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 368871908
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bc57420a3f9426bcaf177fea89f587ffe199692cb6b6daa1a7c14992d6327cc1
|
| 3 |
size 368871908
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 737866955
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:32996fbbfffabbaa0422ca2104929b4ab522fdedbade44c3c5b8d4666df0d633
|
| 3 |
size 737866955
|
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 sha256:580fb8da3e81024a65fdeba249660e74f25c2a71a809eebb9ff29aeaf9dd2f66
|
| 3 |
size 14645
|
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:ed01cb3ead55d9aead6eb70916572ffd5fc6d58c2675a2442748ceab27932dec
|
| 3 |
size 1465
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -2,9 +2,9 @@
|
|
| 2 |
"best_global_step": 2393,
|
| 3 |
"best_metric": 0.7003869916491275,
|
| 4 |
"best_model_checkpoint": "./results_deberta_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": 27.405,
|
| 351 |
"eval_steps_per_second": 1.715,
|
| 352 |
"step": 2393
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 353 |
}
|
| 354 |
],
|
| 355 |
"logging_steps": 50,
|
|
@@ -364,7 +713,7 @@
|
|
| 364 |
"early_stopping_threshold": 0.0
|
| 365 |
},
|
| 366 |
"attributes": {
|
| 367 |
-
"early_stopping_patience_counter":
|
| 368 |
}
|
| 369 |
},
|
| 370 |
"TrainerControl": {
|
|
@@ -378,7 +727,7 @@
|
|
| 378 |
"attributes": {}
|
| 379 |
}
|
| 380 |
},
|
| 381 |
-
"total_flos":
|
| 382 |
"train_batch_size": 16,
|
| 383 |
"trial_name": null,
|
| 384 |
"trial_params": null
|
|
|
|
| 2 |
"best_global_step": 2393,
|
| 3 |
"best_metric": 0.7003869916491275,
|
| 4 |
"best_model_checkpoint": "./results_deberta_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": 27.405,
|
| 351 |
"eval_steps_per_second": 1.715,
|
| 352 |
"step": 2393
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 1.0029251984956122,
|
| 356 |
+
"grad_norm": 2.685546875,
|
| 357 |
+
"learning_rate": 7.083580320094844e-06,
|
| 358 |
+
"loss": 0.6907852172851563,
|
| 359 |
+
"step": 2400
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 1.0238194734642707,
|
| 363 |
+
"grad_norm": 4.87890625,
|
| 364 |
+
"learning_rate": 7.009484291641968e-06,
|
| 365 |
+
"loss": 0.6929679870605469,
|
| 366 |
+
"step": 2450
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 1.0447137484329294,
|
| 370 |
+
"grad_norm": 0.73681640625,
|
| 371 |
+
"learning_rate": 6.935388263189093e-06,
|
| 372 |
+
"loss": 0.6920276641845703,
|
| 373 |
+
"step": 2500
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 1.065608023401588,
|
| 377 |
+
"grad_norm": 1.796875,
|
| 378 |
+
"learning_rate": 6.8612922347362184e-06,
|
| 379 |
+
"loss": 0.6908602905273438,
|
| 380 |
+
"step": 2550
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 1.0865022983702466,
|
| 384 |
+
"grad_norm": 6.328125,
|
| 385 |
+
"learning_rate": 6.787196206283344e-06,
|
| 386 |
+
"loss": 0.6924705505371094,
|
| 387 |
+
"step": 2600
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 1.107396573338905,
|
| 391 |
+
"grad_norm": 0.7470703125,
|
| 392 |
+
"learning_rate": 6.713100177830469e-06,
|
| 393 |
+
"loss": 0.6856446838378907,
|
| 394 |
+
"step": 2650
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 1.1282908483075638,
|
| 398 |
+
"grad_norm": 2.78515625,
|
| 399 |
+
"learning_rate": 6.639004149377594e-06,
|
| 400 |
+
"loss": 0.6915918731689453,
|
| 401 |
+
"step": 2700
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 1.1491851232762222,
|
| 405 |
+
"grad_norm": 0.72802734375,
|
| 406 |
+
"learning_rate": 6.564908120924719e-06,
|
| 407 |
+
"loss": 0.6918558502197265,
|
| 408 |
+
"step": 2750
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 1.170079398244881,
|
| 412 |
+
"grad_norm": 1.701171875,
|
| 413 |
+
"learning_rate": 6.490812092471844e-06,
|
| 414 |
+
"loss": 0.6948513793945312,
|
| 415 |
+
"step": 2800
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 1.1909736732135394,
|
| 419 |
+
"grad_norm": 2.4375,
|
| 420 |
+
"learning_rate": 6.41671606401897e-06,
|
| 421 |
+
"loss": 0.6888138580322266,
|
| 422 |
+
"step": 2850
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 1.2118679481821981,
|
| 426 |
+
"grad_norm": 4.6640625,
|
| 427 |
+
"learning_rate": 6.342620035566094e-06,
|
| 428 |
+
"loss": 0.6906478881835938,
|
| 429 |
+
"step": 2900
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 1.2327622231508566,
|
| 433 |
+
"grad_norm": 2.67578125,
|
| 434 |
+
"learning_rate": 6.268524007113219e-06,
|
| 435 |
+
"loss": 0.6887605285644531,
|
| 436 |
+
"step": 2950
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 1.2536564981195153,
|
| 440 |
+
"grad_norm": 1.1572265625,
|
| 441 |
+
"learning_rate": 6.194427978660344e-06,
|
| 442 |
+
"loss": 0.6858110809326172,
|
| 443 |
+
"step": 3000
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 1.2745507730881738,
|
| 447 |
+
"grad_norm": 1.2041015625,
|
| 448 |
+
"learning_rate": 6.1203319502074694e-06,
|
| 449 |
+
"loss": 0.6906262969970703,
|
| 450 |
+
"step": 3050
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 1.2954450480568325,
|
| 454 |
+
"grad_norm": 3.1953125,
|
| 455 |
+
"learning_rate": 6.046235921754594e-06,
|
| 456 |
+
"loss": 0.6898616027832031,
|
| 457 |
+
"step": 3100
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 1.316339323025491,
|
| 461 |
+
"grad_norm": 0.80615234375,
|
| 462 |
+
"learning_rate": 5.97213989330172e-06,
|
| 463 |
+
"loss": 0.6915110778808594,
|
| 464 |
+
"step": 3150
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 1.3372335979941496,
|
| 468 |
+
"grad_norm": 0.95458984375,
|
| 469 |
+
"learning_rate": 5.898043864848845e-06,
|
| 470 |
+
"loss": 0.69228271484375,
|
| 471 |
+
"step": 3200
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 1.3581278729628081,
|
| 475 |
+
"grad_norm": 6.22265625,
|
| 476 |
+
"learning_rate": 5.82394783639597e-06,
|
| 477 |
+
"loss": 0.6870991516113282,
|
| 478 |
+
"step": 3250
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 1.3790221479314668,
|
| 482 |
+
"grad_norm": 2.3359375,
|
| 483 |
+
"learning_rate": 5.749851807943094e-06,
|
| 484 |
+
"loss": 0.6957393646240234,
|
| 485 |
+
"step": 3300
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 1.3999164229001253,
|
| 489 |
+
"grad_norm": 0.86767578125,
|
| 490 |
+
"learning_rate": 5.675755779490219e-06,
|
| 491 |
+
"loss": 0.6911115264892578,
|
| 492 |
+
"step": 3350
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 1.420810697868784,
|
| 496 |
+
"grad_norm": 1.87109375,
|
| 497 |
+
"learning_rate": 5.601659751037345e-06,
|
| 498 |
+
"loss": 0.6849404907226563,
|
| 499 |
+
"step": 3400
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 1.4417049728374425,
|
| 503 |
+
"grad_norm": 1.0029296875,
|
| 504 |
+
"learning_rate": 5.52756372258447e-06,
|
| 505 |
+
"loss": 0.6912786102294922,
|
| 506 |
+
"step": 3450
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 1.4625992478061012,
|
| 510 |
+
"grad_norm": 1.4169921875,
|
| 511 |
+
"learning_rate": 5.4534676941315945e-06,
|
| 512 |
+
"loss": 0.6927188110351562,
|
| 513 |
+
"step": 3500
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 1.4834935227747597,
|
| 517 |
+
"grad_norm": 2.767578125,
|
| 518 |
+
"learning_rate": 5.37937166567872e-06,
|
| 519 |
+
"loss": 0.688505859375,
|
| 520 |
+
"step": 3550
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 1.5043877977434184,
|
| 524 |
+
"grad_norm": 0.49169921875,
|
| 525 |
+
"learning_rate": 5.3052756372258455e-06,
|
| 526 |
+
"loss": 0.6911637878417969,
|
| 527 |
+
"step": 3600
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 1.525282072712077,
|
| 531 |
+
"grad_norm": 2.513671875,
|
| 532 |
+
"learning_rate": 5.231179608772971e-06,
|
| 533 |
+
"loss": 0.6921497344970703,
|
| 534 |
+
"step": 3650
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 1.5461763476807355,
|
| 538 |
+
"grad_norm": 1.263671875,
|
| 539 |
+
"learning_rate": 5.157083580320096e-06,
|
| 540 |
+
"loss": 0.6940757751464843,
|
| 541 |
+
"step": 3700
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 1.567070622649394,
|
| 545 |
+
"grad_norm": 3.740234375,
|
| 546 |
+
"learning_rate": 5.08298755186722e-06,
|
| 547 |
+
"loss": 0.6916998291015625,
|
| 548 |
+
"step": 3750
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 1.5879648976180527,
|
| 552 |
+
"grad_norm": 1.34765625,
|
| 553 |
+
"learning_rate": 5.008891523414345e-06,
|
| 554 |
+
"loss": 0.6906791687011719,
|
| 555 |
+
"step": 3800
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 1.6088591725867114,
|
| 559 |
+
"grad_norm": 2.775390625,
|
| 560 |
+
"learning_rate": 4.934795494961471e-06,
|
| 561 |
+
"loss": 0.6903327941894531,
|
| 562 |
+
"step": 3850
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 1.6297534475553699,
|
| 566 |
+
"grad_norm": 3.341796875,
|
| 567 |
+
"learning_rate": 4.860699466508595e-06,
|
| 568 |
+
"loss": 0.6916295623779297,
|
| 569 |
+
"step": 3900
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 1.6506477225240284,
|
| 573 |
+
"grad_norm": 1.8291015625,
|
| 574 |
+
"learning_rate": 4.78660343805572e-06,
|
| 575 |
+
"loss": 0.6899072265625,
|
| 576 |
+
"step": 3950
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 1.671541997492687,
|
| 580 |
+
"grad_norm": 2.580078125,
|
| 581 |
+
"learning_rate": 4.7125074096028455e-06,
|
| 582 |
+
"loss": 0.6946562194824218,
|
| 583 |
+
"step": 4000
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 1.6924362724613455,
|
| 587 |
+
"grad_norm": 5.171875,
|
| 588 |
+
"learning_rate": 4.638411381149971e-06,
|
| 589 |
+
"loss": 0.692690200805664,
|
| 590 |
+
"step": 4050
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 1.713330547430004,
|
| 594 |
+
"grad_norm": 0.90478515625,
|
| 595 |
+
"learning_rate": 4.564315352697096e-06,
|
| 596 |
+
"loss": 0.6910479736328125,
|
| 597 |
+
"step": 4100
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 1.7342248223986627,
|
| 601 |
+
"grad_norm": 1.8486328125,
|
| 602 |
+
"learning_rate": 4.490219324244221e-06,
|
| 603 |
+
"loss": 0.6891459655761719,
|
| 604 |
+
"step": 4150
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 1.7551190973673214,
|
| 608 |
+
"grad_norm": 5.5546875,
|
| 609 |
+
"learning_rate": 4.416123295791346e-06,
|
| 610 |
+
"loss": 0.6910189056396484,
|
| 611 |
+
"step": 4200
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 1.77601337233598,
|
| 615 |
+
"grad_norm": 1.09765625,
|
| 616 |
+
"learning_rate": 4.342027267338471e-06,
|
| 617 |
+
"loss": 0.6903411865234375,
|
| 618 |
+
"step": 4250
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 1.7969076473046384,
|
| 622 |
+
"grad_norm": 1.7626953125,
|
| 623 |
+
"learning_rate": 4.267931238885596e-06,
|
| 624 |
+
"loss": 0.6911215209960937,
|
| 625 |
+
"step": 4300
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 1.817801922273297,
|
| 629 |
+
"grad_norm": 7.09765625,
|
| 630 |
+
"learning_rate": 4.193835210432721e-06,
|
| 631 |
+
"loss": 0.6897867584228515,
|
| 632 |
+
"step": 4350
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 1.8386961972419558,
|
| 636 |
+
"grad_norm": 2.671875,
|
| 637 |
+
"learning_rate": 4.119739181979846e-06,
|
| 638 |
+
"loss": 0.6900591278076171,
|
| 639 |
+
"step": 4400
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 1.8595904722106142,
|
| 643 |
+
"grad_norm": 2.34375,
|
| 644 |
+
"learning_rate": 4.045643153526971e-06,
|
| 645 |
+
"loss": 0.6915322875976563,
|
| 646 |
+
"step": 4450
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 1.8804847471792727,
|
| 650 |
+
"grad_norm": 1.58984375,
|
| 651 |
+
"learning_rate": 3.9715471250740965e-06,
|
| 652 |
+
"loss": 0.6882122039794922,
|
| 653 |
+
"step": 4500
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 1.9013790221479314,
|
| 657 |
+
"grad_norm": 1.25,
|
| 658 |
+
"learning_rate": 3.897451096621222e-06,
|
| 659 |
+
"loss": 0.6915691375732422,
|
| 660 |
+
"step": 4550
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 1.9222732971165901,
|
| 664 |
+
"grad_norm": 4.87109375,
|
| 665 |
+
"learning_rate": 3.823355068168346e-06,
|
| 666 |
+
"loss": 0.6912251281738281,
|
| 667 |
+
"step": 4600
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 1.9431675720852486,
|
| 671 |
+
"grad_norm": 6.890625,
|
| 672 |
+
"learning_rate": 3.7492590397154714e-06,
|
| 673 |
+
"loss": 0.6866284942626953,
|
| 674 |
+
"step": 4650
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 1.964061847053907,
|
| 678 |
+
"grad_norm": 3.189453125,
|
| 679 |
+
"learning_rate": 3.675163011262597e-06,
|
| 680 |
+
"loss": 0.6898875427246094,
|
| 681 |
+
"step": 4700
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 1.9849561220225658,
|
| 685 |
+
"grad_norm": 5.70703125,
|
| 686 |
+
"learning_rate": 3.6010669828097216e-06,
|
| 687 |
+
"loss": 0.6908906555175781,
|
| 688 |
+
"step": 4750
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 2.0,
|
| 692 |
+
"eval_accuracy": 0.5389196531187964,
|
| 693 |
+
"eval_f1": 0.7003869916491275,
|
| 694 |
+
"eval_loss": 0.6901159882545471,
|
| 695 |
+
"eval_precision": 0.5389196531187964,
|
| 696 |
+
"eval_recall": 1.0,
|
| 697 |
+
"eval_roc_auc": 0.5,
|
| 698 |
+
"eval_runtime": 347.1335,
|
| 699 |
+
"eval_samples_per_second": 27.572,
|
| 700 |
+
"eval_steps_per_second": 1.726,
|
| 701 |
+
"step": 4786
|
| 702 |
}
|
| 703 |
],
|
| 704 |
"logging_steps": 50,
|
|
|
|
| 713 |
"early_stopping_threshold": 0.0
|
| 714 |
},
|
| 715 |
"attributes": {
|
| 716 |
+
"early_stopping_patience_counter": 1
|
| 717 |
}
|
| 718 |
},
|
| 719 |
"TrainerControl": {
|
|
|
|
| 727 |
"attributes": {}
|
| 728 |
}
|
| 729 |
},
|
| 730 |
+
"total_flos": 1.3062259262328e+16,
|
| 731 |
"train_batch_size": 16,
|
| 732 |
"trial_name": null,
|
| 733 |
"trial_params": null
|