Upload folder using huggingface_hub
Browse files- config.json +1 -1
- pytorch_model-00001-of-00003.bin +1 -1
- pytorch_model-00002-of-00003.bin +1 -1
- pytorch_model-00003-of-00003.bin +1 -1
- trainer_state.json +1893 -357
- training_args.bin +1 -1
config.json
CHANGED
|
@@ -1,5 +1,5 @@
|
|
| 1 |
{
|
| 2 |
-
"_name_or_path": "/workspace/airoboros-
|
| 3 |
"architectures": [
|
| 4 |
"LlamaForCausalLM"
|
| 5 |
],
|
|
|
|
| 1 |
{
|
| 2 |
+
"_name_or_path": "/workspace/airoboros-7b/checkpoint-225",
|
| 3 |
"architectures": [
|
| 4 |
"LlamaForCausalLM"
|
| 5 |
],
|
pytorch_model-00001-of-00003.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 9877989586
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cb48d826dc36d3b01ebf4c8a5b19a2cec8625d72187e7e5d76d300b2fe9399ed
|
| 3 |
size 9877989586
|
pytorch_model-00002-of-00003.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 9894801014
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f72e5f3aede4e4984b3a0d05ac125f189d86f1ccedb0bcf62ced3f9ef3d5ddbc
|
| 3 |
size 9894801014
|
pytorch_model-00003-of-00003.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 7180990649
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d789cb7190b17c2cc43fcf29af45adb3eec1ba248353761e29dacfd512a8e06d
|
| 3 |
size 7180990649
|
trainer_state.json
CHANGED
|
@@ -1,799 +1,2335 @@
|
|
| 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":
|
| 13 |
-
"loss":
|
| 14 |
"step": 1
|
| 15 |
},
|
| 16 |
{
|
| 17 |
"epoch": 0.02,
|
| 18 |
-
"learning_rate":
|
| 19 |
-
"loss":
|
| 20 |
"step": 2
|
| 21 |
},
|
| 22 |
{
|
| 23 |
"epoch": 0.02,
|
| 24 |
-
"learning_rate":
|
| 25 |
-
"loss": 0.
|
| 26 |
"step": 3
|
| 27 |
},
|
| 28 |
{
|
| 29 |
"epoch": 0.03,
|
| 30 |
-
"learning_rate":
|
| 31 |
-
"loss": 0.
|
| 32 |
"step": 4
|
| 33 |
},
|
| 34 |
{
|
| 35 |
"epoch": 0.04,
|
| 36 |
-
"learning_rate":
|
| 37 |
-
"loss": 0.
|
| 38 |
"step": 5
|
| 39 |
},
|
| 40 |
{
|
| 41 |
"epoch": 0.05,
|
| 42 |
-
"learning_rate":
|
| 43 |
-
"loss": 0.
|
| 44 |
"step": 6
|
| 45 |
},
|
| 46 |
{
|
| 47 |
"epoch": 0.05,
|
| 48 |
-
"learning_rate":
|
| 49 |
-
"loss": 0.
|
| 50 |
"step": 7
|
| 51 |
},
|
| 52 |
{
|
| 53 |
"epoch": 0.06,
|
| 54 |
-
"learning_rate":
|
| 55 |
-
"loss": 0.
|
| 56 |
"step": 8
|
| 57 |
},
|
| 58 |
{
|
| 59 |
"epoch": 0.07,
|
| 60 |
-
"learning_rate": 1.
|
| 61 |
-
"loss": 0.
|
| 62 |
"step": 9
|
| 63 |
},
|
| 64 |
{
|
| 65 |
"epoch": 0.08,
|
| 66 |
-
"learning_rate": 1.
|
| 67 |
-
"loss": 0.
|
| 68 |
"step": 10
|
| 69 |
},
|
| 70 |
{
|
| 71 |
-
"epoch": 0.
|
| 72 |
-
"learning_rate": 1.
|
| 73 |
-
"loss": 0.
|
| 74 |
"step": 11
|
| 75 |
},
|
| 76 |
{
|
| 77 |
"epoch": 0.09,
|
| 78 |
-
"learning_rate": 1.
|
| 79 |
-
"loss": 0.
|
| 80 |
"step": 12
|
| 81 |
},
|
| 82 |
{
|
| 83 |
"epoch": 0.1,
|
| 84 |
-
"learning_rate": 1.
|
| 85 |
-
"loss": 0.
|
| 86 |
"step": 13
|
| 87 |
},
|
| 88 |
{
|
| 89 |
"epoch": 0.11,
|
| 90 |
-
"learning_rate": 1.
|
| 91 |
-
"loss": 0.
|
| 92 |
"step": 14
|
| 93 |
},
|
| 94 |
{
|
| 95 |
"epoch": 0.12,
|
| 96 |
-
"learning_rate": 1.
|
| 97 |
-
"loss": 0.
|
| 98 |
"step": 15
|
| 99 |
},
|
| 100 |
{
|
| 101 |
-
"epoch": 0.
|
| 102 |
-
"learning_rate":
|
| 103 |
-
"loss": 0.
|
| 104 |
"step": 16
|
| 105 |
},
|
| 106 |
{
|
| 107 |
"epoch": 0.13,
|
| 108 |
-
"learning_rate": 1.
|
| 109 |
-
"loss": 0.
|
| 110 |
"step": 17
|
| 111 |
},
|
| 112 |
{
|
| 113 |
"epoch": 0.14,
|
| 114 |
-
"learning_rate": 1.
|
| 115 |
-
"loss": 0.
|
| 116 |
"step": 18
|
| 117 |
},
|
| 118 |
{
|
| 119 |
"epoch": 0.15,
|
| 120 |
-
"learning_rate": 1.
|
| 121 |
-
"loss": 0.
|
| 122 |
"step": 19
|
| 123 |
},
|
| 124 |
{
|
| 125 |
-
"epoch": 0.
|
| 126 |
-
"learning_rate": 1.
|
| 127 |
-
"loss": 0.
|
| 128 |
"step": 20
|
| 129 |
},
|
| 130 |
{
|
| 131 |
"epoch": 0.16,
|
| 132 |
-
"learning_rate": 1.
|
| 133 |
-
"loss": 0.
|
| 134 |
"step": 21
|
| 135 |
},
|
| 136 |
{
|
| 137 |
"epoch": 0.17,
|
| 138 |
-
"learning_rate": 1.
|
| 139 |
-
"loss": 0.
|
| 140 |
"step": 22
|
| 141 |
},
|
| 142 |
{
|
| 143 |
"epoch": 0.18,
|
| 144 |
-
"learning_rate": 1.
|
| 145 |
-
"loss": 0.
|
| 146 |
"step": 23
|
| 147 |
},
|
| 148 |
{
|
| 149 |
"epoch": 0.19,
|
| 150 |
-
"learning_rate": 1.
|
| 151 |
-
"loss": 0.
|
| 152 |
"step": 24
|
| 153 |
},
|
| 154 |
{
|
| 155 |
-
"epoch": 0.
|
| 156 |
-
"learning_rate": 1.
|
| 157 |
-
"loss": 0.
|
| 158 |
"step": 25
|
| 159 |
},
|
| 160 |
{
|
| 161 |
"epoch": 0.2,
|
| 162 |
-
"learning_rate": 1.
|
| 163 |
-
"loss": 0.
|
| 164 |
"step": 26
|
| 165 |
},
|
| 166 |
{
|
| 167 |
"epoch": 0.21,
|
| 168 |
-
"learning_rate": 1.
|
| 169 |
-
"loss": 0.
|
| 170 |
"step": 27
|
| 171 |
},
|
| 172 |
{
|
| 173 |
"epoch": 0.22,
|
| 174 |
-
"learning_rate": 1.
|
| 175 |
-
"loss": 0.
|
| 176 |
"step": 28
|
| 177 |
},
|
| 178 |
{
|
| 179 |
-
"epoch": 0.
|
| 180 |
-
"learning_rate": 1.
|
| 181 |
-
"loss": 0.
|
| 182 |
"step": 29
|
| 183 |
},
|
| 184 |
{
|
| 185 |
-
"epoch": 0.
|
| 186 |
-
"learning_rate": 1.
|
| 187 |
-
"loss": 0.
|
| 188 |
"step": 30
|
| 189 |
},
|
| 190 |
{
|
| 191 |
"epoch": 0.24,
|
| 192 |
-
"learning_rate": 1.
|
| 193 |
-
"loss": 0.
|
| 194 |
"step": 31
|
| 195 |
},
|
| 196 |
{
|
| 197 |
"epoch": 0.25,
|
| 198 |
-
"learning_rate": 1.
|
| 199 |
-
"loss": 0.
|
| 200 |
"step": 32
|
| 201 |
},
|
| 202 |
{
|
| 203 |
-
"epoch": 0.
|
| 204 |
-
"learning_rate": 1.
|
| 205 |
-
"loss": 0.
|
| 206 |
"step": 33
|
| 207 |
},
|
| 208 |
{
|
| 209 |
-
"epoch": 0.
|
| 210 |
-
"learning_rate": 1.
|
| 211 |
-
"loss": 0.
|
| 212 |
"step": 34
|
| 213 |
},
|
| 214 |
{
|
| 215 |
"epoch": 0.27,
|
| 216 |
-
"learning_rate": 1.
|
| 217 |
-
"loss": 0.
|
| 218 |
"step": 35
|
| 219 |
},
|
| 220 |
{
|
| 221 |
"epoch": 0.28,
|
| 222 |
-
"learning_rate": 1.
|
| 223 |
-
"loss": 0.
|
| 224 |
"step": 36
|
| 225 |
},
|
| 226 |
{
|
| 227 |
"epoch": 0.29,
|
| 228 |
-
"learning_rate": 1.
|
| 229 |
-
"loss": 0.
|
| 230 |
"step": 37
|
| 231 |
},
|
| 232 |
{
|
| 233 |
-
"epoch": 0.
|
| 234 |
-
"learning_rate": 1.
|
| 235 |
-
"loss": 0.
|
| 236 |
"step": 38
|
| 237 |
},
|
| 238 |
{
|
| 239 |
-
"epoch": 0.
|
| 240 |
-
"learning_rate": 1.
|
| 241 |
-
"loss": 0.
|
| 242 |
"step": 39
|
| 243 |
},
|
| 244 |
{
|
| 245 |
"epoch": 0.31,
|
| 246 |
-
"learning_rate": 1.
|
| 247 |
-
"loss": 0.
|
| 248 |
"step": 40
|
| 249 |
},
|
| 250 |
{
|
| 251 |
"epoch": 0.32,
|
| 252 |
-
"learning_rate": 1.
|
| 253 |
-
"loss": 0.
|
| 254 |
"step": 41
|
| 255 |
},
|
| 256 |
{
|
| 257 |
-
"epoch": 0.
|
| 258 |
-
"learning_rate": 1.
|
| 259 |
-
"loss": 0.
|
| 260 |
"step": 42
|
| 261 |
},
|
| 262 |
{
|
| 263 |
-
"epoch": 0.
|
| 264 |
-
"learning_rate": 1.
|
| 265 |
-
"loss": 0.
|
| 266 |
"step": 43
|
| 267 |
},
|
| 268 |
{
|
| 269 |
-
"epoch": 0.
|
| 270 |
-
"learning_rate": 1.
|
| 271 |
-
"loss": 0.
|
| 272 |
"step": 44
|
| 273 |
},
|
| 274 |
{
|
| 275 |
"epoch": 0.35,
|
| 276 |
-
"learning_rate": 1.
|
| 277 |
-
"loss": 0.
|
| 278 |
"step": 45
|
| 279 |
},
|
| 280 |
{
|
| 281 |
"epoch": 0.36,
|
| 282 |
-
"learning_rate": 1.
|
| 283 |
-
"loss": 0.
|
| 284 |
"step": 46
|
| 285 |
},
|
| 286 |
{
|
| 287 |
-
"epoch": 0.
|
| 288 |
-
"learning_rate": 1.
|
| 289 |
-
"loss": 0.
|
| 290 |
"step": 47
|
| 291 |
},
|
| 292 |
{
|
| 293 |
-
"epoch": 0.
|
| 294 |
-
"learning_rate": 1.
|
| 295 |
-
"loss": 0.
|
| 296 |
"step": 48
|
| 297 |
},
|
| 298 |
{
|
| 299 |
"epoch": 0.38,
|
| 300 |
-
"learning_rate": 1.
|
| 301 |
-
"loss": 0.
|
| 302 |
"step": 49
|
| 303 |
},
|
| 304 |
{
|
| 305 |
"epoch": 0.39,
|
| 306 |
-
"learning_rate": 1.
|
| 307 |
-
"loss": 0.
|
| 308 |
"step": 50
|
| 309 |
},
|
| 310 |
{
|
| 311 |
-
"epoch": 0.
|
| 312 |
-
"learning_rate": 1.
|
| 313 |
-
"loss": 0.
|
| 314 |
"step": 51
|
| 315 |
},
|
| 316 |
{
|
| 317 |
-
"epoch": 0.
|
| 318 |
-
"learning_rate": 1.
|
| 319 |
-
"loss": 0.
|
| 320 |
"step": 52
|
| 321 |
},
|
| 322 |
{
|
| 323 |
-
"epoch": 0.
|
| 324 |
-
"learning_rate": 1.
|
| 325 |
-
"loss": 0.
|
| 326 |
"step": 53
|
| 327 |
},
|
| 328 |
{
|
| 329 |
"epoch": 0.42,
|
| 330 |
-
"learning_rate": 1.
|
| 331 |
-
"loss": 0.
|
| 332 |
"step": 54
|
| 333 |
},
|
| 334 |
{
|
| 335 |
-
"epoch": 0.
|
| 336 |
-
"learning_rate": 1.
|
| 337 |
-
"loss": 0.
|
| 338 |
"step": 55
|
| 339 |
},
|
| 340 |
{
|
| 341 |
-
"epoch": 0.
|
| 342 |
-
"learning_rate": 1.
|
| 343 |
-
"loss": 0.
|
| 344 |
"step": 56
|
| 345 |
},
|
| 346 |
{
|
| 347 |
-
"epoch": 0.
|
| 348 |
-
"learning_rate": 1.
|
| 349 |
-
"loss": 0.
|
| 350 |
"step": 57
|
| 351 |
},
|
| 352 |
{
|
| 353 |
"epoch": 0.45,
|
| 354 |
-
"learning_rate": 1.
|
| 355 |
-
"loss": 0.
|
| 356 |
"step": 58
|
| 357 |
},
|
| 358 |
{
|
| 359 |
"epoch": 0.46,
|
| 360 |
-
"learning_rate": 1.
|
| 361 |
-
"loss": 0.
|
| 362 |
"step": 59
|
| 363 |
},
|
| 364 |
{
|
| 365 |
-
"epoch": 0.
|
| 366 |
-
"learning_rate": 1.
|
| 367 |
-
"loss": 0.
|
| 368 |
"step": 60
|
| 369 |
},
|
| 370 |
{
|
| 371 |
-
"epoch": 0.
|
| 372 |
-
"learning_rate": 1.
|
| 373 |
-
"loss": 0.
|
| 374 |
"step": 61
|
| 375 |
},
|
| 376 |
{
|
| 377 |
-
"epoch": 0.
|
| 378 |
-
"learning_rate": 1.
|
| 379 |
-
"loss": 0.
|
| 380 |
"step": 62
|
| 381 |
},
|
| 382 |
{
|
| 383 |
"epoch": 0.49,
|
| 384 |
-
"learning_rate": 1.
|
| 385 |
-
"loss": 0.
|
| 386 |
"step": 63
|
| 387 |
},
|
| 388 |
{
|
| 389 |
-
"epoch": 0.
|
| 390 |
-
"learning_rate": 1.
|
| 391 |
-
"loss": 0.
|
| 392 |
"step": 64
|
| 393 |
},
|
| 394 |
{
|
| 395 |
-
"epoch": 0.
|
| 396 |
-
"learning_rate": 1.
|
| 397 |
-
"loss": 0.
|
| 398 |
"step": 65
|
| 399 |
},
|
| 400 |
{
|
| 401 |
-
"epoch": 0.
|
| 402 |
-
"learning_rate": 1.
|
| 403 |
-
"loss": 0.
|
| 404 |
"step": 66
|
| 405 |
},
|
| 406 |
{
|
| 407 |
-
"epoch": 0.
|
| 408 |
-
"learning_rate": 1.
|
| 409 |
-
"loss": 0.
|
| 410 |
"step": 67
|
| 411 |
},
|
| 412 |
{
|
| 413 |
"epoch": 0.53,
|
| 414 |
-
"learning_rate":
|
| 415 |
-
"loss": 0.
|
| 416 |
"step": 68
|
| 417 |
},
|
| 418 |
{
|
| 419 |
-
"epoch": 0.
|
| 420 |
-
"learning_rate":
|
| 421 |
-
"loss": 0.
|
| 422 |
"step": 69
|
| 423 |
},
|
| 424 |
{
|
| 425 |
-
"epoch": 0.
|
| 426 |
-
"learning_rate":
|
| 427 |
-
"loss": 0.
|
| 428 |
"step": 70
|
| 429 |
},
|
| 430 |
{
|
| 431 |
-
"epoch": 0.
|
| 432 |
-
"learning_rate":
|
| 433 |
-
"loss": 0.
|
| 434 |
"step": 71
|
| 435 |
},
|
| 436 |
{
|
| 437 |
"epoch": 0.56,
|
| 438 |
-
"learning_rate":
|
| 439 |
-
"loss": 0.
|
| 440 |
"step": 72
|
| 441 |
},
|
| 442 |
{
|
| 443 |
-
"epoch": 0.
|
| 444 |
-
"learning_rate":
|
| 445 |
-
"loss": 0.
|
| 446 |
"step": 73
|
| 447 |
},
|
| 448 |
{
|
| 449 |
-
"epoch": 0.
|
| 450 |
-
"learning_rate":
|
| 451 |
-
"loss": 0.
|
| 452 |
"step": 74
|
| 453 |
},
|
| 454 |
{
|
| 455 |
-
"epoch": 0.
|
| 456 |
-
"learning_rate":
|
| 457 |
-
"loss": 0.
|
| 458 |
"step": 75
|
| 459 |
},
|
| 460 |
{
|
| 461 |
"epoch": 0.59,
|
| 462 |
-
"
|
| 463 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 464 |
"step": 76
|
| 465 |
},
|
| 466 |
{
|
| 467 |
-
"epoch": 0.
|
| 468 |
-
"learning_rate":
|
| 469 |
-
"loss": 0.
|
| 470 |
"step": 77
|
| 471 |
},
|
| 472 |
{
|
| 473 |
-
"epoch": 0.
|
| 474 |
-
"learning_rate":
|
| 475 |
-
"loss": 0.
|
| 476 |
"step": 78
|
| 477 |
},
|
| 478 |
{
|
| 479 |
-
"epoch": 0.
|
| 480 |
-
"learning_rate":
|
| 481 |
-
"loss": 0.
|
| 482 |
"step": 79
|
| 483 |
},
|
| 484 |
{
|
| 485 |
-
"epoch": 0.
|
| 486 |
-
"learning_rate":
|
| 487 |
-
"loss": 0.
|
| 488 |
"step": 80
|
| 489 |
},
|
| 490 |
{
|
| 491 |
-
"epoch": 0.
|
| 492 |
-
"learning_rate":
|
| 493 |
-
"loss": 0.
|
| 494 |
"step": 81
|
| 495 |
},
|
| 496 |
{
|
| 497 |
-
"epoch": 0.
|
| 498 |
-
"learning_rate":
|
| 499 |
-
"loss": 0.
|
| 500 |
"step": 82
|
| 501 |
},
|
| 502 |
{
|
| 503 |
-
"epoch": 0.
|
| 504 |
-
"learning_rate":
|
| 505 |
-
"loss": 0.
|
| 506 |
"step": 83
|
| 507 |
},
|
| 508 |
{
|
| 509 |
-
"epoch": 0.
|
| 510 |
-
"learning_rate":
|
| 511 |
-
"loss": 0.
|
| 512 |
"step": 84
|
| 513 |
},
|
| 514 |
{
|
| 515 |
-
"epoch": 0.
|
| 516 |
-
"learning_rate":
|
| 517 |
-
"loss": 0.
|
| 518 |
"step": 85
|
| 519 |
},
|
| 520 |
{
|
| 521 |
-
"epoch": 0.
|
| 522 |
-
"learning_rate":
|
| 523 |
-
"loss": 0.
|
| 524 |
"step": 86
|
| 525 |
},
|
| 526 |
{
|
| 527 |
-
"epoch": 0.
|
| 528 |
-
"learning_rate":
|
| 529 |
-
"loss": 0.
|
| 530 |
"step": 87
|
| 531 |
},
|
| 532 |
{
|
| 533 |
-
"epoch": 0.
|
| 534 |
-
"learning_rate":
|
| 535 |
-
"loss": 0.
|
| 536 |
"step": 88
|
| 537 |
},
|
| 538 |
{
|
| 539 |
-
"epoch": 0.
|
| 540 |
-
"learning_rate":
|
| 541 |
-
"loss": 0.
|
| 542 |
"step": 89
|
| 543 |
},
|
| 544 |
{
|
| 545 |
-
"epoch": 0.
|
| 546 |
-
"learning_rate":
|
| 547 |
-
"loss": 0.
|
| 548 |
"step": 90
|
| 549 |
},
|
| 550 |
{
|
| 551 |
-
"epoch": 0.
|
| 552 |
-
"learning_rate":
|
| 553 |
-
"loss": 0.
|
| 554 |
"step": 91
|
| 555 |
},
|
| 556 |
{
|
| 557 |
-
"epoch": 0.
|
| 558 |
-
"learning_rate":
|
| 559 |
-
"loss": 0.
|
| 560 |
"step": 92
|
| 561 |
},
|
| 562 |
{
|
| 563 |
-
"epoch": 0.
|
| 564 |
-
"learning_rate":
|
| 565 |
-
"loss": 0.
|
| 566 |
"step": 93
|
| 567 |
},
|
| 568 |
{
|
| 569 |
-
"epoch": 0.
|
| 570 |
-
"learning_rate":
|
| 571 |
-
"loss": 0.
|
| 572 |
"step": 94
|
| 573 |
},
|
| 574 |
{
|
| 575 |
-
"epoch": 0.
|
| 576 |
-
"learning_rate":
|
| 577 |
-
"loss": 0.
|
| 578 |
"step": 95
|
| 579 |
},
|
| 580 |
{
|
| 581 |
-
"epoch": 0.
|
| 582 |
-
"learning_rate":
|
| 583 |
-
"loss": 0.
|
| 584 |
"step": 96
|
| 585 |
},
|
| 586 |
{
|
| 587 |
-
"epoch": 0.
|
| 588 |
-
"learning_rate":
|
| 589 |
-
"loss": 0.
|
| 590 |
"step": 97
|
| 591 |
},
|
| 592 |
{
|
| 593 |
-
"epoch": 0.
|
| 594 |
-
"learning_rate":
|
| 595 |
-
"loss": 0.
|
| 596 |
"step": 98
|
| 597 |
},
|
| 598 |
{
|
| 599 |
-
"epoch": 0.
|
| 600 |
-
"learning_rate":
|
| 601 |
-
"loss": 0.
|
| 602 |
"step": 99
|
| 603 |
},
|
| 604 |
{
|
| 605 |
-
"epoch": 0.
|
| 606 |
-
"learning_rate":
|
| 607 |
-
"loss": 0.
|
| 608 |
"step": 100
|
| 609 |
},
|
| 610 |
{
|
| 611 |
-
"epoch": 0.
|
| 612 |
-
"learning_rate":
|
| 613 |
-
"loss": 0.
|
| 614 |
"step": 101
|
| 615 |
},
|
| 616 |
{
|
| 617 |
-
"epoch": 0.
|
| 618 |
-
"learning_rate":
|
| 619 |
-
"loss": 0.
|
| 620 |
"step": 102
|
| 621 |
},
|
| 622 |
{
|
| 623 |
-
"epoch": 0.
|
| 624 |
-
"learning_rate":
|
| 625 |
-
"loss": 0.
|
| 626 |
"step": 103
|
| 627 |
},
|
| 628 |
{
|
| 629 |
-
"epoch": 0.
|
| 630 |
-
"learning_rate": 1.
|
| 631 |
-
"loss": 0.
|
| 632 |
"step": 104
|
| 633 |
},
|
| 634 |
{
|
| 635 |
-
"epoch": 0.
|
| 636 |
-
"learning_rate": 1.
|
| 637 |
-
"loss": 0.
|
| 638 |
"step": 105
|
| 639 |
},
|
| 640 |
{
|
| 641 |
-
"epoch": 0.
|
| 642 |
-
"learning_rate": 1.
|
| 643 |
-
"loss": 0.
|
| 644 |
"step": 106
|
| 645 |
},
|
| 646 |
{
|
| 647 |
-
"epoch": 0.
|
| 648 |
-
"learning_rate": 1.
|
| 649 |
-
"loss": 0.
|
| 650 |
"step": 107
|
| 651 |
},
|
| 652 |
{
|
| 653 |
-
"epoch": 0.
|
| 654 |
-
"learning_rate": 1.
|
| 655 |
-
"loss": 0.
|
| 656 |
"step": 108
|
| 657 |
},
|
| 658 |
{
|
| 659 |
-
"epoch": 0.
|
| 660 |
-
"learning_rate": 1.
|
| 661 |
-
"loss": 0.
|
| 662 |
"step": 109
|
| 663 |
},
|
| 664 |
{
|
| 665 |
-
"epoch": 0.
|
| 666 |
-
"learning_rate": 1.
|
| 667 |
-
"loss": 0.
|
| 668 |
"step": 110
|
| 669 |
},
|
| 670 |
{
|
| 671 |
-
"epoch": 0.
|
| 672 |
-
"learning_rate": 1.
|
| 673 |
-
"loss": 0.
|
| 674 |
"step": 111
|
| 675 |
},
|
| 676 |
{
|
| 677 |
-
"epoch": 0.
|
| 678 |
-
"learning_rate":
|
| 679 |
-
"loss": 0.
|
| 680 |
"step": 112
|
| 681 |
},
|
| 682 |
{
|
| 683 |
-
"epoch": 0.
|
| 684 |
-
"learning_rate":
|
| 685 |
-
"loss": 0.
|
| 686 |
"step": 113
|
| 687 |
},
|
| 688 |
{
|
| 689 |
-
"epoch": 0.
|
| 690 |
-
"learning_rate":
|
| 691 |
-
"loss": 0.
|
| 692 |
"step": 114
|
| 693 |
},
|
| 694 |
{
|
| 695 |
-
"epoch": 0.
|
| 696 |
-
"learning_rate":
|
| 697 |
-
"loss": 0.
|
| 698 |
"step": 115
|
| 699 |
},
|
| 700 |
{
|
| 701 |
-
"epoch": 0.
|
| 702 |
-
"learning_rate":
|
| 703 |
-
"loss": 0.
|
| 704 |
"step": 116
|
| 705 |
},
|
| 706 |
{
|
| 707 |
-
"epoch": 0.
|
| 708 |
-
"learning_rate":
|
| 709 |
-
"loss": 0.
|
| 710 |
"step": 117
|
| 711 |
},
|
| 712 |
{
|
| 713 |
-
"epoch": 0.
|
| 714 |
-
"learning_rate":
|
| 715 |
-
"loss": 0.
|
| 716 |
"step": 118
|
| 717 |
},
|
| 718 |
{
|
| 719 |
-
"epoch": 0.
|
| 720 |
-
"learning_rate":
|
| 721 |
-
"loss": 0.
|
| 722 |
"step": 119
|
| 723 |
},
|
| 724 |
{
|
| 725 |
-
"epoch": 0.
|
| 726 |
-
"learning_rate":
|
| 727 |
-
"loss": 0.
|
| 728 |
"step": 120
|
| 729 |
},
|
| 730 |
{
|
| 731 |
-
"epoch": 0.
|
| 732 |
-
"learning_rate":
|
| 733 |
-
"loss": 0.
|
| 734 |
"step": 121
|
| 735 |
},
|
| 736 |
{
|
| 737 |
-
"epoch": 0.
|
| 738 |
-
"learning_rate": 1.
|
| 739 |
-
"loss": 0.
|
| 740 |
"step": 122
|
| 741 |
},
|
| 742 |
{
|
| 743 |
-
"epoch": 0.
|
| 744 |
-
"learning_rate": 1.
|
| 745 |
-
"loss": 0.
|
| 746 |
"step": 123
|
| 747 |
},
|
| 748 |
{
|
| 749 |
-
"epoch": 0.
|
| 750 |
-
"learning_rate":
|
| 751 |
-
"loss": 0.
|
| 752 |
"step": 124
|
| 753 |
},
|
| 754 |
{
|
| 755 |
-
"epoch": 0.
|
| 756 |
-
"learning_rate":
|
| 757 |
-
"loss": 0.
|
| 758 |
"step": 125
|
| 759 |
},
|
| 760 |
{
|
| 761 |
-
"epoch": 0.
|
| 762 |
-
"learning_rate":
|
| 763 |
-
"loss": 0.
|
| 764 |
"step": 126
|
| 765 |
},
|
| 766 |
{
|
| 767 |
-
"epoch": 0
|
| 768 |
-
"learning_rate": 1.
|
| 769 |
-
"loss": 0.
|
| 770 |
"step": 127
|
| 771 |
},
|
| 772 |
{
|
| 773 |
-
"epoch": 0
|
| 774 |
-
"learning_rate":
|
| 775 |
-
"loss": 0.
|
| 776 |
"step": 128
|
| 777 |
},
|
| 778 |
{
|
| 779 |
-
"epoch": 1.
|
| 780 |
-
"learning_rate":
|
| 781 |
-
"loss": 0.
|
| 782 |
"step": 129
|
| 783 |
},
|
| 784 |
{
|
| 785 |
-
"epoch": 1.
|
| 786 |
-
"
|
| 787 |
-
"
|
| 788 |
-
"
|
| 789 |
-
|
| 790 |
-
|
| 791 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 792 |
}
|
| 793 |
],
|
| 794 |
-
"max_steps":
|
| 795 |
-
"num_train_epochs":
|
| 796 |
-
"total_flos":
|
| 797 |
"trial_name": null,
|
| 798 |
"trial_params": null
|
| 799 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 2.992156862745098,
|
| 5 |
+
"global_step": 381,
|
| 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": 1.25e-06,
|
| 13 |
+
"loss": 1.031,
|
| 14 |
"step": 1
|
| 15 |
},
|
| 16 |
{
|
| 17 |
"epoch": 0.02,
|
| 18 |
+
"learning_rate": 2.5e-06,
|
| 19 |
+
"loss": 1.0257,
|
| 20 |
"step": 2
|
| 21 |
},
|
| 22 |
{
|
| 23 |
"epoch": 0.02,
|
| 24 |
+
"learning_rate": 3.7500000000000005e-06,
|
| 25 |
+
"loss": 0.8911,
|
| 26 |
"step": 3
|
| 27 |
},
|
| 28 |
{
|
| 29 |
"epoch": 0.03,
|
| 30 |
+
"learning_rate": 5e-06,
|
| 31 |
+
"loss": 0.8772,
|
| 32 |
"step": 4
|
| 33 |
},
|
| 34 |
{
|
| 35 |
"epoch": 0.04,
|
| 36 |
+
"learning_rate": 6.25e-06,
|
| 37 |
+
"loss": 0.8499,
|
| 38 |
"step": 5
|
| 39 |
},
|
| 40 |
{
|
| 41 |
"epoch": 0.05,
|
| 42 |
+
"learning_rate": 7.500000000000001e-06,
|
| 43 |
+
"loss": 0.8031,
|
| 44 |
"step": 6
|
| 45 |
},
|
| 46 |
{
|
| 47 |
"epoch": 0.05,
|
| 48 |
+
"learning_rate": 8.750000000000001e-06,
|
| 49 |
+
"loss": 0.7872,
|
| 50 |
"step": 7
|
| 51 |
},
|
| 52 |
{
|
| 53 |
"epoch": 0.06,
|
| 54 |
+
"learning_rate": 1e-05,
|
| 55 |
+
"loss": 0.7626,
|
| 56 |
"step": 8
|
| 57 |
},
|
| 58 |
{
|
| 59 |
"epoch": 0.07,
|
| 60 |
+
"learning_rate": 1.125e-05,
|
| 61 |
+
"loss": 0.7332,
|
| 62 |
"step": 9
|
| 63 |
},
|
| 64 |
{
|
| 65 |
"epoch": 0.08,
|
| 66 |
+
"learning_rate": 1.25e-05,
|
| 67 |
+
"loss": 0.7248,
|
| 68 |
"step": 10
|
| 69 |
},
|
| 70 |
{
|
| 71 |
+
"epoch": 0.09,
|
| 72 |
+
"learning_rate": 1.375e-05,
|
| 73 |
+
"loss": 0.7121,
|
| 74 |
"step": 11
|
| 75 |
},
|
| 76 |
{
|
| 77 |
"epoch": 0.09,
|
| 78 |
+
"learning_rate": 1.5000000000000002e-05,
|
| 79 |
+
"loss": 0.694,
|
| 80 |
"step": 12
|
| 81 |
},
|
| 82 |
{
|
| 83 |
"epoch": 0.1,
|
| 84 |
+
"learning_rate": 1.6250000000000002e-05,
|
| 85 |
+
"loss": 0.6793,
|
| 86 |
"step": 13
|
| 87 |
},
|
| 88 |
{
|
| 89 |
"epoch": 0.11,
|
| 90 |
+
"learning_rate": 1.7500000000000002e-05,
|
| 91 |
+
"loss": 0.6811,
|
| 92 |
"step": 14
|
| 93 |
},
|
| 94 |
{
|
| 95 |
"epoch": 0.12,
|
| 96 |
+
"learning_rate": 1.8750000000000002e-05,
|
| 97 |
+
"loss": 0.6715,
|
| 98 |
"step": 15
|
| 99 |
},
|
| 100 |
{
|
| 101 |
+
"epoch": 0.13,
|
| 102 |
+
"learning_rate": 2e-05,
|
| 103 |
+
"loss": 0.6723,
|
| 104 |
"step": 16
|
| 105 |
},
|
| 106 |
{
|
| 107 |
"epoch": 0.13,
|
| 108 |
+
"learning_rate": 1.9999629591162658e-05,
|
| 109 |
+
"loss": 0.6675,
|
| 110 |
"step": 17
|
| 111 |
},
|
| 112 |
{
|
| 113 |
"epoch": 0.14,
|
| 114 |
+
"learning_rate": 1.9998518392091163e-05,
|
| 115 |
+
"loss": 0.6539,
|
| 116 |
"step": 18
|
| 117 |
},
|
| 118 |
{
|
| 119 |
"epoch": 0.15,
|
| 120 |
+
"learning_rate": 1.9996666485105115e-05,
|
| 121 |
+
"loss": 0.6561,
|
| 122 |
"step": 19
|
| 123 |
},
|
| 124 |
{
|
| 125 |
+
"epoch": 0.16,
|
| 126 |
+
"learning_rate": 1.999407400739705e-05,
|
| 127 |
+
"loss": 0.6408,
|
| 128 |
"step": 20
|
| 129 |
},
|
| 130 |
{
|
| 131 |
"epoch": 0.16,
|
| 132 |
+
"learning_rate": 1.9990741151022302e-05,
|
| 133 |
+
"loss": 0.6491,
|
| 134 |
"step": 21
|
| 135 |
},
|
| 136 |
{
|
| 137 |
"epoch": 0.17,
|
| 138 |
+
"learning_rate": 1.9986668162884763e-05,
|
| 139 |
+
"loss": 0.6422,
|
| 140 |
"step": 22
|
| 141 |
},
|
| 142 |
{
|
| 143 |
"epoch": 0.18,
|
| 144 |
+
"learning_rate": 1.9981855344718587e-05,
|
| 145 |
+
"loss": 0.6339,
|
| 146 |
"step": 23
|
| 147 |
},
|
| 148 |
{
|
| 149 |
"epoch": 0.19,
|
| 150 |
+
"learning_rate": 1.997630305306586e-05,
|
| 151 |
+
"loss": 0.6329,
|
| 152 |
"step": 24
|
| 153 |
},
|
| 154 |
{
|
| 155 |
+
"epoch": 0.2,
|
| 156 |
+
"learning_rate": 1.997001169925015e-05,
|
| 157 |
+
"loss": 0.6177,
|
| 158 |
"step": 25
|
| 159 |
},
|
| 160 |
{
|
| 161 |
"epoch": 0.2,
|
| 162 |
+
"learning_rate": 1.996298174934608e-05,
|
| 163 |
+
"loss": 0.6184,
|
| 164 |
"step": 26
|
| 165 |
},
|
| 166 |
{
|
| 167 |
"epoch": 0.21,
|
| 168 |
+
"learning_rate": 1.9955213724144754e-05,
|
| 169 |
+
"loss": 0.636,
|
| 170 |
"step": 27
|
| 171 |
},
|
| 172 |
{
|
| 173 |
"epoch": 0.22,
|
| 174 |
+
"learning_rate": 1.994670819911521e-05,
|
| 175 |
+
"loss": 0.6313,
|
| 176 |
"step": 28
|
| 177 |
},
|
| 178 |
{
|
| 179 |
+
"epoch": 0.23,
|
| 180 |
+
"learning_rate": 1.9937465804361783e-05,
|
| 181 |
+
"loss": 0.61,
|
| 182 |
"step": 29
|
| 183 |
},
|
| 184 |
{
|
| 185 |
+
"epoch": 0.24,
|
| 186 |
+
"learning_rate": 1.9927487224577402e-05,
|
| 187 |
+
"loss": 0.6154,
|
| 188 |
"step": 30
|
| 189 |
},
|
| 190 |
{
|
| 191 |
"epoch": 0.24,
|
| 192 |
+
"learning_rate": 1.99167731989929e-05,
|
| 193 |
+
"loss": 0.6012,
|
| 194 |
"step": 31
|
| 195 |
},
|
| 196 |
{
|
| 197 |
"epoch": 0.25,
|
| 198 |
+
"learning_rate": 1.990532452132223e-05,
|
| 199 |
+
"loss": 0.5978,
|
| 200 |
"step": 32
|
| 201 |
},
|
| 202 |
{
|
| 203 |
+
"epoch": 0.26,
|
| 204 |
+
"learning_rate": 1.9893142039703662e-05,
|
| 205 |
+
"loss": 0.6062,
|
| 206 |
"step": 33
|
| 207 |
},
|
| 208 |
{
|
| 209 |
+
"epoch": 0.27,
|
| 210 |
+
"learning_rate": 1.9880226656636977e-05,
|
| 211 |
+
"loss": 0.6004,
|
| 212 |
"step": 34
|
| 213 |
},
|
| 214 |
{
|
| 215 |
"epoch": 0.27,
|
| 216 |
+
"learning_rate": 1.986657932891657e-05,
|
| 217 |
+
"loss": 0.6011,
|
| 218 |
"step": 35
|
| 219 |
},
|
| 220 |
{
|
| 221 |
"epoch": 0.28,
|
| 222 |
+
"learning_rate": 1.9852201067560607e-05,
|
| 223 |
+
"loss": 0.6008,
|
| 224 |
"step": 36
|
| 225 |
},
|
| 226 |
{
|
| 227 |
"epoch": 0.29,
|
| 228 |
+
"learning_rate": 1.98370929377361e-05,
|
| 229 |
+
"loss": 0.5943,
|
| 230 |
"step": 37
|
| 231 |
},
|
| 232 |
{
|
| 233 |
+
"epoch": 0.3,
|
| 234 |
+
"learning_rate": 1.982125605868001e-05,
|
| 235 |
+
"loss": 0.5906,
|
| 236 |
"step": 38
|
| 237 |
},
|
| 238 |
{
|
| 239 |
+
"epoch": 0.31,
|
| 240 |
+
"learning_rate": 1.9804691603616324e-05,
|
| 241 |
+
"loss": 0.5899,
|
| 242 |
"step": 39
|
| 243 |
},
|
| 244 |
{
|
| 245 |
"epoch": 0.31,
|
| 246 |
+
"learning_rate": 1.9787400799669155e-05,
|
| 247 |
+
"loss": 0.5848,
|
| 248 |
"step": 40
|
| 249 |
},
|
| 250 |
{
|
| 251 |
"epoch": 0.32,
|
| 252 |
+
"learning_rate": 1.976938492777182e-05,
|
| 253 |
+
"loss": 0.5921,
|
| 254 |
"step": 41
|
| 255 |
},
|
| 256 |
{
|
| 257 |
+
"epoch": 0.33,
|
| 258 |
+
"learning_rate": 1.9750645322571952e-05,
|
| 259 |
+
"loss": 0.5959,
|
| 260 |
"step": 42
|
| 261 |
},
|
| 262 |
{
|
| 263 |
+
"epoch": 0.34,
|
| 264 |
+
"learning_rate": 1.973118337233262e-05,
|
| 265 |
+
"loss": 0.5933,
|
| 266 |
"step": 43
|
| 267 |
},
|
| 268 |
{
|
| 269 |
+
"epoch": 0.35,
|
| 270 |
+
"learning_rate": 1.9711000518829505e-05,
|
| 271 |
+
"loss": 0.5845,
|
| 272 |
"step": 44
|
| 273 |
},
|
| 274 |
{
|
| 275 |
"epoch": 0.35,
|
| 276 |
+
"learning_rate": 1.9690098257244063e-05,
|
| 277 |
+
"loss": 0.5832,
|
| 278 |
"step": 45
|
| 279 |
},
|
| 280 |
{
|
| 281 |
"epoch": 0.36,
|
| 282 |
+
"learning_rate": 1.9668478136052776e-05,
|
| 283 |
+
"loss": 0.5766,
|
| 284 |
"step": 46
|
| 285 |
},
|
| 286 |
{
|
| 287 |
+
"epoch": 0.37,
|
| 288 |
+
"learning_rate": 1.9646141756912437e-05,
|
| 289 |
+
"loss": 0.5807,
|
| 290 |
"step": 47
|
| 291 |
},
|
| 292 |
{
|
| 293 |
+
"epoch": 0.38,
|
| 294 |
+
"learning_rate": 1.962309077454149e-05,
|
| 295 |
+
"loss": 0.581,
|
| 296 |
"step": 48
|
| 297 |
},
|
| 298 |
{
|
| 299 |
"epoch": 0.38,
|
| 300 |
+
"learning_rate": 1.9599326896597448e-05,
|
| 301 |
+
"loss": 0.5695,
|
| 302 |
"step": 49
|
| 303 |
},
|
| 304 |
{
|
| 305 |
"epoch": 0.39,
|
| 306 |
+
"learning_rate": 1.9574851883550395e-05,
|
| 307 |
+
"loss": 0.5633,
|
| 308 |
"step": 50
|
| 309 |
},
|
| 310 |
{
|
| 311 |
+
"epoch": 0.4,
|
| 312 |
+
"learning_rate": 1.9549667548552557e-05,
|
| 313 |
+
"loss": 0.5697,
|
| 314 |
"step": 51
|
| 315 |
},
|
| 316 |
{
|
| 317 |
+
"epoch": 0.41,
|
| 318 |
+
"learning_rate": 1.9523775757303975e-05,
|
| 319 |
+
"loss": 0.5708,
|
| 320 |
"step": 52
|
| 321 |
},
|
| 322 |
{
|
| 323 |
+
"epoch": 0.42,
|
| 324 |
+
"learning_rate": 1.949717842791432e-05,
|
| 325 |
+
"loss": 0.567,
|
| 326 |
"step": 53
|
| 327 |
},
|
| 328 |
{
|
| 329 |
"epoch": 0.42,
|
| 330 |
+
"learning_rate": 1.9469877530760753e-05,
|
| 331 |
+
"loss": 0.559,
|
| 332 |
"step": 54
|
| 333 |
},
|
| 334 |
{
|
| 335 |
+
"epoch": 0.43,
|
| 336 |
+
"learning_rate": 1.9441875088342e-05,
|
| 337 |
+
"loss": 0.5635,
|
| 338 |
"step": 55
|
| 339 |
},
|
| 340 |
{
|
| 341 |
+
"epoch": 0.44,
|
| 342 |
+
"learning_rate": 1.9413173175128472e-05,
|
| 343 |
+
"loss": 0.5557,
|
| 344 |
"step": 56
|
| 345 |
},
|
| 346 |
{
|
| 347 |
+
"epoch": 0.45,
|
| 348 |
+
"learning_rate": 1.9383773917408644e-05,
|
| 349 |
+
"loss": 0.5641,
|
| 350 |
"step": 57
|
| 351 |
},
|
| 352 |
{
|
| 353 |
"epoch": 0.45,
|
| 354 |
+
"learning_rate": 1.9353679493131486e-05,
|
| 355 |
+
"loss": 0.5536,
|
| 356 |
"step": 58
|
| 357 |
},
|
| 358 |
{
|
| 359 |
"epoch": 0.46,
|
| 360 |
+
"learning_rate": 1.9322892131745135e-05,
|
| 361 |
+
"loss": 0.5511,
|
| 362 |
"step": 59
|
| 363 |
},
|
| 364 |
{
|
| 365 |
+
"epoch": 0.47,
|
| 366 |
+
"learning_rate": 1.9291414114031744e-05,
|
| 367 |
+
"loss": 0.5628,
|
| 368 |
"step": 60
|
| 369 |
},
|
| 370 |
{
|
| 371 |
+
"epoch": 0.48,
|
| 372 |
+
"learning_rate": 1.92592477719385e-05,
|
| 373 |
+
"loss": 0.5494,
|
| 374 |
"step": 61
|
| 375 |
},
|
| 376 |
{
|
| 377 |
+
"epoch": 0.49,
|
| 378 |
+
"learning_rate": 1.9226395488404875e-05,
|
| 379 |
+
"loss": 0.5464,
|
| 380 |
"step": 62
|
| 381 |
},
|
| 382 |
{
|
| 383 |
"epoch": 0.49,
|
| 384 |
+
"learning_rate": 1.9192859697186105e-05,
|
| 385 |
+
"loss": 0.55,
|
| 386 |
"step": 63
|
| 387 |
},
|
| 388 |
{
|
| 389 |
+
"epoch": 0.5,
|
| 390 |
+
"learning_rate": 1.9158642882672873e-05,
|
| 391 |
+
"loss": 0.5601,
|
| 392 |
"step": 64
|
| 393 |
},
|
| 394 |
{
|
| 395 |
+
"epoch": 0.51,
|
| 396 |
+
"learning_rate": 1.9123747579707275e-05,
|
| 397 |
+
"loss": 0.5576,
|
| 398 |
"step": 65
|
| 399 |
},
|
| 400 |
{
|
| 401 |
+
"epoch": 0.52,
|
| 402 |
+
"learning_rate": 1.908817637339503e-05,
|
| 403 |
+
"loss": 0.5481,
|
| 404 |
"step": 66
|
| 405 |
},
|
| 406 |
{
|
| 407 |
+
"epoch": 0.53,
|
| 408 |
+
"learning_rate": 1.9051931898913977e-05,
|
| 409 |
+
"loss": 0.5434,
|
| 410 |
"step": 67
|
| 411 |
},
|
| 412 |
{
|
| 413 |
"epoch": 0.53,
|
| 414 |
+
"learning_rate": 1.9015016841318843e-05,
|
| 415 |
+
"loss": 0.5458,
|
| 416 |
"step": 68
|
| 417 |
},
|
| 418 |
{
|
| 419 |
+
"epoch": 0.54,
|
| 420 |
+
"learning_rate": 1.8977433935342338e-05,
|
| 421 |
+
"loss": 0.5365,
|
| 422 |
"step": 69
|
| 423 |
},
|
| 424 |
{
|
| 425 |
+
"epoch": 0.55,
|
| 426 |
+
"learning_rate": 1.8939185965192572e-05,
|
| 427 |
+
"loss": 0.5435,
|
| 428 |
"step": 70
|
| 429 |
},
|
| 430 |
{
|
| 431 |
+
"epoch": 0.56,
|
| 432 |
+
"learning_rate": 1.890027576434677e-05,
|
| 433 |
+
"loss": 0.5405,
|
| 434 |
"step": 71
|
| 435 |
},
|
| 436 |
{
|
| 437 |
"epoch": 0.56,
|
| 438 |
+
"learning_rate": 1.8860706215341383e-05,
|
| 439 |
+
"loss": 0.5433,
|
| 440 |
"step": 72
|
| 441 |
},
|
| 442 |
{
|
| 443 |
+
"epoch": 0.57,
|
| 444 |
+
"learning_rate": 1.8820480249558538e-05,
|
| 445 |
+
"loss": 0.5465,
|
| 446 |
"step": 73
|
| 447 |
},
|
| 448 |
{
|
| 449 |
+
"epoch": 0.58,
|
| 450 |
+
"learning_rate": 1.8779600847008884e-05,
|
| 451 |
+
"loss": 0.5344,
|
| 452 |
"step": 74
|
| 453 |
},
|
| 454 |
{
|
| 455 |
+
"epoch": 0.59,
|
| 456 |
+
"learning_rate": 1.873807103611081e-05,
|
| 457 |
+
"loss": 0.5386,
|
| 458 |
"step": 75
|
| 459 |
},
|
| 460 |
{
|
| 461 |
"epoch": 0.59,
|
| 462 |
+
"eval_loss": 0.5331106781959534,
|
| 463 |
+
"eval_runtime": 133.6436,
|
| 464 |
+
"eval_samples_per_second": 14.95,
|
| 465 |
+
"eval_steps_per_second": 0.082,
|
| 466 |
+
"step": 75
|
| 467 |
+
},
|
| 468 |
+
{
|
| 469 |
+
"epoch": 0.6,
|
| 470 |
+
"learning_rate": 1.869589389346611e-05,
|
| 471 |
+
"loss": 0.5341,
|
| 472 |
"step": 76
|
| 473 |
},
|
| 474 |
{
|
| 475 |
+
"epoch": 0.6,
|
| 476 |
+
"learning_rate": 1.8653072543632064e-05,
|
| 477 |
+
"loss": 0.5399,
|
| 478 |
"step": 77
|
| 479 |
},
|
| 480 |
{
|
| 481 |
+
"epoch": 0.61,
|
| 482 |
+
"learning_rate": 1.8609610158889943e-05,
|
| 483 |
+
"loss": 0.551,
|
| 484 |
"step": 78
|
| 485 |
},
|
| 486 |
{
|
| 487 |
+
"epoch": 0.62,
|
| 488 |
+
"learning_rate": 1.8565509959010037e-05,
|
| 489 |
+
"loss": 0.5491,
|
| 490 |
"step": 79
|
| 491 |
},
|
| 492 |
{
|
| 493 |
+
"epoch": 0.63,
|
| 494 |
+
"learning_rate": 1.8520775211013094e-05,
|
| 495 |
+
"loss": 0.5274,
|
| 496 |
"step": 80
|
| 497 |
},
|
| 498 |
{
|
| 499 |
+
"epoch": 0.64,
|
| 500 |
+
"learning_rate": 1.8475409228928314e-05,
|
| 501 |
+
"loss": 0.5343,
|
| 502 |
"step": 81
|
| 503 |
},
|
| 504 |
{
|
| 505 |
+
"epoch": 0.64,
|
| 506 |
+
"learning_rate": 1.842941537354783e-05,
|
| 507 |
+
"loss": 0.5207,
|
| 508 |
"step": 82
|
| 509 |
},
|
| 510 |
{
|
| 511 |
+
"epoch": 0.65,
|
| 512 |
+
"learning_rate": 1.8382797052177746e-05,
|
| 513 |
+
"loss": 0.5202,
|
| 514 |
"step": 83
|
| 515 |
},
|
| 516 |
{
|
| 517 |
+
"epoch": 0.66,
|
| 518 |
+
"learning_rate": 1.8335557718385702e-05,
|
| 519 |
+
"loss": 0.5165,
|
| 520 |
"step": 84
|
| 521 |
},
|
| 522 |
{
|
| 523 |
+
"epoch": 0.67,
|
| 524 |
+
"learning_rate": 1.8287700871745036e-05,
|
| 525 |
+
"loss": 0.5268,
|
| 526 |
"step": 85
|
| 527 |
},
|
| 528 |
{
|
| 529 |
+
"epoch": 0.67,
|
| 530 |
+
"learning_rate": 1.8239230057575542e-05,
|
| 531 |
+
"loss": 0.5263,
|
| 532 |
"step": 86
|
| 533 |
},
|
| 534 |
{
|
| 535 |
+
"epoch": 0.68,
|
| 536 |
+
"learning_rate": 1.81901488666808e-05,
|
| 537 |
+
"loss": 0.5139,
|
| 538 |
"step": 87
|
| 539 |
},
|
| 540 |
{
|
| 541 |
+
"epoch": 0.69,
|
| 542 |
+
"learning_rate": 1.814046093508218e-05,
|
| 543 |
+
"loss": 0.525,
|
| 544 |
"step": 88
|
| 545 |
},
|
| 546 |
{
|
| 547 |
+
"epoch": 0.7,
|
| 548 |
+
"learning_rate": 1.8090169943749477e-05,
|
| 549 |
+
"loss": 0.5191,
|
| 550 |
"step": 89
|
| 551 |
},
|
| 552 |
{
|
| 553 |
+
"epoch": 0.71,
|
| 554 |
+
"learning_rate": 1.8039279618328215e-05,
|
| 555 |
+
"loss": 0.5123,
|
| 556 |
"step": 90
|
| 557 |
},
|
| 558 |
{
|
| 559 |
+
"epoch": 0.71,
|
| 560 |
+
"learning_rate": 1.798779372886365e-05,
|
| 561 |
+
"loss": 0.512,
|
| 562 |
"step": 91
|
| 563 |
},
|
| 564 |
{
|
| 565 |
+
"epoch": 0.72,
|
| 566 |
+
"learning_rate": 1.7935716089521474e-05,
|
| 567 |
+
"loss": 0.5088,
|
| 568 |
"step": 92
|
| 569 |
},
|
| 570 |
{
|
| 571 |
+
"epoch": 0.73,
|
| 572 |
+
"learning_rate": 1.7883050558305255e-05,
|
| 573 |
+
"loss": 0.5223,
|
| 574 |
"step": 93
|
| 575 |
},
|
| 576 |
{
|
| 577 |
+
"epoch": 0.74,
|
| 578 |
+
"learning_rate": 1.7829801036770628e-05,
|
| 579 |
+
"loss": 0.5105,
|
| 580 |
"step": 94
|
| 581 |
},
|
| 582 |
{
|
| 583 |
+
"epoch": 0.75,
|
| 584 |
+
"learning_rate": 1.777597146973627e-05,
|
| 585 |
+
"loss": 0.5095,
|
| 586 |
"step": 95
|
| 587 |
},
|
| 588 |
{
|
| 589 |
+
"epoch": 0.75,
|
| 590 |
+
"learning_rate": 1.7721565844991643e-05,
|
| 591 |
+
"loss": 0.5159,
|
| 592 |
"step": 96
|
| 593 |
},
|
| 594 |
{
|
| 595 |
+
"epoch": 0.76,
|
| 596 |
+
"learning_rate": 1.7666588193001595e-05,
|
| 597 |
+
"loss": 0.5157,
|
| 598 |
"step": 97
|
| 599 |
},
|
| 600 |
{
|
| 601 |
+
"epoch": 0.77,
|
| 602 |
+
"learning_rate": 1.7611042586607748e-05,
|
| 603 |
+
"loss": 0.5081,
|
| 604 |
"step": 98
|
| 605 |
},
|
| 606 |
{
|
| 607 |
+
"epoch": 0.78,
|
| 608 |
+
"learning_rate": 1.7554933140726803e-05,
|
| 609 |
+
"loss": 0.5028,
|
| 610 |
"step": 99
|
| 611 |
},
|
| 612 |
{
|
| 613 |
+
"epoch": 0.78,
|
| 614 |
+
"learning_rate": 1.7498264012045686e-05,
|
| 615 |
+
"loss": 0.503,
|
| 616 |
"step": 100
|
| 617 |
},
|
| 618 |
{
|
| 619 |
+
"epoch": 0.79,
|
| 620 |
+
"learning_rate": 1.744103939871361e-05,
|
| 621 |
+
"loss": 0.5053,
|
| 622 |
"step": 101
|
| 623 |
},
|
| 624 |
{
|
| 625 |
+
"epoch": 0.8,
|
| 626 |
+
"learning_rate": 1.738326354003107e-05,
|
| 627 |
+
"loss": 0.5061,
|
| 628 |
"step": 102
|
| 629 |
},
|
| 630 |
{
|
| 631 |
+
"epoch": 0.81,
|
| 632 |
+
"learning_rate": 1.732494071613579e-05,
|
| 633 |
+
"loss": 0.497,
|
| 634 |
"step": 103
|
| 635 |
},
|
| 636 |
{
|
| 637 |
+
"epoch": 0.82,
|
| 638 |
+
"learning_rate": 1.7266075247685656e-05,
|
| 639 |
+
"loss": 0.5025,
|
| 640 |
"step": 104
|
| 641 |
},
|
| 642 |
{
|
| 643 |
+
"epoch": 0.82,
|
| 644 |
+
"learning_rate": 1.720667149553861e-05,
|
| 645 |
+
"loss": 0.5064,
|
| 646 |
"step": 105
|
| 647 |
},
|
| 648 |
{
|
| 649 |
+
"epoch": 0.83,
|
| 650 |
+
"learning_rate": 1.7146733860429614e-05,
|
| 651 |
+
"loss": 0.5102,
|
| 652 |
"step": 106
|
| 653 |
},
|
| 654 |
{
|
| 655 |
+
"epoch": 0.84,
|
| 656 |
+
"learning_rate": 1.70862667826446e-05,
|
| 657 |
+
"loss": 0.4977,
|
| 658 |
"step": 107
|
| 659 |
},
|
| 660 |
{
|
| 661 |
+
"epoch": 0.85,
|
| 662 |
+
"learning_rate": 1.702527474169157e-05,
|
| 663 |
+
"loss": 0.4939,
|
| 664 |
"step": 108
|
| 665 |
},
|
| 666 |
{
|
| 667 |
+
"epoch": 0.85,
|
| 668 |
+
"learning_rate": 1.6963762255968723e-05,
|
| 669 |
+
"loss": 0.4918,
|
| 670 |
"step": 109
|
| 671 |
},
|
| 672 |
{
|
| 673 |
+
"epoch": 0.86,
|
| 674 |
+
"learning_rate": 1.690173388242972e-05,
|
| 675 |
+
"loss": 0.4901,
|
| 676 |
"step": 110
|
| 677 |
},
|
| 678 |
{
|
| 679 |
+
"epoch": 0.87,
|
| 680 |
+
"learning_rate": 1.683919421624611e-05,
|
| 681 |
+
"loss": 0.4966,
|
| 682 |
"step": 111
|
| 683 |
},
|
| 684 |
{
|
| 685 |
+
"epoch": 0.88,
|
| 686 |
+
"learning_rate": 1.677614789046689e-05,
|
| 687 |
+
"loss": 0.5022,
|
| 688 |
"step": 112
|
| 689 |
},
|
| 690 |
{
|
| 691 |
+
"epoch": 0.89,
|
| 692 |
+
"learning_rate": 1.6712599575675318e-05,
|
| 693 |
+
"loss": 0.496,
|
| 694 |
"step": 113
|
| 695 |
},
|
| 696 |
{
|
| 697 |
+
"epoch": 0.89,
|
| 698 |
+
"learning_rate": 1.6648553979642867e-05,
|
| 699 |
+
"loss": 0.5017,
|
| 700 |
"step": 114
|
| 701 |
},
|
| 702 |
{
|
| 703 |
+
"epoch": 0.9,
|
| 704 |
+
"learning_rate": 1.658401584698049e-05,
|
| 705 |
+
"loss": 0.4869,
|
| 706 |
"step": 115
|
| 707 |
},
|
| 708 |
{
|
| 709 |
+
"epoch": 0.91,
|
| 710 |
+
"learning_rate": 1.6518989958787126e-05,
|
| 711 |
+
"loss": 0.4878,
|
| 712 |
"step": 116
|
| 713 |
},
|
| 714 |
{
|
| 715 |
+
"epoch": 0.92,
|
| 716 |
+
"learning_rate": 1.6453481132295507e-05,
|
| 717 |
+
"loss": 0.4862,
|
| 718 |
"step": 117
|
| 719 |
},
|
| 720 |
{
|
| 721 |
+
"epoch": 0.93,
|
| 722 |
+
"learning_rate": 1.6387494220515276e-05,
|
| 723 |
+
"loss": 0.482,
|
| 724 |
"step": 118
|
| 725 |
},
|
| 726 |
{
|
| 727 |
+
"epoch": 0.93,
|
| 728 |
+
"learning_rate": 1.6321034111873487e-05,
|
| 729 |
+
"loss": 0.4941,
|
| 730 |
"step": 119
|
| 731 |
},
|
| 732 |
{
|
| 733 |
+
"epoch": 0.94,
|
| 734 |
+
"learning_rate": 1.6254105729852466e-05,
|
| 735 |
+
"loss": 0.4868,
|
| 736 |
"step": 120
|
| 737 |
},
|
| 738 |
{
|
| 739 |
+
"epoch": 0.95,
|
| 740 |
+
"learning_rate": 1.6186714032625036e-05,
|
| 741 |
+
"loss": 0.483,
|
| 742 |
"step": 121
|
| 743 |
},
|
| 744 |
{
|
| 745 |
+
"epoch": 0.96,
|
| 746 |
+
"learning_rate": 1.6118864012687246e-05,
|
| 747 |
+
"loss": 0.4744,
|
| 748 |
"step": 122
|
| 749 |
},
|
| 750 |
{
|
| 751 |
+
"epoch": 0.96,
|
| 752 |
+
"learning_rate": 1.6050560696488493e-05,
|
| 753 |
+
"loss": 0.4859,
|
| 754 |
"step": 123
|
| 755 |
},
|
| 756 |
{
|
| 757 |
+
"epoch": 0.97,
|
| 758 |
+
"learning_rate": 1.5981809144059167e-05,
|
| 759 |
+
"loss": 0.4762,
|
| 760 |
"step": 124
|
| 761 |
},
|
| 762 |
{
|
| 763 |
+
"epoch": 0.98,
|
| 764 |
+
"learning_rate": 1.5912614448635784e-05,
|
| 765 |
+
"loss": 0.4767,
|
| 766 |
"step": 125
|
| 767 |
},
|
| 768 |
{
|
| 769 |
+
"epoch": 0.99,
|
| 770 |
+
"learning_rate": 1.5842981736283686e-05,
|
| 771 |
+
"loss": 0.4887,
|
| 772 |
"step": 126
|
| 773 |
},
|
| 774 |
{
|
| 775 |
+
"epoch": 1.0,
|
| 776 |
+
"learning_rate": 1.5772916165517275e-05,
|
| 777 |
+
"loss": 0.4775,
|
| 778 |
"step": 127
|
| 779 |
},
|
| 780 |
{
|
| 781 |
+
"epoch": 1.0,
|
| 782 |
+
"learning_rate": 1.5702422926917872e-05,
|
| 783 |
+
"loss": 0.4728,
|
| 784 |
"step": 128
|
| 785 |
},
|
| 786 |
{
|
| 787 |
+
"epoch": 1.01,
|
| 788 |
+
"learning_rate": 1.5631507242749187e-05,
|
| 789 |
+
"loss": 0.4653,
|
| 790 |
"step": 129
|
| 791 |
},
|
| 792 |
{
|
| 793 |
+
"epoch": 1.02,
|
| 794 |
+
"learning_rate": 1.5560174366570448e-05,
|
| 795 |
+
"loss": 0.4678,
|
| 796 |
+
"step": 130
|
| 797 |
+
},
|
| 798 |
+
{
|
| 799 |
+
"epoch": 1.03,
|
| 800 |
+
"learning_rate": 1.5488429582847194e-05,
|
| 801 |
+
"loss": 0.4668,
|
| 802 |
+
"step": 131
|
| 803 |
+
},
|
| 804 |
+
{
|
| 805 |
+
"epoch": 1.04,
|
| 806 |
+
"learning_rate": 1.5416278206559816e-05,
|
| 807 |
+
"loss": 0.4687,
|
| 808 |
+
"step": 132
|
| 809 |
+
},
|
| 810 |
+
{
|
| 811 |
+
"epoch": 1.04,
|
| 812 |
+
"learning_rate": 1.5343725582809793e-05,
|
| 813 |
+
"loss": 0.4735,
|
| 814 |
+
"step": 133
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 1.05,
|
| 818 |
+
"learning_rate": 1.5270777086423724e-05,
|
| 819 |
+
"loss": 0.4634,
|
| 820 |
+
"step": 134
|
| 821 |
+
},
|
| 822 |
+
{
|
| 823 |
+
"epoch": 1.06,
|
| 824 |
+
"learning_rate": 1.5197438121555159e-05,
|
| 825 |
+
"loss": 0.4583,
|
| 826 |
+
"step": 135
|
| 827 |
+
},
|
| 828 |
+
{
|
| 829 |
+
"epoch": 1.07,
|
| 830 |
+
"learning_rate": 1.512371412128424e-05,
|
| 831 |
+
"loss": 0.4587,
|
| 832 |
+
"step": 136
|
| 833 |
+
},
|
| 834 |
+
{
|
| 835 |
+
"epoch": 1.07,
|
| 836 |
+
"learning_rate": 1.5049610547215205e-05,
|
| 837 |
+
"loss": 0.4611,
|
| 838 |
+
"step": 137
|
| 839 |
+
},
|
| 840 |
+
{
|
| 841 |
+
"epoch": 1.08,
|
| 842 |
+
"learning_rate": 1.4975132889071808e-05,
|
| 843 |
+
"loss": 0.4568,
|
| 844 |
+
"step": 138
|
| 845 |
+
},
|
| 846 |
+
{
|
| 847 |
+
"epoch": 1.09,
|
| 848 |
+
"learning_rate": 1.4900286664290593e-05,
|
| 849 |
+
"loss": 0.4538,
|
| 850 |
+
"step": 139
|
| 851 |
+
},
|
| 852 |
+
{
|
| 853 |
+
"epoch": 1.1,
|
| 854 |
+
"learning_rate": 1.4825077417612187e-05,
|
| 855 |
+
"loss": 0.4601,
|
| 856 |
+
"step": 140
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 1.11,
|
| 860 |
+
"learning_rate": 1.4749510720670506e-05,
|
| 861 |
+
"loss": 0.4519,
|
| 862 |
+
"step": 141
|
| 863 |
+
},
|
| 864 |
+
{
|
| 865 |
+
"epoch": 1.11,
|
| 866 |
+
"learning_rate": 1.4673592171580025e-05,
|
| 867 |
+
"loss": 0.4524,
|
| 868 |
+
"step": 142
|
| 869 |
+
},
|
| 870 |
+
{
|
| 871 |
+
"epoch": 1.12,
|
| 872 |
+
"learning_rate": 1.4597327394521044e-05,
|
| 873 |
+
"loss": 0.4527,
|
| 874 |
+
"step": 143
|
| 875 |
+
},
|
| 876 |
+
{
|
| 877 |
+
"epoch": 1.13,
|
| 878 |
+
"learning_rate": 1.4520722039323045e-05,
|
| 879 |
+
"loss": 0.4409,
|
| 880 |
+
"step": 144
|
| 881 |
+
},
|
| 882 |
+
{
|
| 883 |
+
"epoch": 1.14,
|
| 884 |
+
"learning_rate": 1.4443781781046135e-05,
|
| 885 |
+
"loss": 0.4466,
|
| 886 |
+
"step": 145
|
| 887 |
+
},
|
| 888 |
+
{
|
| 889 |
+
"epoch": 1.15,
|
| 890 |
+
"learning_rate": 1.4366512319560642e-05,
|
| 891 |
+
"loss": 0.4597,
|
| 892 |
+
"step": 146
|
| 893 |
+
},
|
| 894 |
+
{
|
| 895 |
+
"epoch": 1.15,
|
| 896 |
+
"learning_rate": 1.4288919379124837e-05,
|
| 897 |
+
"loss": 0.4516,
|
| 898 |
+
"step": 147
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 1.16,
|
| 902 |
+
"learning_rate": 1.4211008707960897e-05,
|
| 903 |
+
"loss": 0.4529,
|
| 904 |
+
"step": 148
|
| 905 |
+
},
|
| 906 |
+
{
|
| 907 |
+
"epoch": 1.17,
|
| 908 |
+
"learning_rate": 1.4132786077829044e-05,
|
| 909 |
+
"loss": 0.4361,
|
| 910 |
+
"step": 149
|
| 911 |
+
},
|
| 912 |
+
{
|
| 913 |
+
"epoch": 1.18,
|
| 914 |
+
"learning_rate": 1.4054257283599974e-05,
|
| 915 |
+
"loss": 0.4501,
|
| 916 |
+
"step": 150
|
| 917 |
+
},
|
| 918 |
+
{
|
| 919 |
+
"epoch": 1.18,
|
| 920 |
+
"eval_loss": 0.4584537148475647,
|
| 921 |
+
"eval_runtime": 133.8838,
|
| 922 |
+
"eval_samples_per_second": 14.923,
|
| 923 |
+
"eval_steps_per_second": 0.082,
|
| 924 |
+
"step": 150
|
| 925 |
+
},
|
| 926 |
+
{
|
| 927 |
+
"epoch": 1.18,
|
| 928 |
+
"learning_rate": 1.3975428142825562e-05,
|
| 929 |
+
"loss": 0.4453,
|
| 930 |
+
"step": 151
|
| 931 |
+
},
|
| 932 |
+
{
|
| 933 |
+
"epoch": 1.19,
|
| 934 |
+
"learning_rate": 1.3896304495307881e-05,
|
| 935 |
+
"loss": 0.457,
|
| 936 |
+
"step": 152
|
| 937 |
+
},
|
| 938 |
+
{
|
| 939 |
+
"epoch": 1.2,
|
| 940 |
+
"learning_rate": 1.3816892202666591e-05,
|
| 941 |
+
"loss": 0.4441,
|
| 942 |
+
"step": 153
|
| 943 |
+
},
|
| 944 |
+
{
|
| 945 |
+
"epoch": 1.21,
|
| 946 |
+
"learning_rate": 1.373719714790469e-05,
|
| 947 |
+
"loss": 0.4449,
|
| 948 |
+
"step": 154
|
| 949 |
+
},
|
| 950 |
+
{
|
| 951 |
+
"epoch": 1.22,
|
| 952 |
+
"learning_rate": 1.3657225234972695e-05,
|
| 953 |
+
"loss": 0.4456,
|
| 954 |
+
"step": 155
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 1.22,
|
| 958 |
+
"learning_rate": 1.3576982388331258e-05,
|
| 959 |
+
"loss": 0.4332,
|
| 960 |
+
"step": 156
|
| 961 |
+
},
|
| 962 |
+
{
|
| 963 |
+
"epoch": 1.23,
|
| 964 |
+
"learning_rate": 1.3496474552512286e-05,
|
| 965 |
+
"loss": 0.4444,
|
| 966 |
+
"step": 157
|
| 967 |
+
},
|
| 968 |
+
{
|
| 969 |
+
"epoch": 1.24,
|
| 970 |
+
"learning_rate": 1.3415707691678557e-05,
|
| 971 |
+
"loss": 0.4381,
|
| 972 |
+
"step": 158
|
| 973 |
+
},
|
| 974 |
+
{
|
| 975 |
+
"epoch": 1.25,
|
| 976 |
+
"learning_rate": 1.333468778918187e-05,
|
| 977 |
+
"loss": 0.4391,
|
| 978 |
+
"step": 159
|
| 979 |
+
},
|
| 980 |
+
{
|
| 981 |
+
"epoch": 1.25,
|
| 982 |
+
"learning_rate": 1.3253420847119804e-05,
|
| 983 |
+
"loss": 0.4386,
|
| 984 |
+
"step": 160
|
| 985 |
+
},
|
| 986 |
+
{
|
| 987 |
+
"epoch": 1.26,
|
| 988 |
+
"learning_rate": 1.3171912885891063e-05,
|
| 989 |
+
"loss": 0.4534,
|
| 990 |
+
"step": 161
|
| 991 |
+
},
|
| 992 |
+
{
|
| 993 |
+
"epoch": 1.27,
|
| 994 |
+
"learning_rate": 1.3090169943749475e-05,
|
| 995 |
+
"loss": 0.4432,
|
| 996 |
+
"step": 162
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 1.28,
|
| 1000 |
+
"learning_rate": 1.3008198076356678e-05,
|
| 1001 |
+
"loss": 0.4248,
|
| 1002 |
+
"step": 163
|
| 1003 |
+
},
|
| 1004 |
+
{
|
| 1005 |
+
"epoch": 1.29,
|
| 1006 |
+
"learning_rate": 1.2926003356333487e-05,
|
| 1007 |
+
"loss": 0.4415,
|
| 1008 |
+
"step": 164
|
| 1009 |
+
},
|
| 1010 |
+
{
|
| 1011 |
+
"epoch": 1.29,
|
| 1012 |
+
"learning_rate": 1.2843591872810039e-05,
|
| 1013 |
+
"loss": 0.4472,
|
| 1014 |
+
"step": 165
|
| 1015 |
+
},
|
| 1016 |
+
{
|
| 1017 |
+
"epoch": 1.3,
|
| 1018 |
+
"learning_rate": 1.2760969730974692e-05,
|
| 1019 |
+
"loss": 0.4346,
|
| 1020 |
+
"step": 166
|
| 1021 |
+
},
|
| 1022 |
+
{
|
| 1023 |
+
"epoch": 1.31,
|
| 1024 |
+
"learning_rate": 1.2678143051621743e-05,
|
| 1025 |
+
"loss": 0.4407,
|
| 1026 |
+
"step": 167
|
| 1027 |
+
},
|
| 1028 |
+
{
|
| 1029 |
+
"epoch": 1.32,
|
| 1030 |
+
"learning_rate": 1.2595117970697998e-05,
|
| 1031 |
+
"loss": 0.4395,
|
| 1032 |
+
"step": 168
|
| 1033 |
+
},
|
| 1034 |
+
{
|
| 1035 |
+
"epoch": 1.33,
|
| 1036 |
+
"learning_rate": 1.2511900638848196e-05,
|
| 1037 |
+
"loss": 0.4393,
|
| 1038 |
+
"step": 169
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 1.33,
|
| 1042 |
+
"learning_rate": 1.2428497220959359e-05,
|
| 1043 |
+
"loss": 0.4309,
|
| 1044 |
+
"step": 170
|
| 1045 |
+
},
|
| 1046 |
+
{
|
| 1047 |
+
"epoch": 1.34,
|
| 1048 |
+
"learning_rate": 1.2344913895704099e-05,
|
| 1049 |
+
"loss": 0.4305,
|
| 1050 |
+
"step": 171
|
| 1051 |
+
},
|
| 1052 |
+
{
|
| 1053 |
+
"epoch": 1.35,
|
| 1054 |
+
"learning_rate": 1.2261156855082882e-05,
|
| 1055 |
+
"loss": 0.4319,
|
| 1056 |
+
"step": 172
|
| 1057 |
+
},
|
| 1058 |
+
{
|
| 1059 |
+
"epoch": 1.36,
|
| 1060 |
+
"learning_rate": 1.217723230396532e-05,
|
| 1061 |
+
"loss": 0.4325,
|
| 1062 |
+
"step": 173
|
| 1063 |
+
},
|
| 1064 |
+
{
|
| 1065 |
+
"epoch": 1.36,
|
| 1066 |
+
"learning_rate": 1.2093146459630488e-05,
|
| 1067 |
+
"loss": 0.4306,
|
| 1068 |
+
"step": 174
|
| 1069 |
+
},
|
| 1070 |
+
{
|
| 1071 |
+
"epoch": 1.37,
|
| 1072 |
+
"learning_rate": 1.2008905551306356e-05,
|
| 1073 |
+
"loss": 0.4294,
|
| 1074 |
+
"step": 175
|
| 1075 |
+
},
|
| 1076 |
+
{
|
| 1077 |
+
"epoch": 1.38,
|
| 1078 |
+
"learning_rate": 1.19245158197083e-05,
|
| 1079 |
+
"loss": 0.4298,
|
| 1080 |
+
"step": 176
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 1.39,
|
| 1084 |
+
"learning_rate": 1.1839983516576802e-05,
|
| 1085 |
+
"loss": 0.4295,
|
| 1086 |
+
"step": 177
|
| 1087 |
+
},
|
| 1088 |
+
{
|
| 1089 |
+
"epoch": 1.4,
|
| 1090 |
+
"learning_rate": 1.1755314904214284e-05,
|
| 1091 |
+
"loss": 0.4254,
|
| 1092 |
+
"step": 178
|
| 1093 |
+
},
|
| 1094 |
+
{
|
| 1095 |
+
"epoch": 1.4,
|
| 1096 |
+
"learning_rate": 1.1670516255021193e-05,
|
| 1097 |
+
"loss": 0.4307,
|
| 1098 |
+
"step": 179
|
| 1099 |
+
},
|
| 1100 |
+
{
|
| 1101 |
+
"epoch": 1.41,
|
| 1102 |
+
"learning_rate": 1.1585593851031346e-05,
|
| 1103 |
+
"loss": 0.4226,
|
| 1104 |
+
"step": 180
|
| 1105 |
+
},
|
| 1106 |
+
{
|
| 1107 |
+
"epoch": 1.42,
|
| 1108 |
+
"learning_rate": 1.1500553983446527e-05,
|
| 1109 |
+
"loss": 0.4245,
|
| 1110 |
+
"step": 181
|
| 1111 |
+
},
|
| 1112 |
+
{
|
| 1113 |
+
"epoch": 1.43,
|
| 1114 |
+
"learning_rate": 1.1415402952170434e-05,
|
| 1115 |
+
"loss": 0.4221,
|
| 1116 |
+
"step": 182
|
| 1117 |
+
},
|
| 1118 |
+
{
|
| 1119 |
+
"epoch": 1.44,
|
| 1120 |
+
"learning_rate": 1.1330147065341962e-05,
|
| 1121 |
+
"loss": 0.4283,
|
| 1122 |
+
"step": 183
|
| 1123 |
+
},
|
| 1124 |
+
{
|
| 1125 |
+
"epoch": 1.44,
|
| 1126 |
+
"learning_rate": 1.1244792638867895e-05,
|
| 1127 |
+
"loss": 0.4213,
|
| 1128 |
+
"step": 184
|
| 1129 |
+
},
|
| 1130 |
+
{
|
| 1131 |
+
"epoch": 1.45,
|
| 1132 |
+
"learning_rate": 1.1159345995955007e-05,
|
| 1133 |
+
"loss": 0.4293,
|
| 1134 |
+
"step": 185
|
| 1135 |
+
},
|
| 1136 |
+
{
|
| 1137 |
+
"epoch": 1.46,
|
| 1138 |
+
"learning_rate": 1.1073813466641633e-05,
|
| 1139 |
+
"loss": 0.4261,
|
| 1140 |
+
"step": 186
|
| 1141 |
+
},
|
| 1142 |
+
{
|
| 1143 |
+
"epoch": 1.47,
|
| 1144 |
+
"learning_rate": 1.0988201387328716e-05,
|
| 1145 |
+
"loss": 0.4278,
|
| 1146 |
+
"step": 187
|
| 1147 |
+
},
|
| 1148 |
+
{
|
| 1149 |
+
"epoch": 1.47,
|
| 1150 |
+
"learning_rate": 1.0902516100310412e-05,
|
| 1151 |
+
"loss": 0.4156,
|
| 1152 |
+
"step": 188
|
| 1153 |
+
},
|
| 1154 |
+
{
|
| 1155 |
+
"epoch": 1.48,
|
| 1156 |
+
"learning_rate": 1.0816763953304228e-05,
|
| 1157 |
+
"loss": 0.4334,
|
| 1158 |
+
"step": 189
|
| 1159 |
+
},
|
| 1160 |
+
{
|
| 1161 |
+
"epoch": 1.49,
|
| 1162 |
+
"learning_rate": 1.0730951298980776e-05,
|
| 1163 |
+
"loss": 0.4124,
|
| 1164 |
+
"step": 190
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 1.5,
|
| 1168 |
+
"learning_rate": 1.0645084494493166e-05,
|
| 1169 |
+
"loss": 0.4088,
|
| 1170 |
+
"step": 191
|
| 1171 |
+
},
|
| 1172 |
+
{
|
| 1173 |
+
"epoch": 1.51,
|
| 1174 |
+
"learning_rate": 1.0559169901006035e-05,
|
| 1175 |
+
"loss": 0.4168,
|
| 1176 |
+
"step": 192
|
| 1177 |
+
},
|
| 1178 |
+
{
|
| 1179 |
+
"epoch": 1.51,
|
| 1180 |
+
"learning_rate": 1.0473213883224321e-05,
|
| 1181 |
+
"loss": 0.4138,
|
| 1182 |
+
"step": 193
|
| 1183 |
+
},
|
| 1184 |
+
{
|
| 1185 |
+
"epoch": 1.52,
|
| 1186 |
+
"learning_rate": 1.0387222808921746e-05,
|
| 1187 |
+
"loss": 0.4206,
|
| 1188 |
+
"step": 194
|
| 1189 |
+
},
|
| 1190 |
+
{
|
| 1191 |
+
"epoch": 1.53,
|
| 1192 |
+
"learning_rate": 1.0301203048469084e-05,
|
| 1193 |
+
"loss": 0.4106,
|
| 1194 |
+
"step": 195
|
| 1195 |
+
},
|
| 1196 |
+
{
|
| 1197 |
+
"epoch": 1.54,
|
| 1198 |
+
"learning_rate": 1.0215160974362224e-05,
|
| 1199 |
+
"loss": 0.4095,
|
| 1200 |
+
"step": 196
|
| 1201 |
+
},
|
| 1202 |
+
{
|
| 1203 |
+
"epoch": 1.55,
|
| 1204 |
+
"learning_rate": 1.0129102960750092e-05,
|
| 1205 |
+
"loss": 0.412,
|
| 1206 |
+
"step": 197
|
| 1207 |
+
},
|
| 1208 |
+
{
|
| 1209 |
+
"epoch": 1.55,
|
| 1210 |
+
"learning_rate": 1.0043035382962443e-05,
|
| 1211 |
+
"loss": 0.4092,
|
| 1212 |
+
"step": 198
|
| 1213 |
+
},
|
| 1214 |
+
{
|
| 1215 |
+
"epoch": 1.56,
|
| 1216 |
+
"learning_rate": 9.956964617037559e-06,
|
| 1217 |
+
"loss": 0.4236,
|
| 1218 |
+
"step": 199
|
| 1219 |
+
},
|
| 1220 |
+
{
|
| 1221 |
+
"epoch": 1.57,
|
| 1222 |
+
"learning_rate": 9.87089703924991e-06,
|
| 1223 |
+
"loss": 0.4139,
|
| 1224 |
+
"step": 200
|
| 1225 |
+
},
|
| 1226 |
+
{
|
| 1227 |
+
"epoch": 1.58,
|
| 1228 |
+
"learning_rate": 9.78483902563778e-06,
|
| 1229 |
+
"loss": 0.4089,
|
| 1230 |
+
"step": 201
|
| 1231 |
+
},
|
| 1232 |
+
{
|
| 1233 |
+
"epoch": 1.58,
|
| 1234 |
+
"learning_rate": 9.69879695153092e-06,
|
| 1235 |
+
"loss": 0.4274,
|
| 1236 |
+
"step": 202
|
| 1237 |
+
},
|
| 1238 |
+
{
|
| 1239 |
+
"epoch": 1.59,
|
| 1240 |
+
"learning_rate": 9.612777191078257e-06,
|
| 1241 |
+
"loss": 0.4096,
|
| 1242 |
+
"step": 203
|
| 1243 |
+
},
|
| 1244 |
+
{
|
| 1245 |
+
"epoch": 1.6,
|
| 1246 |
+
"learning_rate": 9.526786116775682e-06,
|
| 1247 |
+
"loss": 0.4015,
|
| 1248 |
+
"step": 204
|
| 1249 |
+
},
|
| 1250 |
+
{
|
| 1251 |
+
"epoch": 1.61,
|
| 1252 |
+
"learning_rate": 9.440830098993969e-06,
|
| 1253 |
+
"loss": 0.4076,
|
| 1254 |
+
"step": 205
|
| 1255 |
+
},
|
| 1256 |
+
{
|
| 1257 |
+
"epoch": 1.62,
|
| 1258 |
+
"learning_rate": 9.354915505506839e-06,
|
| 1259 |
+
"loss": 0.4053,
|
| 1260 |
+
"step": 206
|
| 1261 |
+
},
|
| 1262 |
+
{
|
| 1263 |
+
"epoch": 1.62,
|
| 1264 |
+
"learning_rate": 9.269048701019226e-06,
|
| 1265 |
+
"loss": 0.4008,
|
| 1266 |
+
"step": 207
|
| 1267 |
+
},
|
| 1268 |
+
{
|
| 1269 |
+
"epoch": 1.63,
|
| 1270 |
+
"learning_rate": 9.183236046695777e-06,
|
| 1271 |
+
"loss": 0.4012,
|
| 1272 |
+
"step": 208
|
| 1273 |
+
},
|
| 1274 |
+
{
|
| 1275 |
+
"epoch": 1.64,
|
| 1276 |
+
"learning_rate": 9.09748389968959e-06,
|
| 1277 |
+
"loss": 0.407,
|
| 1278 |
+
"step": 209
|
| 1279 |
+
},
|
| 1280 |
+
{
|
| 1281 |
+
"epoch": 1.65,
|
| 1282 |
+
"learning_rate": 9.011798612671286e-06,
|
| 1283 |
+
"loss": 0.4024,
|
| 1284 |
+
"step": 210
|
| 1285 |
+
},
|
| 1286 |
+
{
|
| 1287 |
+
"epoch": 1.65,
|
| 1288 |
+
"learning_rate": 8.92618653335837e-06,
|
| 1289 |
+
"loss": 0.3949,
|
| 1290 |
+
"step": 211
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 1.66,
|
| 1294 |
+
"learning_rate": 8.840654004044996e-06,
|
| 1295 |
+
"loss": 0.4151,
|
| 1296 |
+
"step": 212
|
| 1297 |
+
},
|
| 1298 |
+
{
|
| 1299 |
+
"epoch": 1.67,
|
| 1300 |
+
"learning_rate": 8.755207361132109e-06,
|
| 1301 |
+
"loss": 0.4092,
|
| 1302 |
+
"step": 213
|
| 1303 |
+
},
|
| 1304 |
+
{
|
| 1305 |
+
"epoch": 1.68,
|
| 1306 |
+
"learning_rate": 8.669852934658042e-06,
|
| 1307 |
+
"loss": 0.3981,
|
| 1308 |
+
"step": 214
|
| 1309 |
+
},
|
| 1310 |
+
{
|
| 1311 |
+
"epoch": 1.69,
|
| 1312 |
+
"learning_rate": 8.58459704782957e-06,
|
| 1313 |
+
"loss": 0.3998,
|
| 1314 |
+
"step": 215
|
| 1315 |
+
},
|
| 1316 |
+
{
|
| 1317 |
+
"epoch": 1.69,
|
| 1318 |
+
"learning_rate": 8.499446016553475e-06,
|
| 1319 |
+
"loss": 0.3966,
|
| 1320 |
+
"step": 216
|
| 1321 |
+
},
|
| 1322 |
+
{
|
| 1323 |
+
"epoch": 1.7,
|
| 1324 |
+
"learning_rate": 8.414406148968657e-06,
|
| 1325 |
+
"loss": 0.4007,
|
| 1326 |
+
"step": 217
|
| 1327 |
+
},
|
| 1328 |
+
{
|
| 1329 |
+
"epoch": 1.71,
|
| 1330 |
+
"learning_rate": 8.32948374497881e-06,
|
| 1331 |
+
"loss": 0.393,
|
| 1332 |
+
"step": 218
|
| 1333 |
+
},
|
| 1334 |
+
{
|
| 1335 |
+
"epoch": 1.72,
|
| 1336 |
+
"learning_rate": 8.24468509578572e-06,
|
| 1337 |
+
"loss": 0.4059,
|
| 1338 |
+
"step": 219
|
| 1339 |
+
},
|
| 1340 |
+
{
|
| 1341 |
+
"epoch": 1.73,
|
| 1342 |
+
"learning_rate": 8.1600164834232e-06,
|
| 1343 |
+
"loss": 0.3988,
|
| 1344 |
+
"step": 220
|
| 1345 |
+
},
|
| 1346 |
+
{
|
| 1347 |
+
"epoch": 1.73,
|
| 1348 |
+
"learning_rate": 8.075484180291702e-06,
|
| 1349 |
+
"loss": 0.3905,
|
| 1350 |
+
"step": 221
|
| 1351 |
+
},
|
| 1352 |
+
{
|
| 1353 |
+
"epoch": 1.74,
|
| 1354 |
+
"learning_rate": 7.991094448693648e-06,
|
| 1355 |
+
"loss": 0.3898,
|
| 1356 |
+
"step": 222
|
| 1357 |
+
},
|
| 1358 |
+
{
|
| 1359 |
+
"epoch": 1.75,
|
| 1360 |
+
"learning_rate": 7.906853540369514e-06,
|
| 1361 |
+
"loss": 0.4007,
|
| 1362 |
+
"step": 223
|
| 1363 |
+
},
|
| 1364 |
+
{
|
| 1365 |
+
"epoch": 1.76,
|
| 1366 |
+
"learning_rate": 7.822767696034683e-06,
|
| 1367 |
+
"loss": 0.4119,
|
| 1368 |
+
"step": 224
|
| 1369 |
+
},
|
| 1370 |
+
{
|
| 1371 |
+
"epoch": 1.76,
|
| 1372 |
+
"learning_rate": 7.738843144917119e-06,
|
| 1373 |
+
"loss": 0.4006,
|
| 1374 |
+
"step": 225
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 1.76,
|
| 1378 |
+
"eval_loss": 0.4047410488128662,
|
| 1379 |
+
"eval_runtime": 133.554,
|
| 1380 |
+
"eval_samples_per_second": 14.96,
|
| 1381 |
+
"eval_steps_per_second": 0.082,
|
| 1382 |
+
"step": 225
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 1.78,
|
| 1386 |
+
"learning_rate": 7.655086104295904e-06,
|
| 1387 |
+
"loss": 0.3859,
|
| 1388 |
+
"step": 226
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 1.78,
|
| 1392 |
+
"learning_rate": 7.571502779040646e-06,
|
| 1393 |
+
"loss": 0.382,
|
| 1394 |
+
"step": 227
|
| 1395 |
+
},
|
| 1396 |
+
{
|
| 1397 |
+
"epoch": 1.79,
|
| 1398 |
+
"learning_rate": 7.4880993611518095e-06,
|
| 1399 |
+
"loss": 0.4043,
|
| 1400 |
+
"step": 228
|
| 1401 |
+
},
|
| 1402 |
+
{
|
| 1403 |
+
"epoch": 1.8,
|
| 1404 |
+
"learning_rate": 7.404882029302003e-06,
|
| 1405 |
+
"loss": 0.3882,
|
| 1406 |
+
"step": 229
|
| 1407 |
+
},
|
| 1408 |
+
{
|
| 1409 |
+
"epoch": 1.81,
|
| 1410 |
+
"learning_rate": 7.321856948378259e-06,
|
| 1411 |
+
"loss": 0.3896,
|
| 1412 |
+
"step": 230
|
| 1413 |
+
},
|
| 1414 |
+
{
|
| 1415 |
+
"epoch": 1.82,
|
| 1416 |
+
"learning_rate": 7.239030269025311e-06,
|
| 1417 |
+
"loss": 0.3835,
|
| 1418 |
+
"step": 231
|
| 1419 |
+
},
|
| 1420 |
+
{
|
| 1421 |
+
"epoch": 1.82,
|
| 1422 |
+
"learning_rate": 7.156408127189964e-06,
|
| 1423 |
+
"loss": 0.3916,
|
| 1424 |
+
"step": 232
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 1.83,
|
| 1428 |
+
"learning_rate": 7.073996643666516e-06,
|
| 1429 |
+
"loss": 0.3802,
|
| 1430 |
+
"step": 233
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 1.84,
|
| 1434 |
+
"learning_rate": 6.991801923643324e-06,
|
| 1435 |
+
"loss": 0.3859,
|
| 1436 |
+
"step": 234
|
| 1437 |
+
},
|
| 1438 |
+
{
|
| 1439 |
+
"epoch": 1.85,
|
| 1440 |
+
"learning_rate": 6.909830056250527e-06,
|
| 1441 |
+
"loss": 0.3984,
|
| 1442 |
+
"step": 235
|
| 1443 |
+
},
|
| 1444 |
+
{
|
| 1445 |
+
"epoch": 1.85,
|
| 1446 |
+
"learning_rate": 6.8280871141089415e-06,
|
| 1447 |
+
"loss": 0.3867,
|
| 1448 |
+
"step": 236
|
| 1449 |
+
},
|
| 1450 |
+
{
|
| 1451 |
+
"epoch": 1.86,
|
| 1452 |
+
"learning_rate": 6.746579152880201e-06,
|
| 1453 |
+
"loss": 0.3796,
|
| 1454 |
+
"step": 237
|
| 1455 |
+
},
|
| 1456 |
+
{
|
| 1457 |
+
"epoch": 1.87,
|
| 1458 |
+
"learning_rate": 6.665312210818133e-06,
|
| 1459 |
+
"loss": 0.3891,
|
| 1460 |
+
"step": 238
|
| 1461 |
+
},
|
| 1462 |
+
{
|
| 1463 |
+
"epoch": 1.88,
|
| 1464 |
+
"learning_rate": 6.584292308321445e-06,
|
| 1465 |
+
"loss": 0.3887,
|
| 1466 |
+
"step": 239
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 1.89,
|
| 1470 |
+
"learning_rate": 6.503525447487717e-06,
|
| 1471 |
+
"loss": 0.3732,
|
| 1472 |
+
"step": 240
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 1.89,
|
| 1476 |
+
"learning_rate": 6.423017611668745e-06,
|
| 1477 |
+
"loss": 0.3791,
|
| 1478 |
+
"step": 241
|
| 1479 |
+
},
|
| 1480 |
+
{
|
| 1481 |
+
"epoch": 1.9,
|
| 1482 |
+
"learning_rate": 6.342774765027309e-06,
|
| 1483 |
+
"loss": 0.3822,
|
| 1484 |
+
"step": 242
|
| 1485 |
+
},
|
| 1486 |
+
{
|
| 1487 |
+
"epoch": 1.91,
|
| 1488 |
+
"learning_rate": 6.262802852095311e-06,
|
| 1489 |
+
"loss": 0.3835,
|
| 1490 |
+
"step": 243
|
| 1491 |
+
},
|
| 1492 |
+
{
|
| 1493 |
+
"epoch": 1.92,
|
| 1494 |
+
"learning_rate": 6.183107797333411e-06,
|
| 1495 |
+
"loss": 0.3875,
|
| 1496 |
+
"step": 244
|
| 1497 |
+
},
|
| 1498 |
+
{
|
| 1499 |
+
"epoch": 1.93,
|
| 1500 |
+
"learning_rate": 6.103695504692122e-06,
|
| 1501 |
+
"loss": 0.3863,
|
| 1502 |
+
"step": 245
|
| 1503 |
+
},
|
| 1504 |
+
{
|
| 1505 |
+
"epoch": 1.93,
|
| 1506 |
+
"learning_rate": 6.024571857174443e-06,
|
| 1507 |
+
"loss": 0.3782,
|
| 1508 |
+
"step": 246
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 1.94,
|
| 1512 |
+
"learning_rate": 5.94574271640003e-06,
|
| 1513 |
+
"loss": 0.389,
|
| 1514 |
+
"step": 247
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 1.95,
|
| 1518 |
+
"learning_rate": 5.867213922170958e-06,
|
| 1519 |
+
"loss": 0.4024,
|
| 1520 |
+
"step": 248
|
| 1521 |
+
},
|
| 1522 |
+
{
|
| 1523 |
+
"epoch": 1.96,
|
| 1524 |
+
"learning_rate": 5.788991292039103e-06,
|
| 1525 |
+
"loss": 0.3772,
|
| 1526 |
+
"step": 249
|
| 1527 |
+
},
|
| 1528 |
+
{
|
| 1529 |
+
"epoch": 1.96,
|
| 1530 |
+
"learning_rate": 5.711080620875165e-06,
|
| 1531 |
+
"loss": 0.3833,
|
| 1532 |
+
"step": 250
|
| 1533 |
+
},
|
| 1534 |
+
{
|
| 1535 |
+
"epoch": 1.97,
|
| 1536 |
+
"learning_rate": 5.633487680439362e-06,
|
| 1537 |
+
"loss": 0.3753,
|
| 1538 |
+
"step": 251
|
| 1539 |
+
},
|
| 1540 |
+
{
|
| 1541 |
+
"epoch": 1.98,
|
| 1542 |
+
"learning_rate": 5.5562182189538684e-06,
|
| 1543 |
+
"loss": 0.3678,
|
| 1544 |
+
"step": 252
|
| 1545 |
+
},
|
| 1546 |
+
{
|
| 1547 |
+
"epoch": 1.99,
|
| 1548 |
+
"learning_rate": 5.479277960676959e-06,
|
| 1549 |
+
"loss": 0.3741,
|
| 1550 |
+
"step": 253
|
| 1551 |
+
},
|
| 1552 |
+
{
|
| 1553 |
+
"epoch": 2.0,
|
| 1554 |
+
"learning_rate": 5.40267260547896e-06,
|
| 1555 |
+
"loss": 0.3782,
|
| 1556 |
+
"step": 254
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 2.0,
|
| 1560 |
+
"learning_rate": 5.32640782841998e-06,
|
| 1561 |
+
"loss": 0.3763,
|
| 1562 |
+
"step": 255
|
| 1563 |
+
},
|
| 1564 |
+
{
|
| 1565 |
+
"epoch": 2.01,
|
| 1566 |
+
"learning_rate": 5.250489279329501e-06,
|
| 1567 |
+
"loss": 0.3668,
|
| 1568 |
+
"step": 256
|
| 1569 |
+
},
|
| 1570 |
+
{
|
| 1571 |
+
"epoch": 2.02,
|
| 1572 |
+
"learning_rate": 5.174922582387819e-06,
|
| 1573 |
+
"loss": 0.3743,
|
| 1574 |
+
"step": 257
|
| 1575 |
+
},
|
| 1576 |
+
{
|
| 1577 |
+
"epoch": 2.03,
|
| 1578 |
+
"learning_rate": 5.099713335709409e-06,
|
| 1579 |
+
"loss": 0.3786,
|
| 1580 |
+
"step": 258
|
| 1581 |
+
},
|
| 1582 |
+
{
|
| 1583 |
+
"epoch": 2.04,
|
| 1584 |
+
"learning_rate": 5.024867110928193e-06,
|
| 1585 |
+
"loss": 0.3723,
|
| 1586 |
+
"step": 259
|
| 1587 |
+
},
|
| 1588 |
+
{
|
| 1589 |
+
"epoch": 2.04,
|
| 1590 |
+
"learning_rate": 4.950389452784796e-06,
|
| 1591 |
+
"loss": 0.366,
|
| 1592 |
+
"step": 260
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 2.05,
|
| 1596 |
+
"learning_rate": 4.876285878715764e-06,
|
| 1597 |
+
"loss": 0.3774,
|
| 1598 |
+
"step": 261
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 2.06,
|
| 1602 |
+
"learning_rate": 4.802561878444846e-06,
|
| 1603 |
+
"loss": 0.3835,
|
| 1604 |
+
"step": 262
|
| 1605 |
+
},
|
| 1606 |
+
{
|
| 1607 |
+
"epoch": 2.07,
|
| 1608 |
+
"learning_rate": 4.729222913576279e-06,
|
| 1609 |
+
"loss": 0.386,
|
| 1610 |
+
"step": 263
|
| 1611 |
+
},
|
| 1612 |
+
{
|
| 1613 |
+
"epoch": 2.07,
|
| 1614 |
+
"learning_rate": 4.6562744171902144e-06,
|
| 1615 |
+
"loss": 0.3733,
|
| 1616 |
+
"step": 264
|
| 1617 |
+
},
|
| 1618 |
+
{
|
| 1619 |
+
"epoch": 2.08,
|
| 1620 |
+
"learning_rate": 4.583721793440188e-06,
|
| 1621 |
+
"loss": 0.3711,
|
| 1622 |
+
"step": 265
|
| 1623 |
+
},
|
| 1624 |
+
{
|
| 1625 |
+
"epoch": 2.09,
|
| 1626 |
+
"learning_rate": 4.5115704171528105e-06,
|
| 1627 |
+
"loss": 0.3713,
|
| 1628 |
+
"step": 266
|
| 1629 |
+
},
|
| 1630 |
+
{
|
| 1631 |
+
"epoch": 2.1,
|
| 1632 |
+
"learning_rate": 4.439825633429558e-06,
|
| 1633 |
+
"loss": 0.3718,
|
| 1634 |
+
"step": 267
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 2.11,
|
| 1638 |
+
"learning_rate": 4.368492757250814e-06,
|
| 1639 |
+
"loss": 0.3722,
|
| 1640 |
+
"step": 268
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 2.11,
|
| 1644 |
+
"learning_rate": 4.297577073082129e-06,
|
| 1645 |
+
"loss": 0.3598,
|
| 1646 |
+
"step": 269
|
| 1647 |
+
},
|
| 1648 |
+
{
|
| 1649 |
+
"epoch": 2.12,
|
| 1650 |
+
"learning_rate": 4.2270838344827285e-06,
|
| 1651 |
+
"loss": 0.3599,
|
| 1652 |
+
"step": 270
|
| 1653 |
+
},
|
| 1654 |
+
{
|
| 1655 |
+
"epoch": 2.13,
|
| 1656 |
+
"learning_rate": 4.1570182637163155e-06,
|
| 1657 |
+
"loss": 0.3643,
|
| 1658 |
+
"step": 271
|
| 1659 |
+
},
|
| 1660 |
+
{
|
| 1661 |
+
"epoch": 2.14,
|
| 1662 |
+
"learning_rate": 4.087385551364219e-06,
|
| 1663 |
+
"loss": 0.3601,
|
| 1664 |
+
"step": 272
|
| 1665 |
+
},
|
| 1666 |
+
{
|
| 1667 |
+
"epoch": 2.15,
|
| 1668 |
+
"learning_rate": 4.018190855940837e-06,
|
| 1669 |
+
"loss": 0.353,
|
| 1670 |
+
"step": 273
|
| 1671 |
+
},
|
| 1672 |
+
{
|
| 1673 |
+
"epoch": 2.15,
|
| 1674 |
+
"learning_rate": 3.949439303511512e-06,
|
| 1675 |
+
"loss": 0.3632,
|
| 1676 |
+
"step": 274
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"epoch": 2.16,
|
| 1680 |
+
"learning_rate": 3.881135987312758e-06,
|
| 1681 |
+
"loss": 0.3581,
|
| 1682 |
+
"step": 275
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 2.17,
|
| 1686 |
+
"learning_rate": 3.8132859673749688e-06,
|
| 1687 |
+
"loss": 0.3584,
|
| 1688 |
+
"step": 276
|
| 1689 |
+
},
|
| 1690 |
+
{
|
| 1691 |
+
"epoch": 2.18,
|
| 1692 |
+
"learning_rate": 3.7458942701475385e-06,
|
| 1693 |
+
"loss": 0.3581,
|
| 1694 |
+
"step": 277
|
| 1695 |
+
},
|
| 1696 |
+
{
|
| 1697 |
+
"epoch": 2.18,
|
| 1698 |
+
"learning_rate": 3.6789658881265135e-06,
|
| 1699 |
+
"loss": 0.349,
|
| 1700 |
+
"step": 278
|
| 1701 |
+
},
|
| 1702 |
+
{
|
| 1703 |
+
"epoch": 2.19,
|
| 1704 |
+
"learning_rate": 3.6125057794847286e-06,
|
| 1705 |
+
"loss": 0.3578,
|
| 1706 |
+
"step": 279
|
| 1707 |
+
},
|
| 1708 |
+
{
|
| 1709 |
+
"epoch": 2.2,
|
| 1710 |
+
"learning_rate": 3.546518867704499e-06,
|
| 1711 |
+
"loss": 0.3539,
|
| 1712 |
+
"step": 280
|
| 1713 |
+
},
|
| 1714 |
+
{
|
| 1715 |
+
"epoch": 2.21,
|
| 1716 |
+
"learning_rate": 3.4810100412128743e-06,
|
| 1717 |
+
"loss": 0.3551,
|
| 1718 |
+
"step": 281
|
| 1719 |
+
},
|
| 1720 |
+
{
|
| 1721 |
+
"epoch": 2.22,
|
| 1722 |
+
"learning_rate": 3.4159841530195127e-06,
|
| 1723 |
+
"loss": 0.3632,
|
| 1724 |
+
"step": 282
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 2.22,
|
| 1728 |
+
"learning_rate": 3.3514460203571365e-06,
|
| 1729 |
+
"loss": 0.3734,
|
| 1730 |
+
"step": 283
|
| 1731 |
+
},
|
| 1732 |
+
{
|
| 1733 |
+
"epoch": 2.23,
|
| 1734 |
+
"learning_rate": 3.287400424324687e-06,
|
| 1735 |
+
"loss": 0.3569,
|
| 1736 |
+
"step": 284
|
| 1737 |
+
},
|
| 1738 |
+
{
|
| 1739 |
+
"epoch": 2.24,
|
| 1740 |
+
"learning_rate": 3.223852109533112e-06,
|
| 1741 |
+
"loss": 0.3624,
|
| 1742 |
+
"step": 285
|
| 1743 |
+
},
|
| 1744 |
+
{
|
| 1745 |
+
"epoch": 2.25,
|
| 1746 |
+
"learning_rate": 3.1608057837538976e-06,
|
| 1747 |
+
"loss": 0.3573,
|
| 1748 |
+
"step": 286
|
| 1749 |
+
},
|
| 1750 |
+
{
|
| 1751 |
+
"epoch": 2.25,
|
| 1752 |
+
"learning_rate": 3.098266117570282e-06,
|
| 1753 |
+
"loss": 0.3534,
|
| 1754 |
+
"step": 287
|
| 1755 |
+
},
|
| 1756 |
+
{
|
| 1757 |
+
"epoch": 2.26,
|
| 1758 |
+
"learning_rate": 3.0362377440312783e-06,
|
| 1759 |
+
"loss": 0.3663,
|
| 1760 |
+
"step": 288
|
| 1761 |
+
},
|
| 1762 |
+
{
|
| 1763 |
+
"epoch": 2.27,
|
| 1764 |
+
"learning_rate": 2.9747252583084297e-06,
|
| 1765 |
+
"loss": 0.3715,
|
| 1766 |
+
"step": 289
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 2.28,
|
| 1770 |
+
"learning_rate": 2.9137332173554043e-06,
|
| 1771 |
+
"loss": 0.3627,
|
| 1772 |
+
"step": 290
|
| 1773 |
+
},
|
| 1774 |
+
{
|
| 1775 |
+
"epoch": 2.29,
|
| 1776 |
+
"learning_rate": 2.853266139570391e-06,
|
| 1777 |
+
"loss": 0.3691,
|
| 1778 |
+
"step": 291
|
| 1779 |
+
},
|
| 1780 |
+
{
|
| 1781 |
+
"epoch": 2.29,
|
| 1782 |
+
"learning_rate": 2.793328504461391e-06,
|
| 1783 |
+
"loss": 0.3465,
|
| 1784 |
+
"step": 292
|
| 1785 |
+
},
|
| 1786 |
+
{
|
| 1787 |
+
"epoch": 2.3,
|
| 1788 |
+
"learning_rate": 2.733924752314345e-06,
|
| 1789 |
+
"loss": 0.3411,
|
| 1790 |
+
"step": 293
|
| 1791 |
+
},
|
| 1792 |
+
{
|
| 1793 |
+
"epoch": 2.31,
|
| 1794 |
+
"learning_rate": 2.6750592838642144e-06,
|
| 1795 |
+
"loss": 0.3571,
|
| 1796 |
+
"step": 294
|
| 1797 |
+
},
|
| 1798 |
+
{
|
| 1799 |
+
"epoch": 2.32,
|
| 1800 |
+
"learning_rate": 2.616736459968936e-06,
|
| 1801 |
+
"loss": 0.3607,
|
| 1802 |
+
"step": 295
|
| 1803 |
+
},
|
| 1804 |
+
{
|
| 1805 |
+
"epoch": 2.33,
|
| 1806 |
+
"learning_rate": 2.5589606012863968e-06,
|
| 1807 |
+
"loss": 0.3573,
|
| 1808 |
+
"step": 296
|
| 1809 |
+
},
|
| 1810 |
+
{
|
| 1811 |
+
"epoch": 2.33,
|
| 1812 |
+
"learning_rate": 2.5017359879543168e-06,
|
| 1813 |
+
"loss": 0.3487,
|
| 1814 |
+
"step": 297
|
| 1815 |
+
},
|
| 1816 |
+
{
|
| 1817 |
+
"epoch": 2.34,
|
| 1818 |
+
"learning_rate": 2.4450668592731976e-06,
|
| 1819 |
+
"loss": 0.3565,
|
| 1820 |
+
"step": 298
|
| 1821 |
+
},
|
| 1822 |
+
{
|
| 1823 |
+
"epoch": 2.35,
|
| 1824 |
+
"learning_rate": 2.3889574133922532e-06,
|
| 1825 |
+
"loss": 0.3491,
|
| 1826 |
+
"step": 299
|
| 1827 |
+
},
|
| 1828 |
+
{
|
| 1829 |
+
"epoch": 2.36,
|
| 1830 |
+
"learning_rate": 2.3334118069984102e-06,
|
| 1831 |
+
"loss": 0.3525,
|
| 1832 |
+
"step": 300
|
| 1833 |
+
},
|
| 1834 |
+
{
|
| 1835 |
+
"epoch": 2.36,
|
| 1836 |
+
"learning_rate": 2.2784341550083577e-06,
|
| 1837 |
+
"loss": 0.3537,
|
| 1838 |
+
"step": 301
|
| 1839 |
+
},
|
| 1840 |
+
{
|
| 1841 |
+
"epoch": 2.37,
|
| 1842 |
+
"learning_rate": 2.224028530263733e-06,
|
| 1843 |
+
"loss": 0.3438,
|
| 1844 |
+
"step": 302
|
| 1845 |
+
},
|
| 1846 |
+
{
|
| 1847 |
+
"epoch": 2.38,
|
| 1848 |
+
"learning_rate": 2.170198963229372e-06,
|
| 1849 |
+
"loss": 0.3413,
|
| 1850 |
+
"step": 303
|
| 1851 |
+
},
|
| 1852 |
+
{
|
| 1853 |
+
"epoch": 2.39,
|
| 1854 |
+
"learning_rate": 2.116949441694748e-06,
|
| 1855 |
+
"loss": 0.3498,
|
| 1856 |
+
"step": 304
|
| 1857 |
+
},
|
| 1858 |
+
{
|
| 1859 |
+
"epoch": 2.4,
|
| 1860 |
+
"learning_rate": 2.064283910478527e-06,
|
| 1861 |
+
"loss": 0.3392,
|
| 1862 |
+
"step": 305
|
| 1863 |
+
},
|
| 1864 |
+
{
|
| 1865 |
+
"epoch": 2.4,
|
| 1866 |
+
"learning_rate": 2.012206271136353e-06,
|
| 1867 |
+
"loss": 0.3488,
|
| 1868 |
+
"step": 306
|
| 1869 |
+
},
|
| 1870 |
+
{
|
| 1871 |
+
"epoch": 2.41,
|
| 1872 |
+
"learning_rate": 1.960720381671789e-06,
|
| 1873 |
+
"loss": 0.3482,
|
| 1874 |
+
"step": 307
|
| 1875 |
+
},
|
| 1876 |
+
{
|
| 1877 |
+
"epoch": 2.42,
|
| 1878 |
+
"learning_rate": 1.9098300562505266e-06,
|
| 1879 |
+
"loss": 0.3423,
|
| 1880 |
+
"step": 308
|
| 1881 |
+
},
|
| 1882 |
+
{
|
| 1883 |
+
"epoch": 2.43,
|
| 1884 |
+
"learning_rate": 1.8595390649178214e-06,
|
| 1885 |
+
"loss": 0.3491,
|
| 1886 |
+
"step": 309
|
| 1887 |
+
},
|
| 1888 |
+
{
|
| 1889 |
+
"epoch": 2.44,
|
| 1890 |
+
"learning_rate": 1.8098511333192026e-06,
|
| 1891 |
+
"loss": 0.3621,
|
| 1892 |
+
"step": 310
|
| 1893 |
+
},
|
| 1894 |
+
{
|
| 1895 |
+
"epoch": 2.44,
|
| 1896 |
+
"learning_rate": 1.7607699424244583e-06,
|
| 1897 |
+
"loss": 0.3513,
|
| 1898 |
+
"step": 311
|
| 1899 |
+
},
|
| 1900 |
+
{
|
| 1901 |
+
"epoch": 2.45,
|
| 1902 |
+
"learning_rate": 1.712299128254965e-06,
|
| 1903 |
+
"loss": 0.3582,
|
| 1904 |
+
"step": 312
|
| 1905 |
+
},
|
| 1906 |
+
{
|
| 1907 |
+
"epoch": 2.46,
|
| 1908 |
+
"learning_rate": 1.6644422816143024e-06,
|
| 1909 |
+
"loss": 0.3523,
|
| 1910 |
+
"step": 313
|
| 1911 |
+
},
|
| 1912 |
+
{
|
| 1913 |
+
"epoch": 2.47,
|
| 1914 |
+
"learning_rate": 1.6172029478222595e-06,
|
| 1915 |
+
"loss": 0.3587,
|
| 1916 |
+
"step": 314
|
| 1917 |
+
},
|
| 1918 |
+
{
|
| 1919 |
+
"epoch": 2.47,
|
| 1920 |
+
"learning_rate": 1.5705846264521728e-06,
|
| 1921 |
+
"loss": 0.3462,
|
| 1922 |
+
"step": 315
|
| 1923 |
+
},
|
| 1924 |
+
{
|
| 1925 |
+
"epoch": 2.48,
|
| 1926 |
+
"learning_rate": 1.5245907710716912e-06,
|
| 1927 |
+
"loss": 0.3454,
|
| 1928 |
+
"step": 316
|
| 1929 |
+
},
|
| 1930 |
+
{
|
| 1931 |
+
"epoch": 2.49,
|
| 1932 |
+
"learning_rate": 1.4792247889869072e-06,
|
| 1933 |
+
"loss": 0.3502,
|
| 1934 |
+
"step": 317
|
| 1935 |
+
},
|
| 1936 |
+
{
|
| 1937 |
+
"epoch": 2.5,
|
| 1938 |
+
"learning_rate": 1.4344900409899643e-06,
|
| 1939 |
+
"loss": 0.34,
|
| 1940 |
+
"step": 318
|
| 1941 |
+
},
|
| 1942 |
+
{
|
| 1943 |
+
"epoch": 2.51,
|
| 1944 |
+
"learning_rate": 1.3903898411100568e-06,
|
| 1945 |
+
"loss": 0.3509,
|
| 1946 |
+
"step": 319
|
| 1947 |
+
},
|
| 1948 |
+
{
|
| 1949 |
+
"epoch": 2.51,
|
| 1950 |
+
"learning_rate": 1.3469274563679401e-06,
|
| 1951 |
+
"loss": 0.3432,
|
| 1952 |
+
"step": 320
|
| 1953 |
+
},
|
| 1954 |
+
{
|
| 1955 |
+
"epoch": 2.52,
|
| 1956 |
+
"learning_rate": 1.30410610653389e-06,
|
| 1957 |
+
"loss": 0.3499,
|
| 1958 |
+
"step": 321
|
| 1959 |
+
},
|
| 1960 |
+
{
|
| 1961 |
+
"epoch": 2.53,
|
| 1962 |
+
"learning_rate": 1.261928963889194e-06,
|
| 1963 |
+
"loss": 0.3355,
|
| 1964 |
+
"step": 322
|
| 1965 |
+
},
|
| 1966 |
+
{
|
| 1967 |
+
"epoch": 2.54,
|
| 1968 |
+
"learning_rate": 1.2203991529911197e-06,
|
| 1969 |
+
"loss": 0.3335,
|
| 1970 |
+
"step": 323
|
| 1971 |
+
},
|
| 1972 |
+
{
|
| 1973 |
+
"epoch": 2.55,
|
| 1974 |
+
"learning_rate": 1.1795197504414657e-06,
|
| 1975 |
+
"loss": 0.3507,
|
| 1976 |
+
"step": 324
|
| 1977 |
+
},
|
| 1978 |
+
{
|
| 1979 |
+
"epoch": 2.55,
|
| 1980 |
+
"learning_rate": 1.1392937846586216e-06,
|
| 1981 |
+
"loss": 0.3498,
|
| 1982 |
+
"step": 325
|
| 1983 |
+
},
|
| 1984 |
+
{
|
| 1985 |
+
"epoch": 2.56,
|
| 1986 |
+
"learning_rate": 1.0997242356532335e-06,
|
| 1987 |
+
"loss": 0.3567,
|
| 1988 |
+
"step": 326
|
| 1989 |
+
},
|
| 1990 |
+
{
|
| 1991 |
+
"epoch": 2.57,
|
| 1992 |
+
"learning_rate": 1.0608140348074292e-06,
|
| 1993 |
+
"loss": 0.3489,
|
| 1994 |
+
"step": 327
|
| 1995 |
+
},
|
| 1996 |
+
{
|
| 1997 |
+
"epoch": 2.58,
|
| 1998 |
+
"learning_rate": 1.022566064657663e-06,
|
| 1999 |
+
"loss": 0.3462,
|
| 2000 |
+
"step": 328
|
| 2001 |
+
},
|
| 2002 |
+
{
|
| 2003 |
+
"epoch": 2.58,
|
| 2004 |
+
"learning_rate": 9.849831586811597e-07,
|
| 2005 |
+
"loss": 0.3508,
|
| 2006 |
+
"step": 329
|
| 2007 |
+
},
|
| 2008 |
+
{
|
| 2009 |
+
"epoch": 2.59,
|
| 2010 |
+
"learning_rate": 9.48068101086026e-07,
|
| 2011 |
+
"loss": 0.3526,
|
| 2012 |
+
"step": 330
|
| 2013 |
+
},
|
| 2014 |
+
{
|
| 2015 |
+
"epoch": 2.6,
|
| 2016 |
+
"learning_rate": 9.118236266049707e-07,
|
| 2017 |
+
"loss": 0.3413,
|
| 2018 |
+
"step": 331
|
| 2019 |
+
},
|
| 2020 |
+
{
|
| 2021 |
+
"epoch": 2.61,
|
| 2022 |
+
"learning_rate": 8.762524202927281e-07,
|
| 2023 |
+
"loss": 0.3445,
|
| 2024 |
+
"step": 332
|
| 2025 |
+
},
|
| 2026 |
+
{
|
| 2027 |
+
"epoch": 2.62,
|
| 2028 |
+
"learning_rate": 8.413571173271295e-07,
|
| 2029 |
+
"loss": 0.3412,
|
| 2030 |
+
"step": 333
|
| 2031 |
+
},
|
| 2032 |
+
{
|
| 2033 |
+
"epoch": 2.62,
|
| 2034 |
+
"learning_rate": 8.071403028138969e-07,
|
| 2035 |
+
"loss": 0.3449,
|
| 2036 |
+
"step": 334
|
| 2037 |
+
},
|
| 2038 |
+
{
|
| 2039 |
+
"epoch": 2.63,
|
| 2040 |
+
"learning_rate": 7.736045115951252e-07,
|
| 2041 |
+
"loss": 0.3424,
|
| 2042 |
+
"step": 335
|
| 2043 |
+
},
|
| 2044 |
+
{
|
| 2045 |
+
"epoch": 2.64,
|
| 2046 |
+
"learning_rate": 7.40752228061502e-07,
|
| 2047 |
+
"loss": 0.3305,
|
| 2048 |
+
"step": 336
|
| 2049 |
+
},
|
| 2050 |
+
{
|
| 2051 |
+
"epoch": 2.65,
|
| 2052 |
+
"learning_rate": 7.08585885968257e-07,
|
| 2053 |
+
"loss": 0.3497,
|
| 2054 |
+
"step": 337
|
| 2055 |
+
},
|
| 2056 |
+
{
|
| 2057 |
+
"epoch": 2.65,
|
| 2058 |
+
"learning_rate": 6.77107868254867e-07,
|
| 2059 |
+
"loss": 0.3369,
|
| 2060 |
+
"step": 338
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"epoch": 2.66,
|
| 2064 |
+
"learning_rate": 6.463205068685174e-07,
|
| 2065 |
+
"loss": 0.3451,
|
| 2066 |
+
"step": 339
|
| 2067 |
+
},
|
| 2068 |
+
{
|
| 2069 |
+
"epoch": 2.67,
|
| 2070 |
+
"learning_rate": 6.162260825913591e-07,
|
| 2071 |
+
"loss": 0.3509,
|
| 2072 |
+
"step": 340
|
| 2073 |
+
},
|
| 2074 |
+
{
|
| 2075 |
+
"epoch": 2.68,
|
| 2076 |
+
"learning_rate": 5.868268248715292e-07,
|
| 2077 |
+
"loss": 0.3338,
|
| 2078 |
+
"step": 341
|
| 2079 |
+
},
|
| 2080 |
+
{
|
| 2081 |
+
"epoch": 2.69,
|
| 2082 |
+
"learning_rate": 5.581249116580068e-07,
|
| 2083 |
+
"loss": 0.3364,
|
| 2084 |
+
"step": 342
|
| 2085 |
+
},
|
| 2086 |
+
{
|
| 2087 |
+
"epoch": 2.69,
|
| 2088 |
+
"learning_rate": 5.301224692392482e-07,
|
| 2089 |
+
"loss": 0.3436,
|
| 2090 |
+
"step": 343
|
| 2091 |
+
},
|
| 2092 |
+
{
|
| 2093 |
+
"epoch": 2.7,
|
| 2094 |
+
"learning_rate": 5.028215720856821e-07,
|
| 2095 |
+
"loss": 0.3497,
|
| 2096 |
+
"step": 344
|
| 2097 |
+
},
|
| 2098 |
+
{
|
| 2099 |
+
"epoch": 2.71,
|
| 2100 |
+
"learning_rate": 4.762242426960262e-07,
|
| 2101 |
+
"loss": 0.3537,
|
| 2102 |
+
"step": 345
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 2.72,
|
| 2106 |
+
"learning_rate": 4.503324514474483e-07,
|
| 2107 |
+
"loss": 0.348,
|
| 2108 |
+
"step": 346
|
| 2109 |
+
},
|
| 2110 |
+
{
|
| 2111 |
+
"epoch": 2.73,
|
| 2112 |
+
"learning_rate": 4.2514811644960743e-07,
|
| 2113 |
+
"loss": 0.3428,
|
| 2114 |
+
"step": 347
|
| 2115 |
+
},
|
| 2116 |
+
{
|
| 2117 |
+
"epoch": 2.73,
|
| 2118 |
+
"learning_rate": 4.006731034025546e-07,
|
| 2119 |
+
"loss": 0.3347,
|
| 2120 |
+
"step": 348
|
| 2121 |
+
},
|
| 2122 |
+
{
|
| 2123 |
+
"epoch": 2.74,
|
| 2124 |
+
"learning_rate": 3.769092254585138e-07,
|
| 2125 |
+
"loss": 0.3334,
|
| 2126 |
+
"step": 349
|
| 2127 |
+
},
|
| 2128 |
+
{
|
| 2129 |
+
"epoch": 2.75,
|
| 2130 |
+
"learning_rate": 3.538582430875659e-07,
|
| 2131 |
+
"loss": 0.3421,
|
| 2132 |
+
"step": 350
|
| 2133 |
+
},
|
| 2134 |
+
{
|
| 2135 |
+
"epoch": 2.76,
|
| 2136 |
+
"learning_rate": 3.3152186394722506e-07,
|
| 2137 |
+
"loss": 0.3398,
|
| 2138 |
+
"step": 351
|
| 2139 |
+
},
|
| 2140 |
+
{
|
| 2141 |
+
"epoch": 2.76,
|
| 2142 |
+
"learning_rate": 3.099017427559392e-07,
|
| 2143 |
+
"loss": 0.3523,
|
| 2144 |
+
"step": 352
|
| 2145 |
+
},
|
| 2146 |
+
{
|
| 2147 |
+
"epoch": 2.77,
|
| 2148 |
+
"learning_rate": 2.889994811704966e-07,
|
| 2149 |
+
"loss": 0.3462,
|
| 2150 |
+
"step": 353
|
| 2151 |
+
},
|
| 2152 |
+
{
|
| 2153 |
+
"epoch": 2.78,
|
| 2154 |
+
"learning_rate": 2.688166276673809e-07,
|
| 2155 |
+
"loss": 0.3464,
|
| 2156 |
+
"step": 354
|
| 2157 |
+
},
|
| 2158 |
+
{
|
| 2159 |
+
"epoch": 2.79,
|
| 2160 |
+
"learning_rate": 2.493546774280531e-07,
|
| 2161 |
+
"loss": 0.3372,
|
| 2162 |
+
"step": 355
|
| 2163 |
+
},
|
| 2164 |
+
{
|
| 2165 |
+
"epoch": 2.8,
|
| 2166 |
+
"learning_rate": 2.3061507222818303e-07,
|
| 2167 |
+
"loss": 0.348,
|
| 2168 |
+
"step": 356
|
| 2169 |
+
},
|
| 2170 |
+
{
|
| 2171 |
+
"epoch": 2.8,
|
| 2172 |
+
"learning_rate": 2.1259920033084746e-07,
|
| 2173 |
+
"loss": 0.3441,
|
| 2174 |
+
"step": 357
|
| 2175 |
+
},
|
| 2176 |
+
{
|
| 2177 |
+
"epoch": 2.81,
|
| 2178 |
+
"learning_rate": 1.9530839638367993e-07,
|
| 2179 |
+
"loss": 0.3482,
|
| 2180 |
+
"step": 358
|
| 2181 |
+
},
|
| 2182 |
+
{
|
| 2183 |
+
"epoch": 2.82,
|
| 2184 |
+
"learning_rate": 1.7874394131999428e-07,
|
| 2185 |
+
"loss": 0.3477,
|
| 2186 |
+
"step": 359
|
| 2187 |
+
},
|
| 2188 |
+
{
|
| 2189 |
+
"epoch": 2.83,
|
| 2190 |
+
"learning_rate": 1.6290706226390286e-07,
|
| 2191 |
+
"loss": 0.3337,
|
| 2192 |
+
"step": 360
|
| 2193 |
+
},
|
| 2194 |
+
{
|
| 2195 |
+
"epoch": 2.84,
|
| 2196 |
+
"learning_rate": 1.4779893243939358e-07,
|
| 2197 |
+
"loss": 0.3315,
|
| 2198 |
+
"step": 361
|
| 2199 |
+
},
|
| 2200 |
+
{
|
| 2201 |
+
"epoch": 2.84,
|
| 2202 |
+
"learning_rate": 1.334206710834296e-07,
|
| 2203 |
+
"loss": 0.3502,
|
| 2204 |
+
"step": 362
|
| 2205 |
+
},
|
| 2206 |
+
{
|
| 2207 |
+
"epoch": 2.85,
|
| 2208 |
+
"learning_rate": 1.1977334336302437e-07,
|
| 2209 |
+
"loss": 0.339,
|
| 2210 |
+
"step": 363
|
| 2211 |
+
},
|
| 2212 |
+
{
|
| 2213 |
+
"epoch": 2.86,
|
| 2214 |
+
"learning_rate": 1.068579602963371e-07,
|
| 2215 |
+
"loss": 0.3485,
|
| 2216 |
+
"step": 364
|
| 2217 |
+
},
|
| 2218 |
+
{
|
| 2219 |
+
"epoch": 2.87,
|
| 2220 |
+
"learning_rate": 9.467547867777261e-08,
|
| 2221 |
+
"loss": 0.3477,
|
| 2222 |
+
"step": 365
|
| 2223 |
+
},
|
| 2224 |
+
{
|
| 2225 |
+
"epoch": 2.87,
|
| 2226 |
+
"learning_rate": 8.322680100710023e-08,
|
| 2227 |
+
"loss": 0.3462,
|
| 2228 |
+
"step": 366
|
| 2229 |
+
},
|
| 2230 |
+
{
|
| 2231 |
+
"epoch": 2.88,
|
| 2232 |
+
"learning_rate": 7.251277542259849e-08,
|
| 2233 |
+
"loss": 0.3354,
|
| 2234 |
+
"step": 367
|
| 2235 |
+
},
|
| 2236 |
+
{
|
| 2237 |
+
"epoch": 2.89,
|
| 2238 |
+
"learning_rate": 6.253419563821971e-08,
|
| 2239 |
+
"loss": 0.3463,
|
| 2240 |
+
"step": 368
|
| 2241 |
+
},
|
| 2242 |
+
{
|
| 2243 |
+
"epoch": 2.9,
|
| 2244 |
+
"learning_rate": 5.3291800884789356e-08,
|
| 2245 |
+
"loss": 0.3348,
|
| 2246 |
+
"step": 369
|
| 2247 |
+
},
|
| 2248 |
+
{
|
| 2249 |
+
"epoch": 2.91,
|
| 2250 |
+
"learning_rate": 4.478627585524753e-08,
|
| 2251 |
+
"loss": 0.3521,
|
| 2252 |
+
"step": 370
|
| 2253 |
+
},
|
| 2254 |
+
{
|
| 2255 |
+
"epoch": 2.91,
|
| 2256 |
+
"learning_rate": 3.701825065392184e-08,
|
| 2257 |
+
"loss": 0.338,
|
| 2258 |
+
"step": 371
|
| 2259 |
+
},
|
| 2260 |
+
{
|
| 2261 |
+
"epoch": 2.92,
|
| 2262 |
+
"learning_rate": 2.998830074984915e-08,
|
| 2263 |
+
"loss": 0.3485,
|
| 2264 |
+
"step": 372
|
| 2265 |
+
},
|
| 2266 |
+
{
|
| 2267 |
+
"epoch": 2.93,
|
| 2268 |
+
"learning_rate": 2.369694693414304e-08,
|
| 2269 |
+
"loss": 0.3411,
|
| 2270 |
+
"step": 373
|
| 2271 |
+
},
|
| 2272 |
+
{
|
| 2273 |
+
"epoch": 2.94,
|
| 2274 |
+
"learning_rate": 1.8144655281413515e-08,
|
| 2275 |
+
"loss": 0.3366,
|
| 2276 |
+
"step": 374
|
| 2277 |
+
},
|
| 2278 |
+
{
|
| 2279 |
+
"epoch": 2.95,
|
| 2280 |
+
"learning_rate": 1.3331837115241331e-08,
|
| 2281 |
+
"loss": 0.3324,
|
| 2282 |
+
"step": 375
|
| 2283 |
+
},
|
| 2284 |
+
{
|
| 2285 |
+
"epoch": 2.95,
|
| 2286 |
+
"learning_rate": 9.25884897770013e-09,
|
| 2287 |
+
"loss": 0.3386,
|
| 2288 |
+
"step": 376
|
| 2289 |
+
},
|
| 2290 |
+
{
|
| 2291 |
+
"epoch": 2.96,
|
| 2292 |
+
"learning_rate": 5.925992602952013e-09,
|
| 2293 |
+
"loss": 0.3407,
|
| 2294 |
+
"step": 377
|
| 2295 |
+
},
|
| 2296 |
+
{
|
| 2297 |
+
"epoch": 2.97,
|
| 2298 |
+
"learning_rate": 3.333514894887646e-09,
|
| 2299 |
+
"loss": 0.3356,
|
| 2300 |
+
"step": 378
|
| 2301 |
+
},
|
| 2302 |
+
{
|
| 2303 |
+
"epoch": 2.98,
|
| 2304 |
+
"learning_rate": 1.4816079088375567e-09,
|
| 2305 |
+
"loss": 0.3414,
|
| 2306 |
+
"step": 379
|
| 2307 |
+
},
|
| 2308 |
+
{
|
| 2309 |
+
"epoch": 2.98,
|
| 2310 |
+
"learning_rate": 3.7040883734462685e-10,
|
| 2311 |
+
"loss": 0.3343,
|
| 2312 |
+
"step": 380
|
| 2313 |
+
},
|
| 2314 |
+
{
|
| 2315 |
+
"epoch": 2.99,
|
| 2316 |
+
"learning_rate": 0.0,
|
| 2317 |
+
"loss": 0.3428,
|
| 2318 |
+
"step": 381
|
| 2319 |
+
},
|
| 2320 |
+
{
|
| 2321 |
+
"epoch": 2.99,
|
| 2322 |
+
"step": 381,
|
| 2323 |
+
"total_flos": 3.027228363560321e+18,
|
| 2324 |
+
"train_loss": 0.14632100531748274,
|
| 2325 |
+
"train_runtime": 23769.305,
|
| 2326 |
+
"train_samples_per_second": 12.351,
|
| 2327 |
+
"train_steps_per_second": 0.016
|
| 2328 |
}
|
| 2329 |
],
|
| 2330 |
+
"max_steps": 381,
|
| 2331 |
+
"num_train_epochs": 3,
|
| 2332 |
+
"total_flos": 3.027228363560321e+18,
|
| 2333 |
"trial_name": null,
|
| 2334 |
"trial_params": null
|
| 2335 |
}
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 3771
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cddaea0359842ed96f740f1330459b79b0918b7006f55346daf2721b4df05cd1
|
| 3 |
size 3771
|