch":
|
| 315 |
-
"learning_rate": 8.
|
| 316 |
-
"loss":
|
| 317 |
"step": 1250
|
| 318 |
},
|
| 319 |
{
|
| 320 |
-
"epoch":
|
| 321 |
-
"learning_rate": 8.
|
| 322 |
-
"loss":
|
| 323 |
"step": 1275
|
| 324 |
},
|
| 325 |
{
|
| 326 |
-
"epoch":
|
| 327 |
-
"learning_rate": 8.
|
| 328 |
-
"loss":
|
| 329 |
"step": 1300
|
| 330 |
},
|
| 331 |
{
|
| 332 |
-
"epoch":
|
| 333 |
-
"learning_rate": 8.
|
| 334 |
-
"loss":
|
| 335 |
"step": 1325
|
| 336 |
},
|
| 337 |
{
|
| 338 |
-
"epoch":
|
| 339 |
-
"learning_rate": 8.
|
| 340 |
-
"loss":
|
| 341 |
"step": 1350
|
| 342 |
},
|
| 343 |
{
|
| 344 |
-
"epoch":
|
| 345 |
-
"learning_rate": 8.
|
| 346 |
-
"loss":
|
| 347 |
"step": 1375
|
| 348 |
},
|
| 349 |
{
|
| 350 |
-
"epoch":
|
| 351 |
-
"learning_rate": 8.
|
| 352 |
-
"loss":
|
| 353 |
"step": 1400
|
| 354 |
},
|
| 355 |
{
|
| 356 |
-
"epoch":
|
| 357 |
-
"learning_rate": 7.
|
| 358 |
-
"loss":
|
| 359 |
"step": 1425
|
| 360 |
},
|
| 361 |
{
|
| 362 |
-
"epoch":
|
| 363 |
-
"learning_rate": 7.
|
| 364 |
-
"loss":
|
| 365 |
"step": 1450
|
| 366 |
},
|
| 367 |
{
|
| 368 |
-
"epoch":
|
| 369 |
-
"learning_rate": 7.
|
| 370 |
-
"loss":
|
| 371 |
"step": 1475
|
| 372 |
},
|
| 373 |
{
|
| 374 |
-
"epoch":
|
| 375 |
-
"learning_rate": 7.
|
| 376 |
-
"loss":
|
| 377 |
"step": 1500
|
| 378 |
},
|
| 379 |
{
|
| 380 |
-
"epoch":
|
| 381 |
-
"learning_rate": 7.
|
| 382 |
-
"loss":
|
| 383 |
"step": 1525
|
| 384 |
},
|
| 385 |
{
|
| 386 |
-
"epoch":
|
| 387 |
-
"learning_rate": 7.
|
| 388 |
-
"loss":
|
| 389 |
"step": 1550
|
| 390 |
},
|
| 391 |
{
|
| 392 |
-
"epoch":
|
| 393 |
-
"learning_rate": 7.
|
| 394 |
-
"loss":
|
| 395 |
"step": 1575
|
| 396 |
},
|
| 397 |
{
|
| 398 |
-
"epoch":
|
| 399 |
-
"learning_rate": 7.
|
| 400 |
-
"loss":
|
| 401 |
"step": 1600
|
| 402 |
},
|
| 403 |
{
|
| 404 |
-
"epoch":
|
| 405 |
-
"learning_rate": 7.
|
| 406 |
-
"loss":
|
| 407 |
"step": 1625
|
| 408 |
},
|
| 409 |
{
|
| 410 |
-
"epoch":
|
| 411 |
-
"learning_rate": 7.
|
| 412 |
-
"loss":
|
| 413 |
"step": 1650
|
| 414 |
},
|
| 415 |
{
|
| 416 |
-
"epoch":
|
| 417 |
-
"learning_rate": 7.
|
| 418 |
-
"loss":
|
| 419 |
"step": 1675
|
| 420 |
},
|
| 421 |
{
|
| 422 |
-
"epoch":
|
| 423 |
-
"learning_rate": 7.
|
| 424 |
-
"loss":
|
| 425 |
"step": 1700
|
| 426 |
},
|
| 427 |
{
|
| 428 |
-
"epoch":
|
| 429 |
-
"learning_rate": 7.
|
| 430 |
-
"loss":
|
| 431 |
"step": 1725
|
| 432 |
},
|
| 433 |
{
|
| 434 |
-
"epoch":
|
| 435 |
-
"learning_rate": 7.
|
| 436 |
-
"loss":
|
| 437 |
"step": 1750
|
| 438 |
},
|
| 439 |
{
|
| 440 |
-
"epoch":
|
| 441 |
-
"learning_rate": 7.
|
| 442 |
-
"loss":
|
| 443 |
"step": 1775
|
| 444 |
},
|
| 445 |
{
|
| 446 |
-
"epoch":
|
| 447 |
-
"learning_rate": 7.
|
| 448 |
-
"loss":
|
| 449 |
"step": 1800
|
| 450 |
},
|
| 451 |
{
|
| 452 |
-
"epoch":
|
| 453 |
-
"learning_rate": 7.
|
| 454 |
-
"loss":
|
| 455 |
"step": 1825
|
| 456 |
},
|
| 457 |
{
|
| 458 |
-
"epoch":
|
| 459 |
-
"learning_rate": 7.
|
| 460 |
-
"loss":
|
| 461 |
"step": 1850
|
| 462 |
},
|
| 463 |
{
|
| 464 |
-
"epoch":
|
| 465 |
-
"learning_rate": 6.
|
| 466 |
-
"loss":
|
| 467 |
"step": 1875
|
| 468 |
},
|
| 469 |
{
|
| 470 |
-
"epoch":
|
| 471 |
-
"learning_rate": 6.
|
| 472 |
-
"loss":
|
| 473 |
"step": 1900
|
| 474 |
},
|
| 475 |
{
|
| 476 |
-
"epoch":
|
| 477 |
-
"learning_rate": 6.
|
| 478 |
-
"loss": 1.
|
| 479 |
"step": 1925
|
| 480 |
},
|
| 481 |
{
|
| 482 |
-
"epoch":
|
| 483 |
-
"learning_rate": 6.
|
| 484 |
-
"loss": 1.
|
| 485 |
"step": 1950
|
| 486 |
},
|
| 487 |
{
|
| 488 |
-
"epoch":
|
| 489 |
-
"learning_rate": 6.
|
| 490 |
-
"loss": 1.
|
| 491 |
"step": 1975
|
| 492 |
},
|
| 493 |
{
|
| 494 |
-
"epoch":
|
| 495 |
-
"learning_rate": 6.
|
| 496 |
-
"loss": 1.
|
| 497 |
"step": 2000
|
| 498 |
},
|
| 499 |
{
|
| 500 |
-
"epoch":
|
| 501 |
-
"eval_loss": 1.
|
| 502 |
-
"eval_runtime":
|
| 503 |
-
"eval_samples_per_second":
|
| 504 |
-
"eval_steps_per_second":
|
| 505 |
-
"eval_wer":
|
| 506 |
"step": 2000
|
| 507 |
},
|
| 508 |
{
|
| 509 |
-
"epoch":
|
| 510 |
-
"learning_rate": 6.
|
| 511 |
-
"loss": 1.
|
| 512 |
"step": 2025
|
| 513 |
},
|
| 514 |
{
|
| 515 |
-
"epoch":
|
| 516 |
-
"learning_rate": 6.
|
| 517 |
-
"loss": 1.
|
| 518 |
"step": 2050
|
| 519 |
},
|
| 520 |
{
|
| 521 |
-
"epoch":
|
| 522 |
-
"learning_rate": 6.
|
| 523 |
-
"loss": 1.
|
| 524 |
"step": 2075
|
| 525 |
},
|
| 526 |
{
|
| 527 |
-
"epoch":
|
| 528 |
-
"learning_rate": 6.
|
| 529 |
-
"loss": 1.
|
| 530 |
"step": 2100
|
| 531 |
},
|
| 532 |
{
|
| 533 |
-
"epoch":
|
| 534 |
-
"learning_rate": 6.
|
| 535 |
-
"loss": 1.
|
| 536 |
"step": 2125
|
| 537 |
},
|
| 538 |
{
|
| 539 |
-
"epoch":
|
| 540 |
-
"learning_rate": 6.
|
| 541 |
-
"loss": 1.
|
| 542 |
"step": 2150
|
| 543 |
},
|
| 544 |
{
|
| 545 |
-
"epoch":
|
| 546 |
-
"learning_rate": 6.
|
| 547 |
-
"loss": 1.
|
| 548 |
"step": 2175
|
| 549 |
},
|
| 550 |
{
|
| 551 |
-
"epoch":
|
| 552 |
-
"learning_rate": 6.
|
| 553 |
-
"loss": 1.
|
| 554 |
"step": 2200
|
| 555 |
},
|
| 556 |
{
|
| 557 |
-
"epoch":
|
| 558 |
-
"learning_rate": 6.
|
| 559 |
-
"loss": 1.
|
| 560 |
"step": 2225
|
| 561 |
},
|
| 562 |
{
|
| 563 |
-
"epoch":
|
| 564 |
-
"learning_rate": 6.
|
| 565 |
-
"loss": 1.
|
| 566 |
"step": 2250
|
| 567 |
},
|
| 568 |
{
|
| 569 |
-
"epoch":
|
| 570 |
-
"learning_rate": 6.
|
| 571 |
-
"loss": 1.
|
| 572 |
"step": 2275
|
| 573 |
},
|
| 574 |
{
|
| 575 |
-
"epoch":
|
| 576 |
-
"learning_rate": 6.
|
| 577 |
-
"loss": 1.
|
| 578 |
"step": 2300
|
| 579 |
},
|
| 580 |
{
|
| 581 |
-
"epoch":
|
| 582 |
-
"learning_rate": 5.
|
| 583 |
-
"loss": 1.
|
| 584 |
"step": 2325
|
| 585 |
},
|
| 586 |
{
|
| 587 |
-
"epoch":
|
| 588 |
-
"learning_rate": 5.
|
| 589 |
-
"loss": 1.
|
| 590 |
"step": 2350
|
| 591 |
},
|
| 592 |
{
|
| 593 |
-
"epoch":
|
| 594 |
-
"learning_rate": 5.
|
| 595 |
-
"loss": 1.
|
| 596 |
"step": 2375
|
| 597 |
},
|
| 598 |
{
|
| 599 |
-
"epoch":
|
| 600 |
-
"learning_rate": 5.
|
| 601 |
-
"loss": 1.
|
| 602 |
"step": 2400
|
| 603 |
},
|
| 604 |
{
|
| 605 |
-
"epoch":
|
| 606 |
-
"learning_rate": 5.
|
| 607 |
-
"loss": 1.
|
| 608 |
"step": 2425
|
| 609 |
},
|
| 610 |
{
|
| 611 |
-
"epoch":
|
| 612 |
-
"learning_rate": 5.
|
| 613 |
-
"loss": 1.
|
| 614 |
"step": 2450
|
| 615 |
},
|
| 616 |
{
|
| 617 |
-
"epoch":
|
| 618 |
-
"learning_rate": 5.
|
| 619 |
-
"loss": 1.
|
| 620 |
"step": 2475
|
| 621 |
},
|
| 622 |
{
|
| 623 |
-
"epoch":
|
| 624 |
-
"learning_rate": 5.
|
| 625 |
-
"loss": 1.
|
| 626 |
"step": 2500
|
| 627 |
},
|
| 628 |
{
|
| 629 |
-
"epoch":
|
| 630 |
-
"learning_rate": 5.
|
| 631 |
-
"loss": 1.
|
| 632 |
"step": 2525
|
| 633 |
},
|
| 634 |
{
|
| 635 |
-
"epoch":
|
| 636 |
-
"learning_rate": 5.
|
| 637 |
-
"loss": 1.
|
| 638 |
"step": 2550
|
| 639 |
},
|
| 640 |
{
|
| 641 |
-
"epoch":
|
| 642 |
-
"learning_rate": 5.
|
| 643 |
-
"loss": 1.
|
| 644 |
"step": 2575
|
| 645 |
},
|
| 646 |
{
|
| 647 |
-
"epoch":
|
| 648 |
-
"learning_rate": 5.
|
| 649 |
-
"loss": 1.
|
| 650 |
"step": 2600
|
| 651 |
},
|
| 652 |
{
|
| 653 |
-
"epoch":
|
| 654 |
-
"learning_rate": 5.
|
| 655 |
-
"loss": 1.
|
| 656 |
"step": 2625
|
| 657 |
},
|
| 658 |
{
|
| 659 |
-
"epoch":
|
| 660 |
-
"learning_rate": 5.
|
| 661 |
-
"loss": 1.
|
| 662 |
"step": 2650
|
| 663 |
},
|
| 664 |
{
|
| 665 |
-
"epoch":
|
| 666 |
-
"learning_rate": 5.
|
| 667 |
-
"loss": 1.
|
| 668 |
"step": 2675
|
| 669 |
},
|
| 670 |
{
|
| 671 |
-
"epoch":
|
| 672 |
-
"learning_rate": 5.
|
| 673 |
-
"loss": 1.
|
| 674 |
"step": 2700
|
| 675 |
},
|
| 676 |
{
|
| 677 |
-
"epoch":
|
| 678 |
-
"learning_rate": 5.
|
| 679 |
-
"loss": 1.
|
| 680 |
"step": 2725
|
| 681 |
},
|
| 682 |
{
|
| 683 |
-
"epoch":
|
| 684 |
-
"learning_rate": 5.
|
| 685 |
-
"loss": 1.
|
| 686 |
"step": 2750
|
| 687 |
},
|
| 688 |
{
|
| 689 |
-
"epoch":
|
| 690 |
-
"learning_rate": 4.
|
| 691 |
-
"loss": 1.
|
| 692 |
"step": 2775
|
| 693 |
},
|
| 694 |
{
|
| 695 |
-
"epoch":
|
| 696 |
-
"learning_rate": 4.
|
| 697 |
-
"loss": 1.
|
| 698 |
"step": 2800
|
| 699 |
},
|
| 700 |
{
|
| 701 |
-
"epoch":
|
| 702 |
-
"learning_rate": 4.
|
| 703 |
-
"loss": 1.
|
| 704 |
"step": 2825
|
| 705 |
},
|
| 706 |
{
|
| 707 |
-
"epoch":
|
| 708 |
-
"learning_rate": 4.
|
| 709 |
-
"loss": 1.
|
| 710 |
"step": 2850
|
| 711 |
},
|
| 712 |
{
|
| 713 |
-
"epoch":
|
| 714 |
-
"learning_rate": 4.
|
| 715 |
-
"loss": 1.
|
| 716 |
"step": 2875
|
| 717 |
},
|
| 718 |
{
|
| 719 |
-
"epoch":
|
| 720 |
-
"learning_rate": 4.
|
| 721 |
-
"loss": 1.
|
| 722 |
"step": 2900
|
| 723 |
},
|
| 724 |
{
|
| 725 |
-
"epoch":
|
| 726 |
-
"learning_rate": 4.
|
| 727 |
-
"loss": 1.
|
| 728 |
"step": 2925
|
| 729 |
},
|
| 730 |
{
|
| 731 |
-
"epoch":
|
| 732 |
-
"learning_rate": 4.
|
| 733 |
-
"loss": 1.
|
| 734 |
"step": 2950
|
| 735 |
},
|
| 736 |
{
|
| 737 |
-
"epoch":
|
| 738 |
-
"learning_rate": 4.
|
| 739 |
-
"loss": 1.
|
| 740 |
"step": 2975
|
| 741 |
},
|
| 742 |
{
|
| 743 |
-
"epoch":
|
| 744 |
-
"learning_rate": 4.
|
| 745 |
-
"loss": 1.
|
| 746 |
"step": 3000
|
| 747 |
},
|
| 748 |
{
|
| 749 |
-
"epoch":
|
| 750 |
-
"eval_loss":
|
| 751 |
-
"eval_runtime":
|
| 752 |
-
"eval_samples_per_second":
|
| 753 |
-
"eval_steps_per_second":
|
| 754 |
-
"eval_wer":
|
| 755 |
"step": 3000
|
| 756 |
},
|
| 757 |
{
|
| 758 |
-
"epoch":
|
| 759 |
-
"learning_rate": 4.
|
| 760 |
-
"loss": 1.
|
| 761 |
"step": 3025
|
| 762 |
},
|
| 763 |
{
|
| 764 |
-
"epoch":
|
| 765 |
-
"learning_rate": 4.
|
| 766 |
-
"loss":
|
| 767 |
"step": 3050
|
| 768 |
},
|
| 769 |
{
|
| 770 |
-
"epoch":
|
| 771 |
-
"learning_rate": 4.
|
| 772 |
-
"loss": 1.
|
| 773 |
"step": 3075
|
| 774 |
},
|
| 775 |
{
|
| 776 |
-
"epoch":
|
| 777 |
-
"learning_rate": 4.
|
| 778 |
-
"loss": 1.
|
| 779 |
"step": 3100
|
| 780 |
},
|
| 781 |
{
|
| 782 |
-
"epoch":
|
| 783 |
-
"learning_rate": 4.
|
| 784 |
-
"loss":
|
| 785 |
"step": 3125
|
| 786 |
},
|
| 787 |
{
|
| 788 |
-
"epoch":
|
| 789 |
-
"learning_rate": 4.
|
| 790 |
-
"loss": 1.
|
| 791 |
"step": 3150
|
| 792 |
},
|
| 793 |
{
|
| 794 |
-
"epoch":
|
| 795 |
-
"learning_rate": 4.
|
| 796 |
-
"loss":
|
| 797 |
"step": 3175
|
| 798 |
},
|
| 799 |
{
|
| 800 |
-
"epoch":
|
| 801 |
-
"learning_rate": 4.
|
| 802 |
-
"loss":
|
| 803 |
"step": 3200
|
| 804 |
},
|
| 805 |
{
|
| 806 |
-
"epoch":
|
| 807 |
-
"learning_rate": 3.
|
| 808 |
-
"loss":
|
| 809 |
"step": 3225
|
| 810 |
},
|
| 811 |
{
|
| 812 |
-
"epoch":
|
| 813 |
-
"learning_rate": 3.
|
| 814 |
-
"loss":
|
| 815 |
"step": 3250
|
| 816 |
},
|
| 817 |
{
|
| 818 |
-
"epoch":
|
| 819 |
-
"learning_rate": 3.
|
| 820 |
-
"loss":
|
| 821 |
"step": 3275
|
| 822 |
},
|
| 823 |
{
|
| 824 |
-
"epoch":
|
| 825 |
-
"learning_rate": 3.
|
| 826 |
-
"loss":
|
| 827 |
"step": 3300
|
| 828 |
},
|
| 829 |
{
|
| 830 |
-
"epoch":
|
| 831 |
-
"learning_rate": 3.
|
| 832 |
-
"loss":
|
| 833 |
"step": 3325
|
| 834 |
},
|
| 835 |
{
|
| 836 |
-
"epoch":
|
| 837 |
-
"learning_rate": 3.
|
| 838 |
-
"loss":
|
| 839 |
"step": 3350
|
| 840 |
},
|
| 841 |
{
|
| 842 |
-
"epoch":
|
| 843 |
-
"learning_rate": 3.
|
| 844 |
-
"loss":
|
| 845 |
"step": 3375
|
| 846 |
},
|
| 847 |
{
|
| 848 |
-
"epoch":
|
| 849 |
-
"learning_rate": 3.
|
| 850 |
-
"loss":
|
| 851 |
"step": 3400
|
| 852 |
},
|
| 853 |
{
|
| 854 |
-
"epoch":
|
| 855 |
-
"learning_rate": 3.
|
| 856 |
-
"loss":
|
| 857 |
"step": 3425
|
| 858 |
},
|
| 859 |
{
|
| 860 |
-
"epoch":
|
| 861 |
-
"learning_rate": 3.
|
| 862 |
-
"loss":
|
| 863 |
"step": 3450
|
| 864 |
},
|
| 865 |
{
|
| 866 |
-
"epoch":
|
| 867 |
-
"learning_rate": 3.
|
| 868 |
-
"loss":
|
| 869 |
"step": 3475
|
| 870 |
},
|
| 871 |
{
|
| 872 |
-
"epoch":
|
| 873 |
-
"learning_rate": 3.
|
| 874 |
-
"loss":
|
| 875 |
"step": 3500
|
| 876 |
},
|
| 877 |
{
|
| 878 |
-
"epoch":
|
| 879 |
-
"learning_rate": 3.
|
| 880 |
-
"loss":
|
| 881 |
"step": 3525
|
| 882 |
},
|
| 883 |
{
|
| 884 |
-
"epoch":
|
| 885 |
-
"learning_rate": 3.
|
| 886 |
-
"loss":
|
| 887 |
"step": 3550
|
| 888 |
},
|
| 889 |
{
|
| 890 |
-
"epoch":
|
| 891 |
-
"learning_rate": 3.
|
| 892 |
-
"loss":
|
| 893 |
"step": 3575
|
| 894 |
},
|
| 895 |
{
|
| 896 |
-
"epoch":
|
| 897 |
-
"learning_rate": 3.
|
| 898 |
-
"loss":
|
| 899 |
"step": 3600
|
| 900 |
},
|
| 901 |
{
|
| 902 |
-
"epoch":
|
| 903 |
-
"learning_rate": 3.
|
| 904 |
-
"loss":
|
| 905 |
"step": 3625
|
| 906 |
},
|
| 907 |
{
|
| 908 |
-
"epoch":
|
| 909 |
-
"learning_rate": 3.
|
| 910 |
-
"loss":
|
| 911 |
"step": 3650
|
| 912 |
},
|
| 913 |
{
|
| 914 |
-
"epoch":
|
| 915 |
-
"learning_rate": 2.
|
| 916 |
-
"loss":
|
| 917 |
"step": 3675
|
| 918 |
},
|
| 919 |
{
|
| 920 |
-
"epoch":
|
| 921 |
-
"learning_rate": 2.
|
| 922 |
-
"loss":
|
| 923 |
"step": 3700
|
| 924 |
},
|
| 925 |
{
|
| 926 |
-
"epoch":
|
| 927 |
-
"learning_rate": 2.
|
| 928 |
-
"loss":
|
| 929 |
"step": 3725
|
| 930 |
},
|
| 931 |
{
|
| 932 |
-
"epoch":
|
| 933 |
-
"learning_rate": 2.
|
| 934 |
-
"loss":
|
| 935 |
"step": 3750
|
| 936 |
},
|
| 937 |
{
|
| 938 |
-
"epoch":
|
| 939 |
-
"learning_rate": 2.
|
| 940 |
-
"loss":
|
| 941 |
"step": 3775
|
| 942 |
},
|
| 943 |
{
|
| 944 |
-
"epoch":
|
| 945 |
-
"learning_rate": 2.
|
| 946 |
-
"loss":
|
| 947 |
"step": 3800
|
| 948 |
},
|
| 949 |
{
|
| 950 |
-
"epoch":
|
| 951 |
-
"learning_rate": 2.
|
| 952 |
-
"loss":
|
| 953 |
"step": 3825
|
| 954 |
},
|
| 955 |
{
|
| 956 |
-
"epoch":
|
| 957 |
-
"learning_rate": 2.
|
| 958 |
-
"loss":
|
| 959 |
"step": 3850
|
| 960 |
},
|
| 961 |
{
|
| 962 |
-
"epoch":
|
| 963 |
-
"learning_rate": 2.
|
| 964 |
-
"loss":
|
| 965 |
"step": 3875
|
| 966 |
},
|
| 967 |
{
|
| 968 |
-
"epoch":
|
| 969 |
-
"learning_rate": 2.
|
| 970 |
-
"loss":
|
| 971 |
"step": 3900
|
| 972 |
},
|
| 973 |
{
|
| 974 |
-
"epoch":
|
| 975 |
-
"learning_rate": 2.
|
| 976 |
-
"loss":
|
| 977 |
"step": 3925
|
| 978 |
},
|
| 979 |
{
|
| 980 |
-
"epoch":
|
| 981 |
-
"learning_rate": 2.
|
| 982 |
-
"loss":
|
| 983 |
"step": 3950
|
| 984 |
},
|
| 985 |
{
|
| 986 |
-
"epoch":
|
| 987 |
-
"learning_rate": 2.
|
| 988 |
-
"loss":
|
| 989 |
"step": 3975
|
| 990 |
},
|
| 991 |
{
|
| 992 |
-
"epoch":
|
| 993 |
-
"learning_rate": 2.
|
| 994 |
-
"loss":
|
| 995 |
"step": 4000
|
| 996 |
},
|
| 997 |
{
|
| 998 |
-
"epoch":
|
| 999 |
-
"eval_loss":
|
| 1000 |
-
"eval_runtime":
|
| 1001 |
-
"eval_samples_per_second":
|
| 1002 |
-
"eval_steps_per_second":
|
| 1003 |
-
"eval_wer":
|
| 1004 |
"step": 4000
|
| 1005 |
},
|
| 1006 |
{
|
| 1007 |
-
"epoch":
|
| 1008 |
-
"learning_rate": 2.
|
| 1009 |
-
"loss":
|
| 1010 |
"step": 4025
|
| 1011 |
},
|
| 1012 |
{
|
| 1013 |
-
"epoch":
|
| 1014 |
-
"learning_rate": 2.
|
| 1015 |
-
"loss":
|
| 1016 |
"step": 4050
|
| 1017 |
},
|
| 1018 |
{
|
| 1019 |
-
"epoch":
|
| 1020 |
-
"learning_rate": 2.
|
| 1021 |
-
"loss":
|
| 1022 |
"step": 4075
|
| 1023 |
},
|
| 1024 |
{
|
| 1025 |
-
"epoch":
|
| 1026 |
-
"learning_rate": 2.
|
| 1027 |
-
"loss":
|
| 1028 |
"step": 4100
|
| 1029 |
},
|
| 1030 |
{
|
| 1031 |
-
"epoch":
|
| 1032 |
-
"learning_rate": 1.
|
| 1033 |
-
"loss":
|
| 1034 |
"step": 4125
|
| 1035 |
},
|
| 1036 |
{
|
| 1037 |
-
"epoch":
|
| 1038 |
-
"learning_rate": 1.
|
| 1039 |
-
"loss":
|
| 1040 |
"step": 4150
|
| 1041 |
},
|
| 1042 |
{
|
| 1043 |
-
"epoch":
|
| 1044 |
-
"learning_rate": 1.
|
| 1045 |
-
"loss":
|
| 1046 |
"step": 4175
|
| 1047 |
},
|
| 1048 |
{
|
| 1049 |
-
"epoch":
|
| 1050 |
-
"learning_rate": 1.
|
| 1051 |
-
"loss":
|
| 1052 |
"step": 4200
|
| 1053 |
},
|
| 1054 |
{
|
| 1055 |
-
"epoch":
|
| 1056 |
-
"learning_rate": 1.
|
| 1057 |
-
"loss":
|
| 1058 |
"step": 4225
|
| 1059 |
},
|
| 1060 |
{
|
| 1061 |
-
"epoch":
|
| 1062 |
-
"learning_rate": 1.
|
| 1063 |
-
"loss":
|
| 1064 |
"step": 4250
|
| 1065 |
},
|
| 1066 |
{
|
| 1067 |
-
"epoch":
|
| 1068 |
-
"learning_rate": 1.
|
| 1069 |
-
"loss":
|
| 1070 |
"step": 4275
|
| 1071 |
},
|
| 1072 |
{
|
| 1073 |
-
"epoch":
|
| 1074 |
-
"learning_rate": 1.
|
| 1075 |
-
"loss":
|
| 1076 |
"step": 4300
|
| 1077 |
},
|
| 1078 |
{
|
| 1079 |
-
"epoch":
|
| 1080 |
-
"learning_rate": 1.
|
| 1081 |
-
"loss":
|
| 1082 |
"step": 4325
|
| 1083 |
},
|
| 1084 |
{
|
| 1085 |
-
"epoch":
|
| 1086 |
-
"learning_rate": 1.
|
| 1087 |
-
"loss":
|
| 1088 |
"step": 4350
|
| 1089 |
},
|
| 1090 |
{
|
| 1091 |
-
"epoch":
|
| 1092 |
-
"learning_rate": 1.
|
| 1093 |
-
"loss":
|
| 1094 |
"step": 4375
|
| 1095 |
},
|
| 1096 |
{
|
| 1097 |
-
"epoch":
|
| 1098 |
-
"learning_rate": 1.
|
| 1099 |
-
"loss":
|
| 1100 |
"step": 4400
|
| 1101 |
},
|
| 1102 |
{
|
| 1103 |
-
"epoch":
|
| 1104 |
-
"learning_rate": 1.
|
| 1105 |
-
"loss":
|
| 1106 |
"step": 4425
|
| 1107 |
},
|
| 1108 |
{
|
| 1109 |
-
"epoch":
|
| 1110 |
-
"learning_rate": 1.
|
| 1111 |
-
"loss":
|
| 1112 |
"step": 4450
|
| 1113 |
},
|
| 1114 |
{
|
| 1115 |
-
"epoch":
|
| 1116 |
-
"learning_rate": 1.
|
| 1117 |
-
"loss":
|
| 1118 |
"step": 4475
|
| 1119 |
},
|
| 1120 |
{
|
| 1121 |
-
"epoch":
|
| 1122 |
-
"learning_rate": 1.
|
| 1123 |
-
"loss":
|
| 1124 |
"step": 4500
|
| 1125 |
},
|
| 1126 |
{
|
| 1127 |
-
"epoch":
|
| 1128 |
-
"learning_rate": 1.
|
| 1129 |
-
"loss":
|
| 1130 |
"step": 4525
|
| 1131 |
},
|
| 1132 |
{
|
| 1133 |
-
"epoch":
|
| 1134 |
-
"learning_rate": 1.
|
| 1135 |
-
"loss":
|
| 1136 |
"step": 4550
|
| 1137 |
},
|
| 1138 |
{
|
| 1139 |
-
"epoch":
|
| 1140 |
-
"learning_rate": 9.
|
| 1141 |
-
"loss":
|
| 1142 |
"step": 4575
|
| 1143 |
},
|
| 1144 |
{
|
| 1145 |
-
"epoch":
|
| 1146 |
-
"learning_rate": 9.
|
| 1147 |
-
"loss":
|
| 1148 |
"step": 4600
|
| 1149 |
},
|
| 1150 |
{
|
| 1151 |
-
"epoch":
|
| 1152 |
-
"learning_rate": 8.
|
| 1153 |
-
"loss":
|
| 1154 |
"step": 4625
|
| 1155 |
},
|
| 1156 |
{
|
| 1157 |
-
"epoch":
|
| 1158 |
-
"learning_rate": 7.
|
| 1159 |
-
"loss":
|
| 1160 |
"step": 4650
|
| 1161 |
},
|
| 1162 |
{
|
| 1163 |
-
"epoch":
|
| 1164 |
-
"learning_rate": 7.
|
| 1165 |
-
"loss":
|
| 1166 |
"step": 4675
|
| 1167 |
},
|
| 1168 |
{
|
| 1169 |
-
"epoch":
|
| 1170 |
-
"learning_rate": 6.
|
| 1171 |
-
"loss":
|
| 1172 |
"step": 4700
|
| 1173 |
},
|
| 1174 |
{
|
| 1175 |
-
"epoch":
|
| 1176 |
-
"learning_rate": 6.
|
| 1177 |
-
"loss":
|
| 1178 |
"step": 4725
|
| 1179 |
},
|
| 1180 |
{
|
| 1181 |
-
"epoch":
|
| 1182 |
-
"learning_rate": 5.
|
| 1183 |
-
"loss":
|
| 1184 |
"step": 4750
|
| 1185 |
},
|
| 1186 |
{
|
| 1187 |
-
"epoch":
|
| 1188 |
-
"learning_rate": 5.
|
| 1189 |
-
"loss":
|
| 1190 |
"step": 4775
|
| 1191 |
},
|
| 1192 |
{
|
| 1193 |
-
"epoch":
|
| 1194 |
-
"learning_rate": 4.
|
| 1195 |
-
"loss":
|
| 1196 |
"step": 4800
|
| 1197 |
},
|
| 1198 |
{
|
| 1199 |
-
"epoch":
|
| 1200 |
-
"learning_rate": 4.
|
| 1201 |
-
"loss":
|
| 1202 |
"step": 4825
|
| 1203 |
},
|
| 1204 |
{
|
| 1205 |
-
"epoch":
|
| 1206 |
-
"learning_rate": 3.
|
| 1207 |
-
"loss":
|
| 1208 |
"step": 4850
|
| 1209 |
},
|
| 1210 |
{
|
| 1211 |
-
"epoch":
|
| 1212 |
-
"learning_rate": 2.
|
| 1213 |
-
"loss":
|
| 1214 |
"step": 4875
|
| 1215 |
},
|
| 1216 |
{
|
| 1217 |
-
"epoch":
|
| 1218 |
-
"learning_rate": 2.
|
| 1219 |
-
"loss":
|
| 1220 |
"step": 4900
|
| 1221 |
},
|
| 1222 |
{
|
| 1223 |
-
"epoch":
|
| 1224 |
-
"learning_rate": 1.
|
| 1225 |
-
"loss":
|
| 1226 |
"step": 4925
|
| 1227 |
},
|
| 1228 |
{
|
| 1229 |
-
"epoch":
|
| 1230 |
-
"learning_rate": 1.
|
| 1231 |
-
"loss":
|
| 1232 |
"step": 4950
|
| 1233 |
},
|
| 1234 |
{
|
| 1235 |
-
"epoch":
|
| 1236 |
-
"learning_rate":
|
| 1237 |
-
"loss":
|
| 1238 |
"step": 4975
|
| 1239 |
},
|
| 1240 |
{
|
| 1241 |
-
"epoch":
|
| 1242 |
-
"learning_rate": 1.
|
| 1243 |
-
"loss":
|
| 1244 |
"step": 5000
|
| 1245 |
},
|
| 1246 |
{
|
| 1247 |
-
"epoch":
|
| 1248 |
-
"eval_loss":
|
| 1249 |
-
"eval_runtime":
|
| 1250 |
-
"eval_samples_per_second":
|
| 1251 |
-
"eval_steps_per_second":
|
| 1252 |
-
"eval_wer":
|
| 1253 |
"step": 5000
|
| 1254 |
},
|
| 1255 |
{
|
| 1256 |
-
"epoch":
|
| 1257 |
"step": 5000,
|
| 1258 |
-
"total_flos": 9.
|
| 1259 |
-
"train_loss": 1.
|
| 1260 |
-
"train_runtime":
|
| 1261 |
-
"train_samples_per_second":
|
| 1262 |
-
"train_steps_per_second":
|
| 1263 |
}
|
| 1264 |
],
|
| 1265 |
"max_steps": 5000,
|
| 1266 |
-
"num_train_epochs":
|
| 1267 |
-
"total_flos": 9.
|
| 1268 |
"trial_name": null,
|
| 1269 |
"trial_params": null
|
| 1270 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 76.37636988522145,
|
| 3 |
"best_model_checkpoint": "./output/small/yt-special-batch8-tiny/checkpoint-5000",
|
| 4 |
+
"epoch": 0.8588114050154586,
|
| 5 |
"global_step": 5000,
|
| 6 |
"is_hyper_param_search": false,
|
| 7 |
"is_local_process_zero": true,
|
| 8 |
"is_world_process_zero": true,
|
| 9 |
"log_history": [
|
| 10 |
{
|
| 11 |
+
"epoch": 0.0,
|
| 12 |
+
"learning_rate": 4.0000000000000003e-07,
|
| 13 |
+
"loss": 3.0375,
|
| 14 |
"step": 25
|
| 15 |
},
|
| 16 |
{
|
| 17 |
+
"epoch": 0.01,
|
| 18 |
+
"learning_rate": 9.000000000000001e-07,
|
| 19 |
+
"loss": 2.8516,
|
| 20 |
"step": 50
|
| 21 |
},
|
| 22 |
{
|
| 23 |
+
"epoch": 0.01,
|
| 24 |
+
"learning_rate": 1.4000000000000001e-06,
|
| 25 |
+
"loss": 2.8996,
|
| 26 |
"step": 75
|
| 27 |
},
|
| 28 |
{
|
| 29 |
+
"epoch": 0.02,
|
| 30 |
+
"learning_rate": 1.9000000000000002e-06,
|
| 31 |
+
"loss": 2.413,
|
| 32 |
"step": 100
|
| 33 |
},
|
| 34 |
{
|
| 35 |
+
"epoch": 0.02,
|
| 36 |
+
"learning_rate": 2.4000000000000003e-06,
|
| 37 |
+
"loss": 2.2213,
|
| 38 |
"step": 125
|
| 39 |
},
|
| 40 |
{
|
| 41 |
+
"epoch": 0.03,
|
| 42 |
+
"learning_rate": 2.9e-06,
|
| 43 |
+
"loss": 2.0832,
|
| 44 |
"step": 150
|
| 45 |
},
|
| 46 |
{
|
| 47 |
+
"epoch": 0.03,
|
| 48 |
+
"learning_rate": 3.4000000000000005e-06,
|
| 49 |
+
"loss": 1.7823,
|
| 50 |
"step": 175
|
| 51 |
},
|
| 52 |
{
|
| 53 |
+
"epoch": 0.03,
|
| 54 |
+
"learning_rate": 3.900000000000001e-06,
|
| 55 |
+
"loss": 1.7617,
|
| 56 |
"step": 200
|
| 57 |
},
|
| 58 |
{
|
| 59 |
+
"epoch": 0.04,
|
| 60 |
+
"learning_rate": 4.4e-06,
|
| 61 |
+
"loss": 1.8501,
|
| 62 |
"step": 225
|
| 63 |
},
|
| 64 |
{
|
| 65 |
+
"epoch": 0.04,
|
| 66 |
+
"learning_rate": 4.9000000000000005e-06,
|
| 67 |
+
"loss": 1.7876,
|
| 68 |
"step": 250
|
| 69 |
},
|
| 70 |
{
|
| 71 |
+
"epoch": 0.05,
|
| 72 |
+
"learning_rate": 5.400000000000001e-06,
|
| 73 |
+
"loss": 1.7035,
|
| 74 |
"step": 275
|
| 75 |
},
|
| 76 |
{
|
| 77 |
+
"epoch": 0.05,
|
| 78 |
+
"learning_rate": 5.9e-06,
|
| 79 |
+
"loss": 1.7522,
|
| 80 |
"step": 300
|
| 81 |
},
|
| 82 |
{
|
| 83 |
+
"epoch": 0.06,
|
| 84 |
+
"learning_rate": 6.4000000000000006e-06,
|
| 85 |
+
"loss": 1.584,
|
| 86 |
"step": 325
|
| 87 |
},
|
| 88 |
{
|
| 89 |
+
"epoch": 0.06,
|
| 90 |
+
"learning_rate": 6.9e-06,
|
| 91 |
+
"loss": 1.6716,
|
| 92 |
"step": 350
|
| 93 |
},
|
| 94 |
{
|
| 95 |
+
"epoch": 0.06,
|
| 96 |
+
"learning_rate": 7.4e-06,
|
| 97 |
+
"loss": 1.796,
|
| 98 |
"step": 375
|
| 99 |
},
|
| 100 |
{
|
| 101 |
+
"epoch": 0.07,
|
| 102 |
+
"learning_rate": 7.9e-06,
|
| 103 |
+
"loss": 1.5802,
|
| 104 |
"step": 400
|
| 105 |
},
|
| 106 |
{
|
| 107 |
+
"epoch": 0.07,
|
| 108 |
+
"learning_rate": 8.400000000000001e-06,
|
| 109 |
+
"loss": 1.4449,
|
| 110 |
"step": 425
|
| 111 |
},
|
| 112 |
{
|
| 113 |
+
"epoch": 0.08,
|
| 114 |
+
"learning_rate": 8.900000000000001e-06,
|
| 115 |
+
"loss": 1.6765,
|
| 116 |
"step": 450
|
| 117 |
},
|
| 118 |
{
|
| 119 |
+
"epoch": 0.08,
|
| 120 |
+
"learning_rate": 9.4e-06,
|
| 121 |
+
"loss": 1.6156,
|
| 122 |
"step": 475
|
| 123 |
},
|
| 124 |
{
|
| 125 |
+
"epoch": 0.09,
|
| 126 |
+
"learning_rate": 9.9e-06,
|
| 127 |
+
"loss": 1.5887,
|
| 128 |
"step": 500
|
| 129 |
},
|
| 130 |
{
|
| 131 |
+
"epoch": 0.09,
|
| 132 |
+
"learning_rate": 9.955555555555556e-06,
|
| 133 |
+
"loss": 1.5356,
|
| 134 |
"step": 525
|
| 135 |
},
|
| 136 |
{
|
| 137 |
+
"epoch": 0.09,
|
| 138 |
+
"learning_rate": 9.9e-06,
|
| 139 |
+
"loss": 1.5713,
|
| 140 |
"step": 550
|
| 141 |
},
|
| 142 |
{
|
| 143 |
+
"epoch": 0.1,
|
| 144 |
+
"learning_rate": 9.844444444444446e-06,
|
| 145 |
+
"loss": 1.5336,
|
| 146 |
"step": 575
|
| 147 |
},
|
| 148 |
{
|
| 149 |
+
"epoch": 0.1,
|
| 150 |
+
"learning_rate": 9.78888888888889e-06,
|
| 151 |
+
"loss": 1.6768,
|
| 152 |
"step": 600
|
| 153 |
},
|
| 154 |
{
|
| 155 |
+
"epoch": 0.11,
|
| 156 |
+
"learning_rate": 9.733333333333334e-06,
|
| 157 |
+
"loss": 1.4941,
|
| 158 |
"step": 625
|
| 159 |
},
|
| 160 |
{
|
| 161 |
+
"epoch": 0.11,
|
| 162 |
+
"learning_rate": 9.677777777777778e-06,
|
| 163 |
+
"loss": 1.5387,
|
| 164 |
"step": 650
|
| 165 |
},
|
| 166 |
{
|
| 167 |
+
"epoch": 0.12,
|
| 168 |
+
"learning_rate": 9.622222222222222e-06,
|
| 169 |
+
"loss": 1.5816,
|
| 170 |
"step": 675
|
| 171 |
},
|
| 172 |
{
|
| 173 |
+
"epoch": 0.12,
|
| 174 |
+
"learning_rate": 9.566666666666668e-06,
|
| 175 |
+
"loss": 1.4096,
|
| 176 |
"step": 700
|
| 177 |
},
|
| 178 |
{
|
| 179 |
+
"epoch": 0.12,
|
| 180 |
+
"learning_rate": 9.511111111111112e-06,
|
| 181 |
+
"loss": 1.4932,
|
| 182 |
"step": 725
|
| 183 |
},
|
| 184 |
{
|
| 185 |
+
"epoch": 0.13,
|
| 186 |
+
"learning_rate": 9.455555555555557e-06,
|
| 187 |
+
"loss": 1.5254,
|
| 188 |
"step": 750
|
| 189 |
},
|
| 190 |
{
|
| 191 |
+
"epoch": 0.13,
|
| 192 |
+
"learning_rate": 9.4e-06,
|
| 193 |
+
"loss": 1.3965,
|
| 194 |
"step": 775
|
| 195 |
},
|
| 196 |
{
|
| 197 |
+
"epoch": 0.14,
|
| 198 |
+
"learning_rate": 9.344444444444446e-06,
|
| 199 |
+
"loss": 1.4093,
|
| 200 |
"step": 800
|
| 201 |
},
|
| 202 |
{
|
| 203 |
+
"epoch": 0.14,
|
| 204 |
+
"learning_rate": 9.28888888888889e-06,
|
| 205 |
+
"loss": 1.3946,
|
| 206 |
"step": 825
|
| 207 |
},
|
| 208 |
{
|
| 209 |
+
"epoch": 0.15,
|
| 210 |
+
"learning_rate": 9.233333333333334e-06,
|
| 211 |
+
"loss": 1.3491,
|
| 212 |
"step": 850
|
| 213 |
},
|
| 214 |
{
|
| 215 |
+
"epoch": 0.15,
|
| 216 |
+
"learning_rate": 9.17777777777778e-06,
|
| 217 |
+
"loss": 1.4431,
|
| 218 |
"step": 875
|
| 219 |
},
|
| 220 |
{
|
| 221 |
+
"epoch": 0.15,
|
| 222 |
+
"learning_rate": 9.122222222222223e-06,
|
| 223 |
+
"loss": 1.4492,
|
| 224 |
"step": 900
|
| 225 |
},
|
| 226 |
{
|
| 227 |
+
"epoch": 0.16,
|
| 228 |
+
"learning_rate": 9.066666666666667e-06,
|
| 229 |
+
"loss": 1.4466,
|
| 230 |
"step": 925
|
| 231 |
},
|
| 232 |
{
|
| 233 |
+
"epoch": 0.16,
|
| 234 |
+
"learning_rate": 9.011111111111111e-06,
|
| 235 |
+
"loss": 1.3713,
|
| 236 |
"step": 950
|
| 237 |
},
|
| 238 |
{
|
| 239 |
+
"epoch": 0.17,
|
| 240 |
+
"learning_rate": 8.955555555555555e-06,
|
| 241 |
+
"loss": 1.4049,
|
| 242 |
"step": 975
|
| 243 |
},
|
| 244 |
{
|
| 245 |
+
"epoch": 0.17,
|
| 246 |
+
"learning_rate": 8.900000000000001e-06,
|
| 247 |
+
"loss": 1.4295,
|
| 248 |
"step": 1000
|
| 249 |
},
|
| 250 |
{
|
| 251 |
+
"epoch": 0.17,
|
| 252 |
+
"eval_loss": 1.4093937873840332,
|
| 253 |
+
"eval_runtime": 278.5863,
|
| 254 |
+
"eval_samples_per_second": 8.798,
|
| 255 |
+
"eval_steps_per_second": 2.2,
|
| 256 |
+
"eval_wer": 110.15498346410739,
|
| 257 |
"step": 1000
|
| 258 |
},
|
| 259 |
{
|
| 260 |
+
"epoch": 0.18,
|
| 261 |
+
"learning_rate": 8.844444444444445e-06,
|
| 262 |
+
"loss": 1.5387,
|
| 263 |
"step": 1025
|
| 264 |
},
|
| 265 |
{
|
| 266 |
+
"epoch": 0.18,
|
| 267 |
+
"learning_rate": 8.788888888888891e-06,
|
| 268 |
+
"loss": 1.3363,
|
| 269 |
"step": 1050
|
| 270 |
},
|
| 271 |
{
|
| 272 |
+
"epoch": 0.18,
|
| 273 |
+
"learning_rate": 8.733333333333333e-06,
|
| 274 |
+
"loss": 1.4644,
|
| 275 |
"step": 1075
|
| 276 |
},
|
| 277 |
{
|
| 278 |
+
"epoch": 0.19,
|
| 279 |
+
"learning_rate": 8.677777777777779e-06,
|
| 280 |
+
"loss": 1.3795,
|
| 281 |
"step": 1100
|
| 282 |
},
|
| 283 |
{
|
| 284 |
+
"epoch": 0.19,
|
| 285 |
+
"learning_rate": 8.622222222222223e-06,
|
| 286 |
+
"loss": 1.3536,
|
| 287 |
"step": 1125
|
| 288 |
},
|
| 289 |
{
|
| 290 |
+
"epoch": 0.2,
|
| 291 |
+
"learning_rate": 8.566666666666667e-06,
|
| 292 |
+
"loss": 1.5356,
|
| 293 |
"step": 1150
|
| 294 |
},
|
| 295 |
{
|
| 296 |
+
"epoch": 0.2,
|
| 297 |
+
"learning_rate": 8.511111111111113e-06,
|
| 298 |
+
"loss": 1.5368,
|
| 299 |
"step": 1175
|
| 300 |
},
|
| 301 |
{
|
| 302 |
+
"epoch": 0.21,
|
| 303 |
+
"learning_rate": 8.455555555555555e-06,
|
| 304 |
+
"loss": 1.3584,
|
| 305 |
"step": 1200
|
| 306 |
},
|
| 307 |
{
|
| 308 |
+
"epoch": 0.21,
|
| 309 |
+
"learning_rate": 8.400000000000001e-06,
|
| 310 |
+
"loss": 1.3795,
|
| 311 |
"step": 1225
|
| 312 |
},
|
| 313 |
{
|
| 314 |
+
"epoch": 0.21,
|
| 315 |
+
"learning_rate": 8.344444444444445e-06,
|
| 316 |
+
"loss": 1.5068,
|
| 317 |
"step": 1250
|
| 318 |
},
|
| 319 |
{
|
| 320 |
+
"epoch": 0.22,
|
| 321 |
+
"learning_rate": 8.288888888888889e-06,
|
| 322 |
+
"loss": 1.2671,
|
| 323 |
"step": 1275
|
| 324 |
},
|
| 325 |
{
|
| 326 |
+
"epoch": 0.22,
|
| 327 |
+
"learning_rate": 8.233333333333335e-06,
|
| 328 |
+
"loss": 1.4027,
|
| 329 |
"step": 1300
|
| 330 |
},
|
| 331 |
{
|
| 332 |
+
"epoch": 0.23,
|
| 333 |
+
"learning_rate": 8.177777777777779e-06,
|
| 334 |
+
"loss": 1.3407,
|
| 335 |
"step": 1325
|
| 336 |
},
|
| 337 |
{
|
| 338 |
+
"epoch": 0.23,
|
| 339 |
+
"learning_rate": 8.122222222222223e-06,
|
| 340 |
+
"loss": 1.3717,
|
| 341 |
"step": 1350
|
| 342 |
},
|
| 343 |
{
|
| 344 |
+
"epoch": 0.24,
|
| 345 |
+
"learning_rate": 8.066666666666667e-06,
|
| 346 |
+
"loss": 1.4011,
|
| 347 |
"step": 1375
|
| 348 |
},
|
| 349 |
{
|
| 350 |
+
"epoch": 0.24,
|
| 351 |
+
"learning_rate": 8.011111111111113e-06,
|
| 352 |
+
"loss": 1.3397,
|
| 353 |
"step": 1400
|
| 354 |
},
|
| 355 |
{
|
| 356 |
+
"epoch": 0.24,
|
| 357 |
+
"learning_rate": 7.955555555555557e-06,
|
| 358 |
+
"loss": 1.4166,
|
| 359 |
"step": 1425
|
| 360 |
},
|
| 361 |
{
|
| 362 |
+
"epoch": 0.25,
|
| 363 |
+
"learning_rate": 7.9e-06,
|
| 364 |
+
"loss": 1.3022,
|
| 365 |
"step": 1450
|
| 366 |
},
|
| 367 |
{
|
| 368 |
+
"epoch": 0.25,
|
| 369 |
+
"learning_rate": 7.844444444444446e-06,
|
| 370 |
+
"loss": 1.3544,
|
| 371 |
"step": 1475
|
| 372 |
},
|
| 373 |
{
|
| 374 |
+
"epoch": 0.26,
|
| 375 |
+
"learning_rate": 7.788888888888889e-06,
|
| 376 |
+
"loss": 1.3412,
|
| 377 |
"step": 1500
|
| 378 |
},
|
| 379 |
{
|
| 380 |
+
"epoch": 0.26,
|
| 381 |
+
"learning_rate": 7.733333333333334e-06,
|
| 382 |
+
"loss": 1.5112,
|
| 383 |
"step": 1525
|
| 384 |
},
|
| 385 |
{
|
| 386 |
+
"epoch": 0.27,
|
| 387 |
+
"learning_rate": 7.677777777777778e-06,
|
| 388 |
+
"loss": 1.3038,
|
| 389 |
"step": 1550
|
| 390 |
},
|
| 391 |
{
|
| 392 |
+
"epoch": 0.27,
|
| 393 |
+
"learning_rate": 7.622222222222223e-06,
|
| 394 |
+
"loss": 1.3752,
|
| 395 |
"step": 1575
|
| 396 |
},
|
| 397 |
{
|
| 398 |
+
"epoch": 0.27,
|
| 399 |
+
"learning_rate": 7.566666666666667e-06,
|
| 400 |
+
"loss": 1.3446,
|
| 401 |
"step": 1600
|
| 402 |
},
|
| 403 |
{
|
| 404 |
+
"epoch": 0.28,
|
| 405 |
+
"learning_rate": 7.511111111111111e-06,
|
| 406 |
+
"loss": 1.2387,
|
| 407 |
"step": 1625
|
| 408 |
},
|
| 409 |
{
|
| 410 |
+
"epoch": 0.28,
|
| 411 |
+
"learning_rate": 7.455555555555556e-06,
|
| 412 |
+
"loss": 1.3236,
|
| 413 |
"step": 1650
|
| 414 |
},
|
| 415 |
{
|
| 416 |
+
"epoch": 0.29,
|
| 417 |
+
"learning_rate": 7.4e-06,
|
| 418 |
+
"loss": 1.3304,
|
| 419 |
"step": 1675
|
| 420 |
},
|
| 421 |
{
|
| 422 |
+
"epoch": 0.29,
|
| 423 |
+
"learning_rate": 7.344444444444445e-06,
|
| 424 |
+
"loss": 1.2388,
|
| 425 |
"step": 1700
|
| 426 |
},
|
| 427 |
{
|
| 428 |
+
"epoch": 0.3,
|
| 429 |
+
"learning_rate": 7.28888888888889e-06,
|
| 430 |
+
"loss": 1.3252,
|
| 431 |
"step": 1725
|
| 432 |
},
|
| 433 |
{
|
| 434 |
+
"epoch": 0.3,
|
| 435 |
+
"learning_rate": 7.233333333333334e-06,
|
| 436 |
+
"loss": 1.2723,
|
| 437 |
"step": 1750
|
| 438 |
},
|
| 439 |
{
|
| 440 |
+
"epoch": 0.3,
|
| 441 |
+
"learning_rate": 7.177777777777778e-06,
|
| 442 |
+
"loss": 1.3858,
|
| 443 |
"step": 1775
|
| 444 |
},
|
| 445 |
{
|
| 446 |
+
"epoch": 0.31,
|
| 447 |
+
"learning_rate": 7.122222222222222e-06,
|
| 448 |
+
"loss": 1.2257,
|
| 449 |
"step": 1800
|
| 450 |
},
|
| 451 |
{
|
| 452 |
+
"epoch": 0.31,
|
| 453 |
+
"learning_rate": 7.066666666666667e-06,
|
| 454 |
+
"loss": 1.3428,
|
| 455 |
"step": 1825
|
| 456 |
},
|
| 457 |
{
|
| 458 |
+
"epoch": 0.32,
|
| 459 |
+
"learning_rate": 7.011111111111112e-06,
|
| 460 |
+
"loss": 1.3284,
|
| 461 |
"step": 1850
|
| 462 |
},
|
| 463 |
{
|
| 464 |
+
"epoch": 0.32,
|
| 465 |
+
"learning_rate": 6.955555555555557e-06,
|
| 466 |
+
"loss": 1.2452,
|
| 467 |
"step": 1875
|
| 468 |
},
|
| 469 |
{
|
| 470 |
+
"epoch": 0.33,
|
| 471 |
+
"learning_rate": 6.9e-06,
|
| 472 |
+
"loss": 1.3218,
|
| 473 |
"step": 1900
|
| 474 |
},
|
| 475 |
{
|
| 476 |
+
"epoch": 0.33,
|
| 477 |
+
"learning_rate": 6.844444444444445e-06,
|
| 478 |
+
"loss": 1.3243,
|
| 479 |
"step": 1925
|
| 480 |
},
|
| 481 |
{
|
| 482 |
+
"epoch": 0.33,
|
| 483 |
+
"learning_rate": 6.788888888888889e-06,
|
| 484 |
+
"loss": 1.3978,
|
| 485 |
"step": 1950
|
| 486 |
},
|
| 487 |
{
|
| 488 |
+
"epoch": 0.34,
|
| 489 |
+
"learning_rate": 6.733333333333334e-06,
|
| 490 |
+
"loss": 1.2782,
|
| 491 |
"step": 1975
|
| 492 |
},
|
| 493 |
{
|
| 494 |
+
"epoch": 0.34,
|
| 495 |
+
"learning_rate": 6.677777777777779e-06,
|
| 496 |
+
"loss": 1.3042,
|
| 497 |
"step": 2000
|
| 498 |
},
|
| 499 |
{
|
| 500 |
+
"epoch": 0.34,
|
| 501 |
+
"eval_loss": 1.2886050939559937,
|
| 502 |
+
"eval_runtime": 281.1051,
|
| 503 |
+
"eval_samples_per_second": 8.719,
|
| 504 |
+
"eval_steps_per_second": 2.181,
|
| 505 |
+
"eval_wer": 86.29142078983205,
|
| 506 |
"step": 2000
|
| 507 |
},
|
| 508 |
{
|
| 509 |
+
"epoch": 0.35,
|
| 510 |
+
"learning_rate": 6.6222222222222236e-06,
|
| 511 |
+
"loss": 1.2557,
|
| 512 |
"step": 2025
|
| 513 |
},
|
| 514 |
{
|
| 515 |
+
"epoch": 0.35,
|
| 516 |
+
"learning_rate": 6.566666666666667e-06,
|
| 517 |
+
"loss": 1.3275,
|
| 518 |
"step": 2050
|
| 519 |
},
|
| 520 |
{
|
| 521 |
+
"epoch": 0.36,
|
| 522 |
+
"learning_rate": 6.511111111111112e-06,
|
| 523 |
+
"loss": 1.3171,
|
| 524 |
"step": 2075
|
| 525 |
},
|
| 526 |
{
|
| 527 |
+
"epoch": 0.36,
|
| 528 |
+
"learning_rate": 6.455555555555556e-06,
|
| 529 |
+
"loss": 1.2166,
|
| 530 |
"step": 2100
|
| 531 |
},
|
| 532 |
{
|
| 533 |
+
"epoch": 0.36,
|
| 534 |
+
"learning_rate": 6.4000000000000006e-06,
|
| 535 |
+
"loss": 1.1869,
|
| 536 |
"step": 2125
|
| 537 |
},
|
| 538 |
{
|
| 539 |
+
"epoch": 0.37,
|
| 540 |
+
"learning_rate": 6.3444444444444454e-06,
|
| 541 |
+
"loss": 1.3423,
|
| 542 |
"step": 2150
|
| 543 |
},
|
| 544 |
{
|
| 545 |
+
"epoch": 0.37,
|
| 546 |
+
"learning_rate": 6.28888888888889e-06,
|
| 547 |
+
"loss": 1.3287,
|
| 548 |
"step": 2175
|
| 549 |
},
|
| 550 |
{
|
| 551 |
+
"epoch": 0.38,
|
| 552 |
+
"learning_rate": 6.2333333333333335e-06,
|
| 553 |
+
"loss": 1.2774,
|
| 554 |
"step": 2200
|
| 555 |
},
|
| 556 |
{
|
| 557 |
+
"epoch": 0.38,
|
| 558 |
+
"learning_rate": 6.177777777777778e-06,
|
| 559 |
+
"loss": 1.2855,
|
| 560 |
"step": 2225
|
| 561 |
},
|
| 562 |
{
|
| 563 |
+
"epoch": 0.39,
|
| 564 |
+
"learning_rate": 6.1222222222222224e-06,
|
| 565 |
+
"loss": 1.2227,
|
| 566 |
"step": 2250
|
| 567 |
},
|
| 568 |
{
|
| 569 |
+
"epoch": 0.39,
|
| 570 |
+
"learning_rate": 6.066666666666667e-06,
|
| 571 |
+
"loss": 1.198,
|
| 572 |
"step": 2275
|
| 573 |
},
|
| 574 |
{
|
| 575 |
+
"epoch": 0.4,
|
| 576 |
+
"learning_rate": 6.011111111111112e-06,
|
| 577 |
+
"loss": 1.3709,
|
| 578 |
"step": 2300
|
| 579 |
},
|
| 580 |
{
|
| 581 |
+
"epoch": 0.4,
|
| 582 |
+
"learning_rate": 5.955555555555555e-06,
|
| 583 |
+
"loss": 1.3011,
|
| 584 |
"step": 2325
|
| 585 |
},
|
| 586 |
{
|
| 587 |
+
"epoch": 0.4,
|
| 588 |
+
"learning_rate": 5.9e-06,
|
| 589 |
+
"loss": 1.3681,
|
| 590 |
"step": 2350
|
| 591 |
},
|
| 592 |
{
|
| 593 |
+
"epoch": 0.41,
|
| 594 |
+
"learning_rate": 5.844444444444445e-06,
|
| 595 |
+
"loss": 1.3086,
|
| 596 |
"step": 2375
|
| 597 |
},
|
| 598 |
{
|
| 599 |
+
"epoch": 0.41,
|
| 600 |
+
"learning_rate": 5.788888888888889e-06,
|
| 601 |
+
"loss": 1.2167,
|
| 602 |
"step": 2400
|
| 603 |
},
|
| 604 |
{
|
| 605 |
+
"epoch": 0.42,
|
| 606 |
+
"learning_rate": 5.733333333333334e-06,
|
| 607 |
+
"loss": 1.2282,
|
| 608 |
"step": 2425
|
| 609 |
},
|
| 610 |
{
|
| 611 |
+
"epoch": 0.42,
|
| 612 |
+
"learning_rate": 5.677777777777779e-06,
|
| 613 |
+
"loss": 1.266,
|
| 614 |
"step": 2450
|
| 615 |
},
|
| 616 |
{
|
| 617 |
+
"epoch": 0.43,
|
| 618 |
+
"learning_rate": 5.622222222222222e-06,
|
| 619 |
+
"loss": 1.2449,
|
| 620 |
"step": 2475
|
| 621 |
},
|
| 622 |
{
|
| 623 |
+
"epoch": 0.43,
|
| 624 |
+
"learning_rate": 5.566666666666667e-06,
|
| 625 |
+
"loss": 1.2406,
|
| 626 |
"step": 2500
|
| 627 |
},
|
| 628 |
{
|
| 629 |
+
"epoch": 0.43,
|
| 630 |
+
"learning_rate": 5.511111111111112e-06,
|
| 631 |
+
"loss": 1.2414,
|
| 632 |
"step": 2525
|
| 633 |
},
|
| 634 |
{
|
| 635 |
+
"epoch": 0.44,
|
| 636 |
+
"learning_rate": 5.455555555555556e-06,
|
| 637 |
+
"loss": 1.3149,
|
| 638 |
"step": 2550
|
| 639 |
},
|
| 640 |
{
|
| 641 |
+
"epoch": 0.44,
|
| 642 |
+
"learning_rate": 5.400000000000001e-06,
|
| 643 |
+
"loss": 1.2321,
|
| 644 |
"step": 2575
|
| 645 |
},
|
| 646 |
{
|
| 647 |
+
"epoch": 0.45,
|
| 648 |
+
"learning_rate": 5.344444444444446e-06,
|
| 649 |
+
"loss": 1.1998,
|
| 650 |
"step": 2600
|
| 651 |
},
|
| 652 |
{
|
| 653 |
+
"epoch": 0.45,
|
| 654 |
+
"learning_rate": 5.288888888888889e-06,
|
| 655 |
+
"loss": 1.2292,
|
| 656 |
"step": 2625
|
| 657 |
},
|
| 658 |
{
|
| 659 |
+
"epoch": 0.46,
|
| 660 |
+
"learning_rate": 5.233333333333334e-06,
|
| 661 |
+
"loss": 1.3701,
|
| 662 |
"step": 2650
|
| 663 |
},
|
| 664 |
{
|
| 665 |
+
"epoch": 0.46,
|
| 666 |
+
"learning_rate": 5.177777777777779e-06,
|
| 667 |
+
"loss": 1.2846,
|
| 668 |
"step": 2675
|
| 669 |
},
|
| 670 |
{
|
| 671 |
+
"epoch": 0.46,
|
| 672 |
+
"learning_rate": 5.122222222222223e-06,
|
| 673 |
+
"loss": 1.2457,
|
| 674 |
"step": 2700
|
| 675 |
},
|
| 676 |
{
|
| 677 |
+
"epoch": 0.47,
|
| 678 |
+
"learning_rate": 5.0666666666666676e-06,
|
| 679 |
+
"loss": 1.2912,
|
| 680 |
"step": 2725
|
| 681 |
},
|
| 682 |
{
|
| 683 |
+
"epoch": 0.47,
|
| 684 |
+
"learning_rate": 5.011111111111111e-06,
|
| 685 |
+
"loss": 1.2478,
|
| 686 |
"step": 2750
|
| 687 |
},
|
| 688 |
{
|
| 689 |
+
"epoch": 0.48,
|
| 690 |
+
"learning_rate": 4.9555555555555565e-06,
|
| 691 |
+
"loss": 1.3023,
|
| 692 |
"step": 2775
|
| 693 |
},
|
| 694 |
{
|
| 695 |
+
"epoch": 0.48,
|
| 696 |
+
"learning_rate": 4.9000000000000005e-06,
|
| 697 |
+
"loss": 1.1222,
|
| 698 |
"step": 2800
|
| 699 |
},
|
| 700 |
{
|
| 701 |
+
"epoch": 0.49,
|
| 702 |
+
"learning_rate": 4.8444444444444446e-06,
|
| 703 |
+
"loss": 1.1982,
|
| 704 |
"step": 2825
|
| 705 |
},
|
| 706 |
{
|
| 707 |
+
"epoch": 0.49,
|
| 708 |
+
"learning_rate": 4.7888888888888894e-06,
|
| 709 |
+
"loss": 1.2535,
|
| 710 |
"step": 2850
|
| 711 |
},
|
| 712 |
{
|
| 713 |
+
"epoch": 0.49,
|
| 714 |
+
"learning_rate": 4.7333333333333335e-06,
|
| 715 |
+
"loss": 1.307,
|
| 716 |
"step": 2875
|
| 717 |
},
|
| 718 |
{
|
| 719 |
+
"epoch": 0.5,
|
| 720 |
+
"learning_rate": 4.677777777777778e-06,
|
| 721 |
+
"loss": 1.1476,
|
| 722 |
"step": 2900
|
| 723 |
},
|
| 724 |
{
|
| 725 |
+
"epoch": 0.5,
|
| 726 |
+
"learning_rate": 4.622222222222222e-06,
|
| 727 |
+
"loss": 1.3246,
|
| 728 |
"step": 2925
|
| 729 |
},
|
| 730 |
{
|
| 731 |
+
"epoch": 0.51,
|
| 732 |
+
"learning_rate": 4.566666666666667e-06,
|
| 733 |
+
"loss": 1.3723,
|
| 734 |
"step": 2950
|
| 735 |
},
|
| 736 |
{
|
| 737 |
+
"epoch": 0.51,
|
| 738 |
+
"learning_rate": 4.511111111111111e-06,
|
| 739 |
+
"loss": 1.192,
|
| 740 |
"step": 2975
|
| 741 |
},
|
| 742 |
{
|
| 743 |
+
"epoch": 0.52,
|
| 744 |
+
"learning_rate": 4.455555555555555e-06,
|
| 745 |
+
"loss": 1.2212,
|
| 746 |
"step": 3000
|
| 747 |
},
|
| 748 |
{
|
| 749 |
+
"epoch": 0.52,
|
| 750 |
+
"eval_loss": 1.2206178903579712,
|
| 751 |
+
"eval_runtime": 254.8252,
|
| 752 |
+
"eval_samples_per_second": 9.618,
|
| 753 |
+
"eval_steps_per_second": 2.406,
|
| 754 |
+
"eval_wer": 84.11905842682057,
|
| 755 |
"step": 3000
|
| 756 |
},
|
| 757 |
{
|
| 758 |
+
"epoch": 0.52,
|
| 759 |
+
"learning_rate": 4.4e-06,
|
| 760 |
+
"loss": 1.2387,
|
| 761 |
"step": 3025
|
| 762 |
},
|
| 763 |
{
|
| 764 |
+
"epoch": 0.52,
|
| 765 |
+
"learning_rate": 4.344444444444445e-06,
|
| 766 |
+
"loss": 1.1781,
|
| 767 |
"step": 3050
|
| 768 |
},
|
| 769 |
{
|
| 770 |
+
"epoch": 0.53,
|
| 771 |
+
"learning_rate": 4.288888888888889e-06,
|
| 772 |
+
"loss": 1.0837,
|
| 773 |
"step": 3075
|
| 774 |
},
|
| 775 |
{
|
| 776 |
+
"epoch": 0.53,
|
| 777 |
+
"learning_rate": 4.233333333333334e-06,
|
| 778 |
+
"loss": 1.2201,
|
| 779 |
"step": 3100
|
| 780 |
},
|
| 781 |
{
|
| 782 |
+
"epoch": 0.54,
|
| 783 |
+
"learning_rate": 4.177777777777778e-06,
|
| 784 |
+
"loss": 1.3045,
|
| 785 |
"step": 3125
|
| 786 |
},
|
| 787 |
{
|
| 788 |
+
"epoch": 0.54,
|
| 789 |
+
"learning_rate": 4.122222222222222e-06,
|
| 790 |
+
"loss": 1.2479,
|
| 791 |
"step": 3150
|
| 792 |
},
|
| 793 |
{
|
| 794 |
+
"epoch": 0.55,
|
| 795 |
+
"learning_rate": 4.066666666666667e-06,
|
| 796 |
+
"loss": 1.1863,
|
| 797 |
"step": 3175
|
| 798 |
},
|
| 799 |
{
|
| 800 |
+
"epoch": 0.55,
|
| 801 |
+
"learning_rate": 4.011111111111111e-06,
|
| 802 |
+
"loss": 1.3211,
|
| 803 |
"step": 3200
|
| 804 |
},
|
| 805 |
{
|
| 806 |
+
"epoch": 0.55,
|
| 807 |
+
"learning_rate": 3.955555555555556e-06,
|
| 808 |
+
"loss": 1.2037,
|
| 809 |
"step": 3225
|
| 810 |
},
|
| 811 |
{
|
| 812 |
+
"epoch": 0.56,
|
| 813 |
+
"learning_rate": 3.900000000000001e-06,
|
| 814 |
+
"loss": 1.153,
|
| 815 |
"step": 3250
|
| 816 |
},
|
| 817 |
{
|
| 818 |
+
"epoch": 0.56,
|
| 819 |
+
"learning_rate": 3.844444444444445e-06,
|
| 820 |
+
"loss": 1.252,
|
| 821 |
"step": 3275
|
| 822 |
},
|
| 823 |
{
|
| 824 |
+
"epoch": 0.57,
|
| 825 |
+
"learning_rate": 3.7888888888888893e-06,
|
| 826 |
+
"loss": 1.1224,
|
| 827 |
"step": 3300
|
| 828 |
},
|
| 829 |
{
|
| 830 |
+
"epoch": 0.57,
|
| 831 |
+
"learning_rate": 3.7333333333333337e-06,
|
| 832 |
+
"loss": 1.1686,
|
| 833 |
"step": 3325
|
| 834 |
},
|
| 835 |
{
|
| 836 |
+
"epoch": 0.58,
|
| 837 |
+
"learning_rate": 3.6777777777777778e-06,
|
| 838 |
+
"loss": 1.1249,
|
| 839 |
"step": 3350
|
| 840 |
},
|
| 841 |
{
|
| 842 |
+
"epoch": 0.58,
|
| 843 |
+
"learning_rate": 3.6222222222222226e-06,
|
| 844 |
+
"loss": 1.1661,
|
| 845 |
"step": 3375
|
| 846 |
},
|
| 847 |
{
|
| 848 |
+
"epoch": 0.58,
|
| 849 |
+
"learning_rate": 3.566666666666667e-06,
|
| 850 |
+
"loss": 1.1425,
|
| 851 |
"step": 3400
|
| 852 |
},
|
| 853 |
{
|
| 854 |
+
"epoch": 0.59,
|
| 855 |
+
"learning_rate": 3.511111111111111e-06,
|
| 856 |
+
"loss": 1.3145,
|
| 857 |
"step": 3425
|
| 858 |
},
|
| 859 |
{
|
| 860 |
+
"epoch": 0.59,
|
| 861 |
+
"learning_rate": 3.455555555555556e-06,
|
| 862 |
+
"loss": 1.212,
|
| 863 |
"step": 3450
|
| 864 |
},
|
| 865 |
{
|
| 866 |
+
"epoch": 0.6,
|
| 867 |
+
"learning_rate": 3.4000000000000005e-06,
|
| 868 |
+
"loss": 1.3946,
|
| 869 |
"step": 3475
|
| 870 |
},
|
| 871 |
{
|
| 872 |
+
"epoch": 0.6,
|
| 873 |
+
"learning_rate": 3.3444444444444445e-06,
|
| 874 |
+
"loss": 1.1685,
|
| 875 |
"step": 3500
|
| 876 |
},
|
| 877 |
{
|
| 878 |
+
"epoch": 0.61,
|
| 879 |
+
"learning_rate": 3.2888888888888894e-06,
|
| 880 |
+
"loss": 1.253,
|
| 881 |
"step": 3525
|
| 882 |
},
|
| 883 |
{
|
| 884 |
+
"epoch": 0.61,
|
| 885 |
+
"learning_rate": 3.2333333333333334e-06,
|
| 886 |
+
"loss": 1.0476,
|
| 887 |
"step": 3550
|
| 888 |
},
|
| 889 |
{
|
| 890 |
+
"epoch": 0.61,
|
| 891 |
+
"learning_rate": 3.177777777777778e-06,
|
| 892 |
+
"loss": 1.2093,
|
| 893 |
"step": 3575
|
| 894 |
},
|
| 895 |
{
|
| 896 |
+
"epoch": 0.62,
|
| 897 |
+
"learning_rate": 3.1222222222222228e-06,
|
| 898 |
+
"loss": 1.1977,
|
| 899 |
"step": 3600
|
| 900 |
},
|
| 901 |
{
|
| 902 |
+
"epoch": 0.62,
|
| 903 |
+
"learning_rate": 3.066666666666667e-06,
|
| 904 |
+
"loss": 1.2082,
|
| 905 |
"step": 3625
|
| 906 |
},
|
| 907 |
{
|
| 908 |
+
"epoch": 0.63,
|
| 909 |
+
"learning_rate": 3.0111111111111113e-06,
|
| 910 |
+
"loss": 1.1935,
|
| 911 |
"step": 3650
|
| 912 |
},
|
| 913 |
{
|
| 914 |
+
"epoch": 0.63,
|
| 915 |
+
"learning_rate": 2.955555555555556e-06,
|
| 916 |
+
"loss": 1.2793,
|
| 917 |
"step": 3675
|
| 918 |
},
|
| 919 |
{
|
| 920 |
+
"epoch": 0.64,
|
| 921 |
+
"learning_rate": 2.9e-06,
|
| 922 |
+
"loss": 1.2551,
|
| 923 |
"step": 3700
|
| 924 |
},
|
| 925 |
{
|
| 926 |
+
"epoch": 0.64,
|
| 927 |
+
"learning_rate": 2.8444444444444446e-06,
|
| 928 |
+
"loss": 1.0723,
|
| 929 |
"step": 3725
|
| 930 |
},
|
| 931 |
{
|
| 932 |
+
"epoch": 0.64,
|
| 933 |
+
"learning_rate": 2.788888888888889e-06,
|
| 934 |
+
"loss": 1.2802,
|
| 935 |
"step": 3750
|
| 936 |
},
|
| 937 |
{
|
| 938 |
+
"epoch": 0.65,
|
| 939 |
+
"learning_rate": 2.7333333333333336e-06,
|
| 940 |
+
"loss": 1.2181,
|
| 941 |
"step": 3775
|
| 942 |
},
|
| 943 |
{
|
| 944 |
+
"epoch": 0.65,
|
| 945 |
+
"learning_rate": 2.677777777777778e-06,
|
| 946 |
+
"loss": 1.1853,
|
| 947 |
"step": 3800
|
| 948 |
},
|
| 949 |
{
|
| 950 |
+
"epoch": 0.66,
|
| 951 |
+
"learning_rate": 2.6222222222222225e-06,
|
| 952 |
+
"loss": 1.1862,
|
| 953 |
"step": 3825
|
| 954 |
},
|
| 955 |
{
|
| 956 |
+
"epoch": 0.66,
|
| 957 |
+
"learning_rate": 2.566666666666667e-06,
|
| 958 |
+
"loss": 1.2022,
|
| 959 |
"step": 3850
|
| 960 |
},
|
| 961 |
{
|
| 962 |
+
"epoch": 0.67,
|
| 963 |
+
"learning_rate": 2.5111111111111114e-06,
|
| 964 |
+
"loss": 1.261,
|
| 965 |
"step": 3875
|
| 966 |
},
|
| 967 |
{
|
| 968 |
+
"epoch": 0.67,
|
| 969 |
+
"learning_rate": 2.455555555555556e-06,
|
| 970 |
+
"loss": 1.2105,
|
| 971 |
"step": 3900
|
| 972 |
},
|
| 973 |
{
|
| 974 |
+
"epoch": 0.67,
|
| 975 |
+
"learning_rate": 2.4000000000000003e-06,
|
| 976 |
+
"loss": 1.0427,
|
| 977 |
"step": 3925
|
| 978 |
},
|
| 979 |
{
|
| 980 |
+
"epoch": 0.68,
|
| 981 |
+
"learning_rate": 2.3444444444444448e-06,
|
| 982 |
+
"loss": 1.293,
|
| 983 |
"step": 3950
|
| 984 |
},
|
| 985 |
{
|
| 986 |
+
"epoch": 0.68,
|
| 987 |
+
"learning_rate": 2.2888888888888892e-06,
|
| 988 |
+
"loss": 1.1512,
|
| 989 |
"step": 3975
|
| 990 |
},
|
| 991 |
{
|
| 992 |
+
"epoch": 0.69,
|
| 993 |
+
"learning_rate": 2.2333333333333333e-06,
|
| 994 |
+
"loss": 1.1306,
|
| 995 |
"step": 4000
|
| 996 |
},
|
| 997 |
{
|
| 998 |
+
"epoch": 0.69,
|
| 999 |
+
"eval_loss": 1.1814459562301636,
|
| 1000 |
+
"eval_runtime": 194.4571,
|
| 1001 |
+
"eval_samples_per_second": 12.604,
|
| 1002 |
+
"eval_steps_per_second": 3.152,
|
| 1003 |
+
"eval_wer": 78.15316775825173,
|
| 1004 |
"step": 4000
|
| 1005 |
},
|
| 1006 |
{
|
| 1007 |
+
"epoch": 0.69,
|
| 1008 |
+
"learning_rate": 2.1777777777777777e-06,
|
| 1009 |
+
"loss": 1.2084,
|
| 1010 |
"step": 4025
|
| 1011 |
},
|
| 1012 |
{
|
| 1013 |
+
"epoch": 0.7,
|
| 1014 |
+
"learning_rate": 2.1222222222222226e-06,
|
| 1015 |
+
"loss": 1.2137,
|
| 1016 |
"step": 4050
|
| 1017 |
},
|
| 1018 |
{
|
| 1019 |
+
"epoch": 0.7,
|
| 1020 |
+
"learning_rate": 2.0666666666666666e-06,
|
| 1021 |
+
"loss": 1.195,
|
| 1022 |
"step": 4075
|
| 1023 |
},
|
| 1024 |
{
|
| 1025 |
+
"epoch": 0.7,
|
| 1026 |
+
"learning_rate": 2.011111111111111e-06,
|
| 1027 |
+
"loss": 1.0943,
|
| 1028 |
"step": 4100
|
| 1029 |
},
|
| 1030 |
{
|
| 1031 |
+
"epoch": 0.71,
|
| 1032 |
+
"learning_rate": 1.955555555555556e-06,
|
| 1033 |
+
"loss": 1.2158,
|
| 1034 |
"step": 4125
|
| 1035 |
},
|
| 1036 |
{
|
| 1037 |
+
"epoch": 0.71,
|
| 1038 |
+
"learning_rate": 1.9000000000000002e-06,
|
| 1039 |
+
"loss": 1.1431,
|
| 1040 |
"step": 4150
|
| 1041 |
},
|
| 1042 |
{
|
| 1043 |
+
"epoch": 0.72,
|
| 1044 |
+
"learning_rate": 1.8444444444444445e-06,
|
| 1045 |
+
"loss": 1.2699,
|
| 1046 |
"step": 4175
|
| 1047 |
},
|
| 1048 |
{
|
| 1049 |
+
"epoch": 0.72,
|
| 1050 |
+
"learning_rate": 1.788888888888889e-06,
|
| 1051 |
+
"loss": 1.1721,
|
| 1052 |
"step": 4200
|
| 1053 |
},
|
| 1054 |
{
|
| 1055 |
+
"epoch": 0.73,
|
| 1056 |
+
"learning_rate": 1.7333333333333336e-06,
|
| 1057 |
+
"loss": 1.2271,
|
| 1058 |
"step": 4225
|
| 1059 |
},
|
| 1060 |
{
|
| 1061 |
+
"epoch": 0.73,
|
| 1062 |
+
"learning_rate": 1.6777777777777779e-06,
|
| 1063 |
+
"loss": 1.1629,
|
| 1064 |
"step": 4250
|
| 1065 |
},
|
| 1066 |
{
|
| 1067 |
+
"epoch": 0.73,
|
| 1068 |
+
"learning_rate": 1.6222222222222223e-06,
|
| 1069 |
+
"loss": 1.3203,
|
| 1070 |
"step": 4275
|
| 1071 |
},
|
| 1072 |
{
|
| 1073 |
+
"epoch": 0.74,
|
| 1074 |
+
"learning_rate": 1.566666666666667e-06,
|
| 1075 |
+
"loss": 1.1082,
|
| 1076 |
"step": 4300
|
| 1077 |
},
|
| 1078 |
{
|
| 1079 |
+
"epoch": 0.74,
|
| 1080 |
+
"learning_rate": 1.5111111111111112e-06,
|
| 1081 |
+
"loss": 1.1323,
|
| 1082 |
"step": 4325
|
| 1083 |
},
|
| 1084 |
{
|
| 1085 |
+
"epoch": 0.75,
|
| 1086 |
+
"learning_rate": 1.4555555555555557e-06,
|
| 1087 |
+
"loss": 1.335,
|
| 1088 |
"step": 4350
|
| 1089 |
},
|
| 1090 |
{
|
| 1091 |
+
"epoch": 0.75,
|
| 1092 |
+
"learning_rate": 1.4000000000000001e-06,
|
| 1093 |
+
"loss": 1.0591,
|
| 1094 |
"step": 4375
|
| 1095 |
},
|
| 1096 |
{
|
| 1097 |
+
"epoch": 0.76,
|
| 1098 |
+
"learning_rate": 1.3444444444444446e-06,
|
| 1099 |
+
"loss": 1.3213,
|
| 1100 |
"step": 4400
|
| 1101 |
},
|
| 1102 |
{
|
| 1103 |
+
"epoch": 0.76,
|
| 1104 |
+
"learning_rate": 1.288888888888889e-06,
|
| 1105 |
+
"loss": 1.2386,
|
| 1106 |
"step": 4425
|
| 1107 |
},
|
| 1108 |
{
|
| 1109 |
+
"epoch": 0.76,
|
| 1110 |
+
"learning_rate": 1.2333333333333335e-06,
|
| 1111 |
+
"loss": 1.1389,
|
| 1112 |
"step": 4450
|
| 1113 |
},
|
| 1114 |
{
|
| 1115 |
+
"epoch": 0.77,
|
| 1116 |
+
"learning_rate": 1.1777777777777778e-06,
|
| 1117 |
+
"loss": 1.3361,
|
| 1118 |
"step": 4475
|
| 1119 |
},
|
| 1120 |
{
|
| 1121 |
+
"epoch": 0.77,
|
| 1122 |
+
"learning_rate": 1.1222222222222222e-06,
|
| 1123 |
+
"loss": 1.1915,
|
| 1124 |
"step": 4500
|
| 1125 |
},
|
| 1126 |
{
|
| 1127 |
+
"epoch": 0.78,
|
| 1128 |
+
"learning_rate": 1.066666666666667e-06,
|
| 1129 |
+
"loss": 1.067,
|
| 1130 |
"step": 4525
|
| 1131 |
},
|
| 1132 |
{
|
| 1133 |
+
"epoch": 0.78,
|
| 1134 |
+
"learning_rate": 1.0111111111111111e-06,
|
| 1135 |
+
"loss": 1.1499,
|
| 1136 |
"step": 4550
|
| 1137 |
},
|
| 1138 |
{
|
| 1139 |
+
"epoch": 0.79,
|
| 1140 |
+
"learning_rate": 9.555555555555556e-07,
|
| 1141 |
+
"loss": 1.2431,
|
| 1142 |
"step": 4575
|
| 1143 |
},
|
| 1144 |
{
|
| 1145 |
+
"epoch": 0.79,
|
| 1146 |
+
"learning_rate": 9.000000000000001e-07,
|
| 1147 |
+
"loss": 1.2343,
|
| 1148 |
"step": 4600
|
| 1149 |
},
|
| 1150 |
{
|
| 1151 |
+
"epoch": 0.79,
|
| 1152 |
+
"learning_rate": 8.444444444444445e-07,
|
| 1153 |
+
"loss": 1.152,
|
| 1154 |
"step": 4625
|
| 1155 |
},
|
| 1156 |
{
|
| 1157 |
+
"epoch": 0.8,
|
| 1158 |
+
"learning_rate": 7.888888888888889e-07,
|
| 1159 |
+
"loss": 1.2086,
|
| 1160 |
"step": 4650
|
| 1161 |
},
|
| 1162 |
{
|
| 1163 |
+
"epoch": 0.8,
|
| 1164 |
+
"learning_rate": 7.333333333333334e-07,
|
| 1165 |
+
"loss": 1.229,
|
| 1166 |
"step": 4675
|
| 1167 |
},
|
| 1168 |
{
|
| 1169 |
+
"epoch": 0.81,
|
| 1170 |
+
"learning_rate": 6.777777777777779e-07,
|
| 1171 |
+
"loss": 1.2757,
|
| 1172 |
"step": 4700
|
| 1173 |
},
|
| 1174 |
{
|
| 1175 |
+
"epoch": 0.81,
|
| 1176 |
+
"learning_rate": 6.222222222222223e-07,
|
| 1177 |
+
"loss": 1.1664,
|
| 1178 |
"step": 4725
|
| 1179 |
},
|
| 1180 |
{
|
| 1181 |
+
"epoch": 0.82,
|
| 1182 |
+
"learning_rate": 5.666666666666667e-07,
|
| 1183 |
+
"loss": 1.0384,
|
| 1184 |
"step": 4750
|
| 1185 |
},
|
| 1186 |
{
|
| 1187 |
+
"epoch": 0.82,
|
| 1188 |
+
"learning_rate": 5.111111111111112e-07,
|
| 1189 |
+
"loss": 1.2596,
|
| 1190 |
"step": 4775
|
| 1191 |
},
|
| 1192 |
{
|
| 1193 |
+
"epoch": 0.82,
|
| 1194 |
+
"learning_rate": 4.5555555555555563e-07,
|
| 1195 |
+
"loss": 1.1237,
|
| 1196 |
"step": 4800
|
| 1197 |
},
|
| 1198 |
{
|
| 1199 |
+
"epoch": 0.83,
|
| 1200 |
+
"learning_rate": 4.0000000000000003e-07,
|
| 1201 |
+
"loss": 1.0471,
|
| 1202 |
"step": 4825
|
| 1203 |
},
|
| 1204 |
{
|
| 1205 |
+
"epoch": 0.83,
|
| 1206 |
+
"learning_rate": 3.444444444444445e-07,
|
| 1207 |
+
"loss": 1.0879,
|
| 1208 |
"step": 4850
|
| 1209 |
},
|
| 1210 |
{
|
| 1211 |
+
"epoch": 0.84,
|
| 1212 |
+
"learning_rate": 2.888888888888889e-07,
|
| 1213 |
+
"loss": 1.1853,
|
| 1214 |
"step": 4875
|
| 1215 |
},
|
| 1216 |
{
|
| 1217 |
+
"epoch": 0.84,
|
| 1218 |
+
"learning_rate": 2.3333333333333336e-07,
|
| 1219 |
+
"loss": 1.1763,
|
| 1220 |
"step": 4900
|
| 1221 |
},
|
| 1222 |
{
|
| 1223 |
+
"epoch": 0.85,
|
| 1224 |
+
"learning_rate": 1.777777777777778e-07,
|
| 1225 |
+
"loss": 1.1031,
|
| 1226 |
"step": 4925
|
| 1227 |
},
|
| 1228 |
{
|
| 1229 |
+
"epoch": 0.85,
|
| 1230 |
+
"learning_rate": 1.2222222222222225e-07,
|
| 1231 |
+
"loss": 1.1508,
|
| 1232 |
"step": 4950
|
| 1233 |
},
|
| 1234 |
{
|
| 1235 |
+
"epoch": 0.85,
|
| 1236 |
+
"learning_rate": 6.666666666666668e-08,
|
| 1237 |
+
"loss": 1.0186,
|
| 1238 |
"step": 4975
|
| 1239 |
},
|
| 1240 |
{
|
| 1241 |
+
"epoch": 0.86,
|
| 1242 |
+
"learning_rate": 1.1111111111111112e-08,
|
| 1243 |
+
"loss": 1.1333,
|
| 1244 |
"step": 5000
|
| 1245 |
},
|
| 1246 |
{
|
| 1247 |
+
"epoch": 0.86,
|
| 1248 |
+
"eval_loss": 1.1651121377944946,
|
| 1249 |
+
"eval_runtime": 180.3205,
|
| 1250 |
+
"eval_samples_per_second": 13.592,
|
| 1251 |
+
"eval_steps_per_second": 3.4,
|
| 1252 |
+
"eval_wer": 76.37636988522145,
|
| 1253 |
"step": 5000
|
| 1254 |
},
|
| 1255 |
{
|
| 1256 |
+
"epoch": 0.86,
|
| 1257 |
"step": 5000,
|
| 1258 |
+
"total_flos": 9.847554048e+17,
|
| 1259 |
+
"train_loss": 1.343574110031128,
|
| 1260 |
+
"train_runtime": 2849.8517,
|
| 1261 |
+
"train_samples_per_second": 14.036,
|
| 1262 |
+
"train_steps_per_second": 1.754
|
| 1263 |
}
|
| 1264 |
],
|
| 1265 |
"max_steps": 5000,
|
| 1266 |
+
"num_train_epochs": 1,
|
| 1267 |
+
"total_flos": 9.847554048e+17,
|
| 1268 |
"trial_name": null,
|
| 1269 |
"trial_params": null
|
| 1270 |
}
|