Training in progress, epoch 2, checkpoint
Browse files
last-checkpoint/model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 498612824
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3597be7f38ca31c14f1645560a9ad68b700eca2390d2d5d010f4410debe47901
|
| 3 |
size 498612824
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 997348747
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8ad8d52b875e29d393979c3637e814c8ed7a68ca6941a162b2faa9bf43317d1b
|
| 3 |
size 997348747
|
last-checkpoint/rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14645
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2d8f98c42da24b5a5937db7695e7230d70157ebd07836d9e13ea06997d7dfc52
|
| 3 |
size 14645
|
last-checkpoint/scaler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1383
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d5c4224dc8a078df7a12c542aa90115ed7c8e4f973bf685055c317752c437b7b
|
| 3 |
size 1383
|
last-checkpoint/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1465
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:dba91d05f1f723fd97370b2f6d92b20cba8a1ad8076164d6d0112d9ad0660db4
|
| 3 |
size 1465
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -2,9 +2,9 @@
|
|
| 2 |
"best_global_step": 2393,
|
| 3 |
"best_metric": 0.9541757587780202,
|
| 4 |
"best_model_checkpoint": "./results_roberta_framing/checkpoint-2393",
|
| 5 |
-
"epoch":
|
| 6 |
"eval_steps": 500,
|
| 7 |
-
"global_step":
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
|
@@ -350,6 +350,355 @@
|
|
| 350 |
"eval_samples_per_second": 408.199,
|
| 351 |
"eval_steps_per_second": 25.547,
|
| 352 |
"step": 2393
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 353 |
}
|
| 354 |
],
|
| 355 |
"logging_steps": 50,
|
|
@@ -369,7 +718,7 @@
|
|
| 369 |
"attributes": {}
|
| 370 |
}
|
| 371 |
},
|
| 372 |
-
"total_flos":
|
| 373 |
"train_batch_size": 16,
|
| 374 |
"trial_name": null,
|
| 375 |
"trial_params": null
|
|
|
|
| 2 |
"best_global_step": 2393,
|
| 3 |
"best_metric": 0.9541757587780202,
|
| 4 |
"best_model_checkpoint": "./results_roberta_framing/checkpoint-2393",
|
| 5 |
+
"epoch": 2.0,
|
| 6 |
"eval_steps": 500,
|
| 7 |
+
"global_step": 4786,
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
|
|
|
| 350 |
"eval_samples_per_second": 408.199,
|
| 351 |
"eval_steps_per_second": 25.547,
|
| 352 |
"step": 2393
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 1.0029251984956122,
|
| 356 |
+
"grad_norm": 13.274906158447266,
|
| 357 |
+
"learning_rate": 1.3316617913358407e-05,
|
| 358 |
+
"loss": 0.12662715911865235,
|
| 359 |
+
"step": 2400
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 1.0238194734642707,
|
| 363 |
+
"grad_norm": 9.537711143493652,
|
| 364 |
+
"learning_rate": 1.3177322746900683e-05,
|
| 365 |
+
"loss": 0.1331467056274414,
|
| 366 |
+
"step": 2450
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 1.0447137484329294,
|
| 370 |
+
"grad_norm": 0.07932496070861816,
|
| 371 |
+
"learning_rate": 1.303802758044296e-05,
|
| 372 |
+
"loss": 0.07700533390045167,
|
| 373 |
+
"step": 2500
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 1.065608023401588,
|
| 377 |
+
"grad_norm": 22.845592498779297,
|
| 378 |
+
"learning_rate": 1.2898732413985235e-05,
|
| 379 |
+
"loss": 0.1407607078552246,
|
| 380 |
+
"step": 2550
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 1.0865022983702466,
|
| 384 |
+
"grad_norm": 1.8887174129486084,
|
| 385 |
+
"learning_rate": 1.2759437247527511e-05,
|
| 386 |
+
"loss": 0.14862350463867188,
|
| 387 |
+
"step": 2600
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 1.107396573338905,
|
| 391 |
+
"grad_norm": 21.409822463989258,
|
| 392 |
+
"learning_rate": 1.2620142081069789e-05,
|
| 393 |
+
"loss": 0.1066131591796875,
|
| 394 |
+
"step": 2650
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 1.1282908483075638,
|
| 398 |
+
"grad_norm": 9.874686241149902,
|
| 399 |
+
"learning_rate": 1.2480846914612063e-05,
|
| 400 |
+
"loss": 0.1277271842956543,
|
| 401 |
+
"step": 2700
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 1.1491851232762222,
|
| 405 |
+
"grad_norm": 32.69708251953125,
|
| 406 |
+
"learning_rate": 1.234155174815434e-05,
|
| 407 |
+
"loss": 0.10314902305603027,
|
| 408 |
+
"step": 2750
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 1.170079398244881,
|
| 412 |
+
"grad_norm": 40.25661849975586,
|
| 413 |
+
"learning_rate": 1.2202256581696617e-05,
|
| 414 |
+
"loss": 0.1239256763458252,
|
| 415 |
+
"step": 2800
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 1.1909736732135394,
|
| 419 |
+
"grad_norm": 0.40529486536979675,
|
| 420 |
+
"learning_rate": 1.2062961415238891e-05,
|
| 421 |
+
"loss": 0.13213342666625977,
|
| 422 |
+
"step": 2850
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 1.2118679481821981,
|
| 426 |
+
"grad_norm": 9.9985933303833,
|
| 427 |
+
"learning_rate": 1.1923666248781167e-05,
|
| 428 |
+
"loss": 0.1504405117034912,
|
| 429 |
+
"step": 2900
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 1.2327622231508566,
|
| 433 |
+
"grad_norm": 5.924779415130615,
|
| 434 |
+
"learning_rate": 1.1784371082323445e-05,
|
| 435 |
+
"loss": 0.11065505981445313,
|
| 436 |
+
"step": 2950
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 1.2536564981195153,
|
| 440 |
+
"grad_norm": 7.170877933502197,
|
| 441 |
+
"learning_rate": 1.164507591586572e-05,
|
| 442 |
+
"loss": 0.14150952339172362,
|
| 443 |
+
"step": 3000
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 1.2745507730881738,
|
| 447 |
+
"grad_norm": 7.508606433868408,
|
| 448 |
+
"learning_rate": 1.1505780749407995e-05,
|
| 449 |
+
"loss": 0.12164650917053223,
|
| 450 |
+
"step": 3050
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 1.2954450480568325,
|
| 454 |
+
"grad_norm": 2.847343683242798,
|
| 455 |
+
"learning_rate": 1.1366485582950273e-05,
|
| 456 |
+
"loss": 0.1189763069152832,
|
| 457 |
+
"step": 3100
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 1.316339323025491,
|
| 461 |
+
"grad_norm": 0.18800053000450134,
|
| 462 |
+
"learning_rate": 1.1227190416492549e-05,
|
| 463 |
+
"loss": 0.12123541831970215,
|
| 464 |
+
"step": 3150
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 1.3372335979941496,
|
| 468 |
+
"grad_norm": 9.2164306640625,
|
| 469 |
+
"learning_rate": 1.1087895250034826e-05,
|
| 470 |
+
"loss": 0.1492047691345215,
|
| 471 |
+
"step": 3200
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 1.3581278729628081,
|
| 475 |
+
"grad_norm": 3.368389844894409,
|
| 476 |
+
"learning_rate": 1.09486000835771e-05,
|
| 477 |
+
"loss": 0.12039382934570313,
|
| 478 |
+
"step": 3250
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 1.3790221479314668,
|
| 482 |
+
"grad_norm": 15.344005584716797,
|
| 483 |
+
"learning_rate": 1.0809304917119377e-05,
|
| 484 |
+
"loss": 0.17583621978759767,
|
| 485 |
+
"step": 3300
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 1.3999164229001253,
|
| 489 |
+
"grad_norm": 0.2738807797431946,
|
| 490 |
+
"learning_rate": 1.0670009750661654e-05,
|
| 491 |
+
"loss": 0.11576725006103515,
|
| 492 |
+
"step": 3350
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 1.420810697868784,
|
| 496 |
+
"grad_norm": 0.5546051263809204,
|
| 497 |
+
"learning_rate": 1.053071458420393e-05,
|
| 498 |
+
"loss": 0.09594018936157227,
|
| 499 |
+
"step": 3400
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 1.4417049728374425,
|
| 503 |
+
"grad_norm": 6.343163013458252,
|
| 504 |
+
"learning_rate": 1.0391419417746204e-05,
|
| 505 |
+
"loss": 0.15849972724914552,
|
| 506 |
+
"step": 3450
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 1.4625992478061012,
|
| 510 |
+
"grad_norm": 0.42352578043937683,
|
| 511 |
+
"learning_rate": 1.0252124251288482e-05,
|
| 512 |
+
"loss": 0.13367960929870606,
|
| 513 |
+
"step": 3500
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 1.4834935227747597,
|
| 517 |
+
"grad_norm": 0.37602102756500244,
|
| 518 |
+
"learning_rate": 1.0112829084830758e-05,
|
| 519 |
+
"loss": 0.16289531707763671,
|
| 520 |
+
"step": 3550
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 1.5043877977434184,
|
| 524 |
+
"grad_norm": 17.831235885620117,
|
| 525 |
+
"learning_rate": 9.973533918373032e-06,
|
| 526 |
+
"loss": 0.13586066246032716,
|
| 527 |
+
"step": 3600
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 1.525282072712077,
|
| 531 |
+
"grad_norm": 3.6913325786590576,
|
| 532 |
+
"learning_rate": 9.83423875191531e-06,
|
| 533 |
+
"loss": 0.1345800304412842,
|
| 534 |
+
"step": 3650
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 1.5461763476807355,
|
| 538 |
+
"grad_norm": 0.21697098016738892,
|
| 539 |
+
"learning_rate": 9.694943585457586e-06,
|
| 540 |
+
"loss": 0.13902374267578124,
|
| 541 |
+
"step": 3700
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 1.567070622649394,
|
| 545 |
+
"grad_norm": 8.487095832824707,
|
| 546 |
+
"learning_rate": 9.55564841899986e-06,
|
| 547 |
+
"loss": 0.0743906021118164,
|
| 548 |
+
"step": 3750
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 1.5879648976180527,
|
| 552 |
+
"grad_norm": 16.863210678100586,
|
| 553 |
+
"learning_rate": 9.416353252542138e-06,
|
| 554 |
+
"loss": 0.14349569320678712,
|
| 555 |
+
"step": 3800
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 1.6088591725867114,
|
| 559 |
+
"grad_norm": 16.5794677734375,
|
| 560 |
+
"learning_rate": 9.277058086084414e-06,
|
| 561 |
+
"loss": 0.1369628143310547,
|
| 562 |
+
"step": 3850
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 1.6297534475553699,
|
| 566 |
+
"grad_norm": 15.142793655395508,
|
| 567 |
+
"learning_rate": 9.13776291962669e-06,
|
| 568 |
+
"loss": 0.14577944755554198,
|
| 569 |
+
"step": 3900
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 1.6506477225240284,
|
| 573 |
+
"grad_norm": 0.13267600536346436,
|
| 574 |
+
"learning_rate": 8.998467753168966e-06,
|
| 575 |
+
"loss": 0.1497889518737793,
|
| 576 |
+
"step": 3950
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 1.671541997492687,
|
| 580 |
+
"grad_norm": 6.159015655517578,
|
| 581 |
+
"learning_rate": 8.859172586711242e-06,
|
| 582 |
+
"loss": 0.11653865814208984,
|
| 583 |
+
"step": 4000
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 1.6924362724613455,
|
| 587 |
+
"grad_norm": 8.872601509094238,
|
| 588 |
+
"learning_rate": 8.719877420253518e-06,
|
| 589 |
+
"loss": 0.10274889945983887,
|
| 590 |
+
"step": 4050
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 1.713330547430004,
|
| 594 |
+
"grad_norm": 9.018329620361328,
|
| 595 |
+
"learning_rate": 8.580582253795794e-06,
|
| 596 |
+
"loss": 0.11837130546569824,
|
| 597 |
+
"step": 4100
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 1.7342248223986627,
|
| 601 |
+
"grad_norm": 14.152815818786621,
|
| 602 |
+
"learning_rate": 8.44128708733807e-06,
|
| 603 |
+
"loss": 0.16490467071533202,
|
| 604 |
+
"step": 4150
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 1.7551190973673214,
|
| 608 |
+
"grad_norm": 14.32945442199707,
|
| 609 |
+
"learning_rate": 8.301991920880346e-06,
|
| 610 |
+
"loss": 0.1048060703277588,
|
| 611 |
+
"step": 4200
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 1.77601337233598,
|
| 615 |
+
"grad_norm": 6.154604434967041,
|
| 616 |
+
"learning_rate": 8.162696754422622e-06,
|
| 617 |
+
"loss": 0.16529315948486328,
|
| 618 |
+
"step": 4250
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 1.7969076473046384,
|
| 622 |
+
"grad_norm": 0.08094485104084015,
|
| 623 |
+
"learning_rate": 8.0234015879649e-06,
|
| 624 |
+
"loss": 0.1271997547149658,
|
| 625 |
+
"step": 4300
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 1.817801922273297,
|
| 629 |
+
"grad_norm": 0.6882657408714294,
|
| 630 |
+
"learning_rate": 7.884106421507174e-06,
|
| 631 |
+
"loss": 0.13333623886108398,
|
| 632 |
+
"step": 4350
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 1.8386961972419558,
|
| 636 |
+
"grad_norm": 0.6915619373321533,
|
| 637 |
+
"learning_rate": 7.74481125504945e-06,
|
| 638 |
+
"loss": 0.13177142143249512,
|
| 639 |
+
"step": 4400
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 1.8595904722106142,
|
| 643 |
+
"grad_norm": 0.04286617785692215,
|
| 644 |
+
"learning_rate": 7.605516088591726e-06,
|
| 645 |
+
"loss": 0.13941577911376954,
|
| 646 |
+
"step": 4450
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 1.8804847471792727,
|
| 650 |
+
"grad_norm": 0.24726086854934692,
|
| 651 |
+
"learning_rate": 7.466220922134002e-06,
|
| 652 |
+
"loss": 0.09164502143859864,
|
| 653 |
+
"step": 4500
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 1.9013790221479314,
|
| 657 |
+
"grad_norm": 19.788192749023438,
|
| 658 |
+
"learning_rate": 7.326925755676278e-06,
|
| 659 |
+
"loss": 0.16358205795288086,
|
| 660 |
+
"step": 4550
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 1.9222732971165901,
|
| 664 |
+
"grad_norm": 0.27708864212036133,
|
| 665 |
+
"learning_rate": 7.187630589218555e-06,
|
| 666 |
+
"loss": 0.08937115669250488,
|
| 667 |
+
"step": 4600
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 1.9431675720852486,
|
| 671 |
+
"grad_norm": 0.12228301912546158,
|
| 672 |
+
"learning_rate": 7.04833542276083e-06,
|
| 673 |
+
"loss": 0.10156312942504883,
|
| 674 |
+
"step": 4650
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 1.964061847053907,
|
| 678 |
+
"grad_norm": 0.11465885490179062,
|
| 679 |
+
"learning_rate": 6.909040256303107e-06,
|
| 680 |
+
"loss": 0.1402867317199707,
|
| 681 |
+
"step": 4700
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 1.9849561220225658,
|
| 685 |
+
"grad_norm": 0.2828459143638611,
|
| 686 |
+
"learning_rate": 6.769745089845383e-06,
|
| 687 |
+
"loss": 0.1601162338256836,
|
| 688 |
+
"step": 4750
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 2.0,
|
| 692 |
+
"eval_accuracy": 0.948699195486365,
|
| 693 |
+
"eval_f1": 0.9528564570331253,
|
| 694 |
+
"eval_loss": 0.16155333817005157,
|
| 695 |
+
"eval_precision": 0.9438843446832794,
|
| 696 |
+
"eval_recall": 0.9620007754943777,
|
| 697 |
+
"eval_roc_auc": 0.9904143280362305,
|
| 698 |
+
"eval_runtime": 23.4021,
|
| 699 |
+
"eval_samples_per_second": 408.98,
|
| 700 |
+
"eval_steps_per_second": 25.596,
|
| 701 |
+
"step": 4786
|
| 702 |
}
|
| 703 |
],
|
| 704 |
"logging_steps": 50,
|
|
|
|
| 718 |
"attributes": {}
|
| 719 |
}
|
| 720 |
},
|
| 721 |
+
"total_flos": 1.35284510223192e+16,
|
| 722 |
"train_batch_size": 16,
|
| 723 |
"trial_name": null,
|
| 724 |
"trial_params": null
|