Training in progress, step 150, checkpoint
Browse files
last-checkpoint/adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 34456
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:16f0dc6e431f7b6b90f510918993168b5d4375476fd32863a49883b0971a02d8
|
| 3 |
size 34456
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 76154
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a3c868a8450cb66f75df4e2345a41bd57abf7584e688e0350757894ed84bab76
|
| 3 |
size 76154
|
last-checkpoint/rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14244
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:144921e75f42a1460d625a0de872fe0aecbbb880937260ed02766dacab7c2f59
|
| 3 |
size 14244
|
last-checkpoint/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1064
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f4c9c807f0681c8b7e53ada9b6ec3dba530d303de7da0d0a0562a3d8d0bbba08
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": 11.5,
|
| 3 |
"best_model_checkpoint": "miner_id_24/checkpoint-50",
|
| 4 |
-
"epoch": 0.
|
| 5 |
"eval_steps": 50,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -373,6 +373,722 @@
|
|
| 373 |
"eval_samples_per_second": 154.523,
|
| 374 |
"eval_steps_per_second": 77.261,
|
| 375 |
"step": 50
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 376 |
}
|
| 377 |
],
|
| 378 |
"logging_steps": 1,
|
|
@@ -387,7 +1103,7 @@
|
|
| 387 |
"early_stopping_threshold": 0.0
|
| 388 |
},
|
| 389 |
"attributes": {
|
| 390 |
-
"early_stopping_patience_counter":
|
| 391 |
}
|
| 392 |
},
|
| 393 |
"TrainerControl": {
|
|
@@ -401,7 +1117,7 @@
|
|
| 401 |
"attributes": {}
|
| 402 |
}
|
| 403 |
},
|
| 404 |
-
"total_flos":
|
| 405 |
"train_batch_size": 8,
|
| 406 |
"trial_name": null,
|
| 407 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": 11.5,
|
| 3 |
"best_model_checkpoint": "miner_id_24/checkpoint-50",
|
| 4 |
+
"epoch": 0.06737031214911296,
|
| 5 |
"eval_steps": 50,
|
| 6 |
+
"global_step": 150,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 373 |
"eval_samples_per_second": 154.523,
|
| 374 |
"eval_steps_per_second": 77.261,
|
| 375 |
"step": 50
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 0.022905906130698407,
|
| 379 |
+
"grad_norm": 1.5592015188303776e-05,
|
| 380 |
+
"learning_rate": 9.285836503510562e-05,
|
| 381 |
+
"loss": 46.0,
|
| 382 |
+
"step": 51
|
| 383 |
+
},
|
| 384 |
+
{
|
| 385 |
+
"epoch": 0.023355041545025827,
|
| 386 |
+
"grad_norm": 1.280045307794353e-05,
|
| 387 |
+
"learning_rate": 9.24024048078213e-05,
|
| 388 |
+
"loss": 46.0,
|
| 389 |
+
"step": 52
|
| 390 |
+
},
|
| 391 |
+
{
|
| 392 |
+
"epoch": 0.023804176959353246,
|
| 393 |
+
"grad_norm": 2.069097536150366e-05,
|
| 394 |
+
"learning_rate": 9.193352839727121e-05,
|
| 395 |
+
"loss": 46.0,
|
| 396 |
+
"step": 53
|
| 397 |
+
},
|
| 398 |
+
{
|
| 399 |
+
"epoch": 0.024253312373680665,
|
| 400 |
+
"grad_norm": 2.5162336896755733e-05,
|
| 401 |
+
"learning_rate": 9.145187862775209e-05,
|
| 402 |
+
"loss": 46.0,
|
| 403 |
+
"step": 54
|
| 404 |
+
},
|
| 405 |
+
{
|
| 406 |
+
"epoch": 0.024702447788008085,
|
| 407 |
+
"grad_norm": 3.217296762159094e-05,
|
| 408 |
+
"learning_rate": 9.09576022144496e-05,
|
| 409 |
+
"loss": 46.0,
|
| 410 |
+
"step": 55
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 0.025151583202335504,
|
| 414 |
+
"grad_norm": 3.4474760468583554e-05,
|
| 415 |
+
"learning_rate": 9.045084971874738e-05,
|
| 416 |
+
"loss": 46.0,
|
| 417 |
+
"step": 56
|
| 418 |
+
},
|
| 419 |
+
{
|
| 420 |
+
"epoch": 0.025600718616662924,
|
| 421 |
+
"grad_norm": 3.134864164167084e-05,
|
| 422 |
+
"learning_rate": 8.993177550236464e-05,
|
| 423 |
+
"loss": 46.0,
|
| 424 |
+
"step": 57
|
| 425 |
+
},
|
| 426 |
+
{
|
| 427 |
+
"epoch": 0.026049854030990343,
|
| 428 |
+
"grad_norm": 2.7241056159255095e-05,
|
| 429 |
+
"learning_rate": 8.940053768033609e-05,
|
| 430 |
+
"loss": 46.0,
|
| 431 |
+
"step": 58
|
| 432 |
+
},
|
| 433 |
+
{
|
| 434 |
+
"epoch": 0.026498989445317762,
|
| 435 |
+
"grad_norm": 3.1343828595709056e-05,
|
| 436 |
+
"learning_rate": 8.885729807284856e-05,
|
| 437 |
+
"loss": 46.0,
|
| 438 |
+
"step": 59
|
| 439 |
+
},
|
| 440 |
+
{
|
| 441 |
+
"epoch": 0.026948124859645182,
|
| 442 |
+
"grad_norm": 2.3026888811727986e-05,
|
| 443 |
+
"learning_rate": 8.83022221559489e-05,
|
| 444 |
+
"loss": 46.0,
|
| 445 |
+
"step": 60
|
| 446 |
+
},
|
| 447 |
+
{
|
| 448 |
+
"epoch": 0.0273972602739726,
|
| 449 |
+
"grad_norm": 4.045252353535034e-05,
|
| 450 |
+
"learning_rate": 8.773547901113862e-05,
|
| 451 |
+
"loss": 46.0,
|
| 452 |
+
"step": 61
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 0.027846395688300024,
|
| 456 |
+
"grad_norm": 5.758315455750562e-05,
|
| 457 |
+
"learning_rate": 8.715724127386972e-05,
|
| 458 |
+
"loss": 46.0,
|
| 459 |
+
"step": 62
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 0.028295531102627444,
|
| 463 |
+
"grad_norm": 2.3458678697352298e-05,
|
| 464 |
+
"learning_rate": 8.656768508095853e-05,
|
| 465 |
+
"loss": 46.0,
|
| 466 |
+
"step": 63
|
| 467 |
+
},
|
| 468 |
+
{
|
| 469 |
+
"epoch": 0.028744666516954863,
|
| 470 |
+
"grad_norm": 2.0376308384584263e-05,
|
| 471 |
+
"learning_rate": 8.596699001693255e-05,
|
| 472 |
+
"loss": 46.0,
|
| 473 |
+
"step": 64
|
| 474 |
+
},
|
| 475 |
+
{
|
| 476 |
+
"epoch": 0.029193801931282282,
|
| 477 |
+
"grad_norm": 3.670160367619246e-05,
|
| 478 |
+
"learning_rate": 8.535533905932738e-05,
|
| 479 |
+
"loss": 46.0,
|
| 480 |
+
"step": 65
|
| 481 |
+
},
|
| 482 |
+
{
|
| 483 |
+
"epoch": 0.0296429373456097,
|
| 484 |
+
"grad_norm": 3.101183392573148e-05,
|
| 485 |
+
"learning_rate": 8.473291852294987e-05,
|
| 486 |
+
"loss": 46.0,
|
| 487 |
+
"step": 66
|
| 488 |
+
},
|
| 489 |
+
{
|
| 490 |
+
"epoch": 0.03009207275993712,
|
| 491 |
+
"grad_norm": 3.032161112059839e-05,
|
| 492 |
+
"learning_rate": 8.409991800312493e-05,
|
| 493 |
+
"loss": 46.0,
|
| 494 |
+
"step": 67
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 0.03054120817426454,
|
| 498 |
+
"grad_norm": 3.268122236477211e-05,
|
| 499 |
+
"learning_rate": 8.345653031794292e-05,
|
| 500 |
+
"loss": 46.0,
|
| 501 |
+
"step": 68
|
| 502 |
+
},
|
| 503 |
+
{
|
| 504 |
+
"epoch": 0.03099034358859196,
|
| 505 |
+
"grad_norm": 3.1394203688250855e-05,
|
| 506 |
+
"learning_rate": 8.280295144952536e-05,
|
| 507 |
+
"loss": 46.0,
|
| 508 |
+
"step": 69
|
| 509 |
+
},
|
| 510 |
+
{
|
| 511 |
+
"epoch": 0.03143947900291938,
|
| 512 |
+
"grad_norm": 3.980135443271138e-05,
|
| 513 |
+
"learning_rate": 8.213938048432697e-05,
|
| 514 |
+
"loss": 46.0,
|
| 515 |
+
"step": 70
|
| 516 |
+
},
|
| 517 |
+
{
|
| 518 |
+
"epoch": 0.0318886144172468,
|
| 519 |
+
"grad_norm": 3.355582521180622e-05,
|
| 520 |
+
"learning_rate": 8.146601955249188e-05,
|
| 521 |
+
"loss": 46.0,
|
| 522 |
+
"step": 71
|
| 523 |
+
},
|
| 524 |
+
{
|
| 525 |
+
"epoch": 0.03233774983157422,
|
| 526 |
+
"grad_norm": 4.067844929522835e-05,
|
| 527 |
+
"learning_rate": 8.07830737662829e-05,
|
| 528 |
+
"loss": 46.0,
|
| 529 |
+
"step": 72
|
| 530 |
+
},
|
| 531 |
+
{
|
| 532 |
+
"epoch": 0.03278688524590164,
|
| 533 |
+
"grad_norm": 5.312855864758603e-05,
|
| 534 |
+
"learning_rate": 8.009075115760243e-05,
|
| 535 |
+
"loss": 46.0,
|
| 536 |
+
"step": 73
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 0.03323602066022906,
|
| 540 |
+
"grad_norm": 2.6573492505121976e-05,
|
| 541 |
+
"learning_rate": 7.938926261462366e-05,
|
| 542 |
+
"loss": 46.0,
|
| 543 |
+
"step": 74
|
| 544 |
+
},
|
| 545 |
+
{
|
| 546 |
+
"epoch": 0.03368515607455648,
|
| 547 |
+
"grad_norm": 6.407412729458883e-05,
|
| 548 |
+
"learning_rate": 7.86788218175523e-05,
|
| 549 |
+
"loss": 46.0,
|
| 550 |
+
"step": 75
|
| 551 |
+
},
|
| 552 |
+
{
|
| 553 |
+
"epoch": 0.034134291488883896,
|
| 554 |
+
"grad_norm": 4.746004560729489e-05,
|
| 555 |
+
"learning_rate": 7.795964517353735e-05,
|
| 556 |
+
"loss": 46.0,
|
| 557 |
+
"step": 76
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"epoch": 0.03458342690321132,
|
| 561 |
+
"grad_norm": 3.733126141014509e-05,
|
| 562 |
+
"learning_rate": 7.723195175075136e-05,
|
| 563 |
+
"loss": 46.0,
|
| 564 |
+
"step": 77
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"epoch": 0.035032562317538735,
|
| 568 |
+
"grad_norm": 8.151544898282737e-05,
|
| 569 |
+
"learning_rate": 7.649596321166024e-05,
|
| 570 |
+
"loss": 46.0,
|
| 571 |
+
"step": 78
|
| 572 |
+
},
|
| 573 |
+
{
|
| 574 |
+
"epoch": 0.03548169773186616,
|
| 575 |
+
"grad_norm": 2.8419528462109156e-05,
|
| 576 |
+
"learning_rate": 7.575190374550272e-05,
|
| 577 |
+
"loss": 46.0,
|
| 578 |
+
"step": 79
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 0.03593083314619358,
|
| 582 |
+
"grad_norm": 3.4233249607495964e-05,
|
| 583 |
+
"learning_rate": 7.500000000000001e-05,
|
| 584 |
+
"loss": 46.0,
|
| 585 |
+
"step": 80
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 0.036379968560520996,
|
| 589 |
+
"grad_norm": 5.3122937970329076e-05,
|
| 590 |
+
"learning_rate": 7.424048101231686e-05,
|
| 591 |
+
"loss": 46.0,
|
| 592 |
+
"step": 81
|
| 593 |
+
},
|
| 594 |
+
{
|
| 595 |
+
"epoch": 0.03682910397484842,
|
| 596 |
+
"grad_norm": 5.805513501400128e-05,
|
| 597 |
+
"learning_rate": 7.347357813929454e-05,
|
| 598 |
+
"loss": 46.0,
|
| 599 |
+
"step": 82
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 0.037278239389175835,
|
| 603 |
+
"grad_norm": 2.634433076309506e-05,
|
| 604 |
+
"learning_rate": 7.269952498697734e-05,
|
| 605 |
+
"loss": 46.0,
|
| 606 |
+
"step": 83
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"epoch": 0.03772737480350326,
|
| 610 |
+
"grad_norm": 5.438110383693129e-05,
|
| 611 |
+
"learning_rate": 7.191855733945387e-05,
|
| 612 |
+
"loss": 46.0,
|
| 613 |
+
"step": 84
|
| 614 |
+
},
|
| 615 |
+
{
|
| 616 |
+
"epoch": 0.038176510217830674,
|
| 617 |
+
"grad_norm": 5.9794343542307615e-05,
|
| 618 |
+
"learning_rate": 7.113091308703498e-05,
|
| 619 |
+
"loss": 46.0,
|
| 620 |
+
"step": 85
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 0.0386256456321581,
|
| 624 |
+
"grad_norm": 8.437032374786213e-05,
|
| 625 |
+
"learning_rate": 7.033683215379002e-05,
|
| 626 |
+
"loss": 46.0,
|
| 627 |
+
"step": 86
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 0.03907478104648551,
|
| 631 |
+
"grad_norm": 9.161300113191828e-05,
|
| 632 |
+
"learning_rate": 6.953655642446368e-05,
|
| 633 |
+
"loss": 46.0,
|
| 634 |
+
"step": 87
|
| 635 |
+
},
|
| 636 |
+
{
|
| 637 |
+
"epoch": 0.039523916460812936,
|
| 638 |
+
"grad_norm": 9.054856491275132e-05,
|
| 639 |
+
"learning_rate": 6.873032967079561e-05,
|
| 640 |
+
"loss": 46.0,
|
| 641 |
+
"step": 88
|
| 642 |
+
},
|
| 643 |
+
{
|
| 644 |
+
"epoch": 0.03997305187514035,
|
| 645 |
+
"grad_norm": 7.056965841911733e-05,
|
| 646 |
+
"learning_rate": 6.7918397477265e-05,
|
| 647 |
+
"loss": 46.0,
|
| 648 |
+
"step": 89
|
| 649 |
+
},
|
| 650 |
+
{
|
| 651 |
+
"epoch": 0.040422187289467774,
|
| 652 |
+
"grad_norm": 5.0735634431475773e-05,
|
| 653 |
+
"learning_rate": 6.710100716628344e-05,
|
| 654 |
+
"loss": 46.0,
|
| 655 |
+
"step": 90
|
| 656 |
+
},
|
| 657 |
+
{
|
| 658 |
+
"epoch": 0.0408713227037952,
|
| 659 |
+
"grad_norm": 4.1233262891182676e-05,
|
| 660 |
+
"learning_rate": 6.627840772285784e-05,
|
| 661 |
+
"loss": 46.0,
|
| 662 |
+
"step": 91
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 0.04132045811812261,
|
| 666 |
+
"grad_norm": 7.487056427635252e-05,
|
| 667 |
+
"learning_rate": 6.545084971874738e-05,
|
| 668 |
+
"loss": 46.0,
|
| 669 |
+
"step": 92
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"epoch": 0.041769593532450036,
|
| 673 |
+
"grad_norm": 5.445545684779063e-05,
|
| 674 |
+
"learning_rate": 6.461858523613684e-05,
|
| 675 |
+
"loss": 46.0,
|
| 676 |
+
"step": 93
|
| 677 |
+
},
|
| 678 |
+
{
|
| 679 |
+
"epoch": 0.04221872894677745,
|
| 680 |
+
"grad_norm": 7.229465700220317e-05,
|
| 681 |
+
"learning_rate": 6.378186779084995e-05,
|
| 682 |
+
"loss": 46.0,
|
| 683 |
+
"step": 94
|
| 684 |
+
},
|
| 685 |
+
{
|
| 686 |
+
"epoch": 0.042667864361104875,
|
| 687 |
+
"grad_norm": 6.044282781658694e-05,
|
| 688 |
+
"learning_rate": 6.294095225512603e-05,
|
| 689 |
+
"loss": 46.0,
|
| 690 |
+
"step": 95
|
| 691 |
+
},
|
| 692 |
+
{
|
| 693 |
+
"epoch": 0.04311699977543229,
|
| 694 |
+
"grad_norm": 7.44121425668709e-05,
|
| 695 |
+
"learning_rate": 6.209609477998338e-05,
|
| 696 |
+
"loss": 46.0,
|
| 697 |
+
"step": 96
|
| 698 |
+
},
|
| 699 |
+
{
|
| 700 |
+
"epoch": 0.043566135189759714,
|
| 701 |
+
"grad_norm": 6.77663556416519e-05,
|
| 702 |
+
"learning_rate": 6.124755271719325e-05,
|
| 703 |
+
"loss": 46.0,
|
| 704 |
+
"step": 97
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 0.04401527060408713,
|
| 708 |
+
"grad_norm": 4.348663060227409e-05,
|
| 709 |
+
"learning_rate": 6.0395584540887963e-05,
|
| 710 |
+
"loss": 46.0,
|
| 711 |
+
"step": 98
|
| 712 |
+
},
|
| 713 |
+
{
|
| 714 |
+
"epoch": 0.04446440601841455,
|
| 715 |
+
"grad_norm": 9.123215568251908e-05,
|
| 716 |
+
"learning_rate": 5.9540449768827246e-05,
|
| 717 |
+
"loss": 46.0,
|
| 718 |
+
"step": 99
|
| 719 |
+
},
|
| 720 |
+
{
|
| 721 |
+
"epoch": 0.04491354143274197,
|
| 722 |
+
"grad_norm": 5.847312422702089e-05,
|
| 723 |
+
"learning_rate": 5.868240888334653e-05,
|
| 724 |
+
"loss": 46.0,
|
| 725 |
+
"step": 100
|
| 726 |
+
},
|
| 727 |
+
{
|
| 728 |
+
"epoch": 0.04491354143274197,
|
| 729 |
+
"eval_loss": 11.5,
|
| 730 |
+
"eval_runtime": 24.4817,
|
| 731 |
+
"eval_samples_per_second": 153.176,
|
| 732 |
+
"eval_steps_per_second": 76.588,
|
| 733 |
+
"step": 100
|
| 734 |
+
},
|
| 735 |
+
{
|
| 736 |
+
"epoch": 0.04536267684706939,
|
| 737 |
+
"grad_norm": 6.564732029801235e-05,
|
| 738 |
+
"learning_rate": 5.782172325201155e-05,
|
| 739 |
+
"loss": 46.0,
|
| 740 |
+
"step": 101
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 0.045811812261396814,
|
| 744 |
+
"grad_norm": 9.073306864593178e-05,
|
| 745 |
+
"learning_rate": 5.695865504800327e-05,
|
| 746 |
+
"loss": 46.0,
|
| 747 |
+
"step": 102
|
| 748 |
+
},
|
| 749 |
+
{
|
| 750 |
+
"epoch": 0.04626094767572423,
|
| 751 |
+
"grad_norm": 5.6581779062980786e-05,
|
| 752 |
+
"learning_rate": 5.6093467170257374e-05,
|
| 753 |
+
"loss": 46.0,
|
| 754 |
+
"step": 103
|
| 755 |
+
},
|
| 756 |
+
{
|
| 757 |
+
"epoch": 0.04671008309005165,
|
| 758 |
+
"grad_norm": 4.856206942349672e-05,
|
| 759 |
+
"learning_rate": 5.522642316338268e-05,
|
| 760 |
+
"loss": 46.0,
|
| 761 |
+
"step": 104
|
| 762 |
+
},
|
| 763 |
+
{
|
| 764 |
+
"epoch": 0.04715921850437907,
|
| 765 |
+
"grad_norm": 0.00014515251677948982,
|
| 766 |
+
"learning_rate": 5.435778713738292e-05,
|
| 767 |
+
"loss": 46.0,
|
| 768 |
+
"step": 105
|
| 769 |
+
},
|
| 770 |
+
{
|
| 771 |
+
"epoch": 0.04760835391870649,
|
| 772 |
+
"grad_norm": 4.484998862608336e-05,
|
| 773 |
+
"learning_rate": 5.348782368720626e-05,
|
| 774 |
+
"loss": 46.0,
|
| 775 |
+
"step": 106
|
| 776 |
+
},
|
| 777 |
+
{
|
| 778 |
+
"epoch": 0.04805748933303391,
|
| 779 |
+
"grad_norm": 5.431281897472218e-05,
|
| 780 |
+
"learning_rate": 5.26167978121472e-05,
|
| 781 |
+
"loss": 46.0,
|
| 782 |
+
"step": 107
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 0.04850662474736133,
|
| 786 |
+
"grad_norm": 5.2316147048259154e-05,
|
| 787 |
+
"learning_rate": 5.174497483512506e-05,
|
| 788 |
+
"loss": 46.0,
|
| 789 |
+
"step": 108
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"epoch": 0.04895576016168875,
|
| 793 |
+
"grad_norm": 0.00020991968631278723,
|
| 794 |
+
"learning_rate": 5.0872620321864185e-05,
|
| 795 |
+
"loss": 46.0,
|
| 796 |
+
"step": 109
|
| 797 |
+
},
|
| 798 |
+
{
|
| 799 |
+
"epoch": 0.04940489557601617,
|
| 800 |
+
"grad_norm": 0.0001119155203923583,
|
| 801 |
+
"learning_rate": 5e-05,
|
| 802 |
+
"loss": 46.0,
|
| 803 |
+
"step": 110
|
| 804 |
+
},
|
| 805 |
+
{
|
| 806 |
+
"epoch": 0.049854030990343585,
|
| 807 |
+
"grad_norm": 3.5669108910951763e-05,
|
| 808 |
+
"learning_rate": 4.912737967813583e-05,
|
| 809 |
+
"loss": 46.0,
|
| 810 |
+
"step": 111
|
| 811 |
+
},
|
| 812 |
+
{
|
| 813 |
+
"epoch": 0.05030316640467101,
|
| 814 |
+
"grad_norm": 8.696071745362133e-05,
|
| 815 |
+
"learning_rate": 4.825502516487497e-05,
|
| 816 |
+
"loss": 46.0,
|
| 817 |
+
"step": 112
|
| 818 |
+
},
|
| 819 |
+
{
|
| 820 |
+
"epoch": 0.05075230181899843,
|
| 821 |
+
"grad_norm": 4.1281185986008495e-05,
|
| 822 |
+
"learning_rate": 4.738320218785281e-05,
|
| 823 |
+
"loss": 46.0,
|
| 824 |
+
"step": 113
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 0.05120143723332585,
|
| 828 |
+
"grad_norm": 7.59501926950179e-05,
|
| 829 |
+
"learning_rate": 4.6512176312793736e-05,
|
| 830 |
+
"loss": 46.0,
|
| 831 |
+
"step": 114
|
| 832 |
+
},
|
| 833 |
+
{
|
| 834 |
+
"epoch": 0.05165057264765327,
|
| 835 |
+
"grad_norm": 4.041286592837423e-05,
|
| 836 |
+
"learning_rate": 4.564221286261709e-05,
|
| 837 |
+
"loss": 46.0,
|
| 838 |
+
"step": 115
|
| 839 |
+
},
|
| 840 |
+
{
|
| 841 |
+
"epoch": 0.052099708061980686,
|
| 842 |
+
"grad_norm": 0.00012405925372149795,
|
| 843 |
+
"learning_rate": 4.477357683661734e-05,
|
| 844 |
+
"loss": 46.0,
|
| 845 |
+
"step": 116
|
| 846 |
+
},
|
| 847 |
+
{
|
| 848 |
+
"epoch": 0.05254884347630811,
|
| 849 |
+
"grad_norm": 4.541510497801937e-05,
|
| 850 |
+
"learning_rate": 4.390653282974264e-05,
|
| 851 |
+
"loss": 46.0,
|
| 852 |
+
"step": 117
|
| 853 |
+
},
|
| 854 |
+
{
|
| 855 |
+
"epoch": 0.052997978890635525,
|
| 856 |
+
"grad_norm": 3.6235865991329774e-05,
|
| 857 |
+
"learning_rate": 4.3041344951996746e-05,
|
| 858 |
+
"loss": 46.0,
|
| 859 |
+
"step": 118
|
| 860 |
+
},
|
| 861 |
+
{
|
| 862 |
+
"epoch": 0.05344711430496295,
|
| 863 |
+
"grad_norm": 0.0001397922169417143,
|
| 864 |
+
"learning_rate": 4.2178276747988446e-05,
|
| 865 |
+
"loss": 46.0,
|
| 866 |
+
"step": 119
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 0.053896249719290364,
|
| 870 |
+
"grad_norm": 4.858409010921605e-05,
|
| 871 |
+
"learning_rate": 4.131759111665349e-05,
|
| 872 |
+
"loss": 46.0,
|
| 873 |
+
"step": 120
|
| 874 |
+
},
|
| 875 |
+
{
|
| 876 |
+
"epoch": 0.054345385133617786,
|
| 877 |
+
"grad_norm": 6.559174653375521e-05,
|
| 878 |
+
"learning_rate": 4.045955023117276e-05,
|
| 879 |
+
"loss": 46.0,
|
| 880 |
+
"step": 121
|
| 881 |
+
},
|
| 882 |
+
{
|
| 883 |
+
"epoch": 0.0547945205479452,
|
| 884 |
+
"grad_norm": 9.03177133295685e-05,
|
| 885 |
+
"learning_rate": 3.960441545911204e-05,
|
| 886 |
+
"loss": 46.0,
|
| 887 |
+
"step": 122
|
| 888 |
+
},
|
| 889 |
+
{
|
| 890 |
+
"epoch": 0.055243655962272625,
|
| 891 |
+
"grad_norm": 7.547564746346325e-05,
|
| 892 |
+
"learning_rate": 3.875244728280676e-05,
|
| 893 |
+
"loss": 46.0,
|
| 894 |
+
"step": 123
|
| 895 |
+
},
|
| 896 |
+
{
|
| 897 |
+
"epoch": 0.05569279137660005,
|
| 898 |
+
"grad_norm": 8.072968194028363e-05,
|
| 899 |
+
"learning_rate": 3.790390522001662e-05,
|
| 900 |
+
"loss": 46.0,
|
| 901 |
+
"step": 124
|
| 902 |
+
},
|
| 903 |
+
{
|
| 904 |
+
"epoch": 0.056141926790927464,
|
| 905 |
+
"grad_norm": 0.00010309958452126011,
|
| 906 |
+
"learning_rate": 3.705904774487396e-05,
|
| 907 |
+
"loss": 46.0,
|
| 908 |
+
"step": 125
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 0.05659106220525489,
|
| 912 |
+
"grad_norm": 8.730621630093083e-05,
|
| 913 |
+
"learning_rate": 3.6218132209150045e-05,
|
| 914 |
+
"loss": 46.0,
|
| 915 |
+
"step": 126
|
| 916 |
+
},
|
| 917 |
+
{
|
| 918 |
+
"epoch": 0.0570401976195823,
|
| 919 |
+
"grad_norm": 7.893842848716304e-05,
|
| 920 |
+
"learning_rate": 3.5381414763863166e-05,
|
| 921 |
+
"loss": 46.0,
|
| 922 |
+
"step": 127
|
| 923 |
+
},
|
| 924 |
+
{
|
| 925 |
+
"epoch": 0.057489333033909726,
|
| 926 |
+
"grad_norm": 7.924601959530264e-05,
|
| 927 |
+
"learning_rate": 3.4549150281252636e-05,
|
| 928 |
+
"loss": 46.0,
|
| 929 |
+
"step": 128
|
| 930 |
+
},
|
| 931 |
+
{
|
| 932 |
+
"epoch": 0.05793846844823714,
|
| 933 |
+
"grad_norm": 9.8539145255927e-05,
|
| 934 |
+
"learning_rate": 3.372159227714218e-05,
|
| 935 |
+
"loss": 46.0,
|
| 936 |
+
"step": 129
|
| 937 |
+
},
|
| 938 |
+
{
|
| 939 |
+
"epoch": 0.058387603862564565,
|
| 940 |
+
"grad_norm": 0.00011634702241281047,
|
| 941 |
+
"learning_rate": 3.289899283371657e-05,
|
| 942 |
+
"loss": 46.0,
|
| 943 |
+
"step": 130
|
| 944 |
+
},
|
| 945 |
+
{
|
| 946 |
+
"epoch": 0.05883673927689198,
|
| 947 |
+
"grad_norm": 9.868943743640557e-05,
|
| 948 |
+
"learning_rate": 3.2081602522734986e-05,
|
| 949 |
+
"loss": 46.0,
|
| 950 |
+
"step": 131
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 0.0592858746912194,
|
| 954 |
+
"grad_norm": 0.0001640073605813086,
|
| 955 |
+
"learning_rate": 3.12696703292044e-05,
|
| 956 |
+
"loss": 46.0,
|
| 957 |
+
"step": 132
|
| 958 |
+
},
|
| 959 |
+
{
|
| 960 |
+
"epoch": 0.05973501010554682,
|
| 961 |
+
"grad_norm": 0.00012898287968710065,
|
| 962 |
+
"learning_rate": 3.046344357553632e-05,
|
| 963 |
+
"loss": 46.0,
|
| 964 |
+
"step": 133
|
| 965 |
+
},
|
| 966 |
+
{
|
| 967 |
+
"epoch": 0.06018414551987424,
|
| 968 |
+
"grad_norm": 8.394714677706361e-05,
|
| 969 |
+
"learning_rate": 2.9663167846209998e-05,
|
| 970 |
+
"loss": 46.0,
|
| 971 |
+
"step": 134
|
| 972 |
+
},
|
| 973 |
+
{
|
| 974 |
+
"epoch": 0.060633280934201665,
|
| 975 |
+
"grad_norm": 0.00014036337961442769,
|
| 976 |
+
"learning_rate": 2.886908691296504e-05,
|
| 977 |
+
"loss": 46.0,
|
| 978 |
+
"step": 135
|
| 979 |
+
},
|
| 980 |
+
{
|
| 981 |
+
"epoch": 0.06108241634852908,
|
| 982 |
+
"grad_norm": 8.231115498347208e-05,
|
| 983 |
+
"learning_rate": 2.8081442660546125e-05,
|
| 984 |
+
"loss": 46.0,
|
| 985 |
+
"step": 136
|
| 986 |
+
},
|
| 987 |
+
{
|
| 988 |
+
"epoch": 0.061531551762856504,
|
| 989 |
+
"grad_norm": 0.00014053141057956964,
|
| 990 |
+
"learning_rate": 2.7300475013022663e-05,
|
| 991 |
+
"loss": 46.0,
|
| 992 |
+
"step": 137
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 0.06198068717718392,
|
| 996 |
+
"grad_norm": 9.784067515283823e-05,
|
| 997 |
+
"learning_rate": 2.6526421860705473e-05,
|
| 998 |
+
"loss": 46.0,
|
| 999 |
+
"step": 138
|
| 1000 |
+
},
|
| 1001 |
+
{
|
| 1002 |
+
"epoch": 0.06242982259151134,
|
| 1003 |
+
"grad_norm": 9.869643690763041e-05,
|
| 1004 |
+
"learning_rate": 2.575951898768315e-05,
|
| 1005 |
+
"loss": 46.0,
|
| 1006 |
+
"step": 139
|
| 1007 |
+
},
|
| 1008 |
+
{
|
| 1009 |
+
"epoch": 0.06287895800583876,
|
| 1010 |
+
"grad_norm": 7.148009171942249e-05,
|
| 1011 |
+
"learning_rate": 2.500000000000001e-05,
|
| 1012 |
+
"loss": 46.0,
|
| 1013 |
+
"step": 140
|
| 1014 |
+
},
|
| 1015 |
+
{
|
| 1016 |
+
"epoch": 0.06332809342016618,
|
| 1017 |
+
"grad_norm": 0.00010294223466189578,
|
| 1018 |
+
"learning_rate": 2.4248096254497288e-05,
|
| 1019 |
+
"loss": 46.0,
|
| 1020 |
+
"step": 141
|
| 1021 |
+
},
|
| 1022 |
+
{
|
| 1023 |
+
"epoch": 0.0637772288344936,
|
| 1024 |
+
"grad_norm": 9.365833102492616e-05,
|
| 1025 |
+
"learning_rate": 2.350403678833976e-05,
|
| 1026 |
+
"loss": 46.0,
|
| 1027 |
+
"step": 142
|
| 1028 |
+
},
|
| 1029 |
+
{
|
| 1030 |
+
"epoch": 0.06422636424882101,
|
| 1031 |
+
"grad_norm": 6.49421926937066e-05,
|
| 1032 |
+
"learning_rate": 2.2768048249248648e-05,
|
| 1033 |
+
"loss": 46.0,
|
| 1034 |
+
"step": 143
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 0.06467549966314844,
|
| 1038 |
+
"grad_norm": 0.0001447576651116833,
|
| 1039 |
+
"learning_rate": 2.2040354826462668e-05,
|
| 1040 |
+
"loss": 46.0,
|
| 1041 |
+
"step": 144
|
| 1042 |
+
},
|
| 1043 |
+
{
|
| 1044 |
+
"epoch": 0.06512463507747586,
|
| 1045 |
+
"grad_norm": 8.269389218185097e-05,
|
| 1046 |
+
"learning_rate": 2.132117818244771e-05,
|
| 1047 |
+
"loss": 46.0,
|
| 1048 |
+
"step": 145
|
| 1049 |
+
},
|
| 1050 |
+
{
|
| 1051 |
+
"epoch": 0.06557377049180328,
|
| 1052 |
+
"grad_norm": 8.899991371436045e-05,
|
| 1053 |
+
"learning_rate": 2.061073738537635e-05,
|
| 1054 |
+
"loss": 46.0,
|
| 1055 |
+
"step": 146
|
| 1056 |
+
},
|
| 1057 |
+
{
|
| 1058 |
+
"epoch": 0.0660229059061307,
|
| 1059 |
+
"grad_norm": 6.125797517597675e-05,
|
| 1060 |
+
"learning_rate": 1.9909248842397584e-05,
|
| 1061 |
+
"loss": 46.0,
|
| 1062 |
+
"step": 147
|
| 1063 |
+
},
|
| 1064 |
+
{
|
| 1065 |
+
"epoch": 0.06647204132045811,
|
| 1066 |
+
"grad_norm": 0.00010218821262242272,
|
| 1067 |
+
"learning_rate": 1.9216926233717085e-05,
|
| 1068 |
+
"loss": 46.0,
|
| 1069 |
+
"step": 148
|
| 1070 |
+
},
|
| 1071 |
+
{
|
| 1072 |
+
"epoch": 0.06692117673478554,
|
| 1073 |
+
"grad_norm": 5.742968642152846e-05,
|
| 1074 |
+
"learning_rate": 1.8533980447508137e-05,
|
| 1075 |
+
"loss": 46.0,
|
| 1076 |
+
"step": 149
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 0.06737031214911296,
|
| 1080 |
+
"grad_norm": 0.0001050592545652762,
|
| 1081 |
+
"learning_rate": 1.7860619515673033e-05,
|
| 1082 |
+
"loss": 46.0,
|
| 1083 |
+
"step": 150
|
| 1084 |
+
},
|
| 1085 |
+
{
|
| 1086 |
+
"epoch": 0.06737031214911296,
|
| 1087 |
+
"eval_loss": 11.5,
|
| 1088 |
+
"eval_runtime": 24.4902,
|
| 1089 |
+
"eval_samples_per_second": 153.122,
|
| 1090 |
+
"eval_steps_per_second": 76.561,
|
| 1091 |
+
"step": 150
|
| 1092 |
}
|
| 1093 |
],
|
| 1094 |
"logging_steps": 1,
|
|
|
|
| 1103 |
"early_stopping_threshold": 0.0
|
| 1104 |
},
|
| 1105 |
"attributes": {
|
| 1106 |
+
"early_stopping_patience_counter": 2
|
| 1107 |
}
|
| 1108 |
},
|
| 1109 |
"TrainerControl": {
|
|
|
|
| 1117 |
"attributes": {}
|
| 1118 |
}
|
| 1119 |
},
|
| 1120 |
+
"total_flos": 24319837470720.0,
|
| 1121 |
"train_batch_size": 8,
|
| 1122 |
"trial_name": null,
|
| 1123 |
"trial_params": null
|