Commit
•
028d774
1
Parent(s):
4cb5124
Training in progress, step 553, 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 100198584
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7c7efe866276ae7cbe3d104678cd00bf3527bbd6e30a9a945a5ebecfee27e759
|
| 3 |
size 100198584
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 50675604
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:32b9dc35758135f884494cc0e53f601a5e3e91671cbbf46618714b389006b98e
|
| 3 |
size 50675604
|
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 sha256:6e166349555ca5e7760bd934ba6aeb783f7b623356bfc480bba3fd1729b099ba
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch": 0.
|
| 5 |
"eval_steps": 500,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -3332,6 +3332,552 @@
|
|
| 3332 |
"learning_rate": 7.1575342465753425e-06,
|
| 3333 |
"loss": 1.2028,
|
| 3334 |
"step": 475
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3335 |
}
|
| 3336 |
],
|
| 3337 |
"logging_steps": 1,
|
|
@@ -3351,7 +3897,7 @@
|
|
| 3351 |
"attributes": {}
|
| 3352 |
}
|
| 3353 |
},
|
| 3354 |
-
"total_flos":
|
| 3355 |
"train_batch_size": 4,
|
| 3356 |
"trial_name": null,
|
| 3357 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.8079627431284814,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 553,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 3332 |
"learning_rate": 7.1575342465753425e-06,
|
| 3333 |
"loss": 1.2028,
|
| 3334 |
"step": 475
|
| 3335 |
+
},
|
| 3336 |
+
{
|
| 3337 |
+
"epoch": 0.6954616016802119,
|
| 3338 |
+
"grad_norm": 0.12074082344770432,
|
| 3339 |
+
"learning_rate": 7.123287671232877e-06,
|
| 3340 |
+
"loss": 1.3148,
|
| 3341 |
+
"step": 476
|
| 3342 |
+
},
|
| 3343 |
+
{
|
| 3344 |
+
"epoch": 0.6969226554652543,
|
| 3345 |
+
"grad_norm": 0.11308849602937698,
|
| 3346 |
+
"learning_rate": 7.089041095890411e-06,
|
| 3347 |
+
"loss": 1.2337,
|
| 3348 |
+
"step": 477
|
| 3349 |
+
},
|
| 3350 |
+
{
|
| 3351 |
+
"epoch": 0.6983837092502968,
|
| 3352 |
+
"grad_norm": 0.11705081909894943,
|
| 3353 |
+
"learning_rate": 7.054794520547946e-06,
|
| 3354 |
+
"loss": 1.1486,
|
| 3355 |
+
"step": 478
|
| 3356 |
+
},
|
| 3357 |
+
{
|
| 3358 |
+
"epoch": 0.6998447630353393,
|
| 3359 |
+
"grad_norm": 0.11467244476079941,
|
| 3360 |
+
"learning_rate": 7.020547945205481e-06,
|
| 3361 |
+
"loss": 1.3139,
|
| 3362 |
+
"step": 479
|
| 3363 |
+
},
|
| 3364 |
+
{
|
| 3365 |
+
"epoch": 0.7013058168203817,
|
| 3366 |
+
"grad_norm": 0.11780110001564026,
|
| 3367 |
+
"learning_rate": 6.9863013698630145e-06,
|
| 3368 |
+
"loss": 1.2417,
|
| 3369 |
+
"step": 480
|
| 3370 |
+
},
|
| 3371 |
+
{
|
| 3372 |
+
"epoch": 0.7027668706054242,
|
| 3373 |
+
"grad_norm": 0.11144915223121643,
|
| 3374 |
+
"learning_rate": 6.952054794520548e-06,
|
| 3375 |
+
"loss": 1.1892,
|
| 3376 |
+
"step": 481
|
| 3377 |
+
},
|
| 3378 |
+
{
|
| 3379 |
+
"epoch": 0.7042279243904667,
|
| 3380 |
+
"grad_norm": 0.13650654256343842,
|
| 3381 |
+
"learning_rate": 6.917808219178082e-06,
|
| 3382 |
+
"loss": 1.0756,
|
| 3383 |
+
"step": 482
|
| 3384 |
+
},
|
| 3385 |
+
{
|
| 3386 |
+
"epoch": 0.7056889781755091,
|
| 3387 |
+
"grad_norm": 0.11797624081373215,
|
| 3388 |
+
"learning_rate": 6.883561643835617e-06,
|
| 3389 |
+
"loss": 1.275,
|
| 3390 |
+
"step": 483
|
| 3391 |
+
},
|
| 3392 |
+
{
|
| 3393 |
+
"epoch": 0.7071500319605516,
|
| 3394 |
+
"grad_norm": 0.1288430392742157,
|
| 3395 |
+
"learning_rate": 6.849315068493151e-06,
|
| 3396 |
+
"loss": 1.1719,
|
| 3397 |
+
"step": 484
|
| 3398 |
+
},
|
| 3399 |
+
{
|
| 3400 |
+
"epoch": 0.7086110857455941,
|
| 3401 |
+
"grad_norm": 0.11288215219974518,
|
| 3402 |
+
"learning_rate": 6.815068493150685e-06,
|
| 3403 |
+
"loss": 1.1841,
|
| 3404 |
+
"step": 485
|
| 3405 |
+
},
|
| 3406 |
+
{
|
| 3407 |
+
"epoch": 0.7100721395306365,
|
| 3408 |
+
"grad_norm": 0.11503782123327255,
|
| 3409 |
+
"learning_rate": 6.78082191780822e-06,
|
| 3410 |
+
"loss": 1.1977,
|
| 3411 |
+
"step": 486
|
| 3412 |
+
},
|
| 3413 |
+
{
|
| 3414 |
+
"epoch": 0.7115331933156789,
|
| 3415 |
+
"grad_norm": 0.11059726774692535,
|
| 3416 |
+
"learning_rate": 6.746575342465754e-06,
|
| 3417 |
+
"loss": 1.1771,
|
| 3418 |
+
"step": 487
|
| 3419 |
+
},
|
| 3420 |
+
{
|
| 3421 |
+
"epoch": 0.7129942471007213,
|
| 3422 |
+
"grad_norm": 0.12105996906757355,
|
| 3423 |
+
"learning_rate": 6.712328767123288e-06,
|
| 3424 |
+
"loss": 1.218,
|
| 3425 |
+
"step": 488
|
| 3426 |
+
},
|
| 3427 |
+
{
|
| 3428 |
+
"epoch": 0.7144553008857638,
|
| 3429 |
+
"grad_norm": 0.1285051554441452,
|
| 3430 |
+
"learning_rate": 6.678082191780823e-06,
|
| 3431 |
+
"loss": 1.1423,
|
| 3432 |
+
"step": 489
|
| 3433 |
+
},
|
| 3434 |
+
{
|
| 3435 |
+
"epoch": 0.7159163546708063,
|
| 3436 |
+
"grad_norm": 0.11841081827878952,
|
| 3437 |
+
"learning_rate": 6.643835616438357e-06,
|
| 3438 |
+
"loss": 1.1433,
|
| 3439 |
+
"step": 490
|
| 3440 |
+
},
|
| 3441 |
+
{
|
| 3442 |
+
"epoch": 0.7173774084558487,
|
| 3443 |
+
"grad_norm": 0.11378856003284454,
|
| 3444 |
+
"learning_rate": 6.609589041095891e-06,
|
| 3445 |
+
"loss": 1.1941,
|
| 3446 |
+
"step": 491
|
| 3447 |
+
},
|
| 3448 |
+
{
|
| 3449 |
+
"epoch": 0.7188384622408912,
|
| 3450 |
+
"grad_norm": 0.10600030422210693,
|
| 3451 |
+
"learning_rate": 6.5753424657534245e-06,
|
| 3452 |
+
"loss": 1.188,
|
| 3453 |
+
"step": 492
|
| 3454 |
+
},
|
| 3455 |
+
{
|
| 3456 |
+
"epoch": 0.7202995160259337,
|
| 3457 |
+
"grad_norm": 0.10563358664512634,
|
| 3458 |
+
"learning_rate": 6.541095890410959e-06,
|
| 3459 |
+
"loss": 1.2165,
|
| 3460 |
+
"step": 493
|
| 3461 |
+
},
|
| 3462 |
+
{
|
| 3463 |
+
"epoch": 0.7217605698109761,
|
| 3464 |
+
"grad_norm": 0.11092449724674225,
|
| 3465 |
+
"learning_rate": 6.506849315068494e-06,
|
| 3466 |
+
"loss": 1.1956,
|
| 3467 |
+
"step": 494
|
| 3468 |
+
},
|
| 3469 |
+
{
|
| 3470 |
+
"epoch": 0.7232216235960186,
|
| 3471 |
+
"grad_norm": 0.11748078465461731,
|
| 3472 |
+
"learning_rate": 6.472602739726028e-06,
|
| 3473 |
+
"loss": 1.1741,
|
| 3474 |
+
"step": 495
|
| 3475 |
+
},
|
| 3476 |
+
{
|
| 3477 |
+
"epoch": 0.7246826773810611,
|
| 3478 |
+
"grad_norm": 0.12683749198913574,
|
| 3479 |
+
"learning_rate": 6.438356164383563e-06,
|
| 3480 |
+
"loss": 1.166,
|
| 3481 |
+
"step": 496
|
| 3482 |
+
},
|
| 3483 |
+
{
|
| 3484 |
+
"epoch": 0.7261437311661035,
|
| 3485 |
+
"grad_norm": 0.10706394910812378,
|
| 3486 |
+
"learning_rate": 6.4041095890410965e-06,
|
| 3487 |
+
"loss": 1.2278,
|
| 3488 |
+
"step": 497
|
| 3489 |
+
},
|
| 3490 |
+
{
|
| 3491 |
+
"epoch": 0.727604784951146,
|
| 3492 |
+
"grad_norm": 0.10801483690738678,
|
| 3493 |
+
"learning_rate": 6.36986301369863e-06,
|
| 3494 |
+
"loss": 1.1693,
|
| 3495 |
+
"step": 498
|
| 3496 |
+
},
|
| 3497 |
+
{
|
| 3498 |
+
"epoch": 0.7290658387361885,
|
| 3499 |
+
"grad_norm": 0.12337271869182587,
|
| 3500 |
+
"learning_rate": 6.335616438356164e-06,
|
| 3501 |
+
"loss": 1.229,
|
| 3502 |
+
"step": 499
|
| 3503 |
+
},
|
| 3504 |
+
{
|
| 3505 |
+
"epoch": 0.7305268925212309,
|
| 3506 |
+
"grad_norm": 0.13341547548770905,
|
| 3507 |
+
"learning_rate": 6.301369863013699e-06,
|
| 3508 |
+
"loss": 1.1274,
|
| 3509 |
+
"step": 500
|
| 3510 |
+
},
|
| 3511 |
+
{
|
| 3512 |
+
"epoch": 0.7319879463062734,
|
| 3513 |
+
"grad_norm": 0.11283931136131287,
|
| 3514 |
+
"learning_rate": 6.267123287671233e-06,
|
| 3515 |
+
"loss": 1.1867,
|
| 3516 |
+
"step": 501
|
| 3517 |
+
},
|
| 3518 |
+
{
|
| 3519 |
+
"epoch": 0.7334490000913159,
|
| 3520 |
+
"grad_norm": 0.1254453957080841,
|
| 3521 |
+
"learning_rate": 6.2328767123287685e-06,
|
| 3522 |
+
"loss": 1.2634,
|
| 3523 |
+
"step": 502
|
| 3524 |
+
},
|
| 3525 |
+
{
|
| 3526 |
+
"epoch": 0.7349100538763583,
|
| 3527 |
+
"grad_norm": 0.125976100564003,
|
| 3528 |
+
"learning_rate": 6.198630136986302e-06,
|
| 3529 |
+
"loss": 1.0729,
|
| 3530 |
+
"step": 503
|
| 3531 |
+
},
|
| 3532 |
+
{
|
| 3533 |
+
"epoch": 0.7363711076614008,
|
| 3534 |
+
"grad_norm": 0.10732964426279068,
|
| 3535 |
+
"learning_rate": 6.164383561643836e-06,
|
| 3536 |
+
"loss": 1.2072,
|
| 3537 |
+
"step": 504
|
| 3538 |
+
},
|
| 3539 |
+
{
|
| 3540 |
+
"epoch": 0.7378321614464433,
|
| 3541 |
+
"grad_norm": 0.12129033356904984,
|
| 3542 |
+
"learning_rate": 6.13013698630137e-06,
|
| 3543 |
+
"loss": 1.1197,
|
| 3544 |
+
"step": 505
|
| 3545 |
+
},
|
| 3546 |
+
{
|
| 3547 |
+
"epoch": 0.7392932152314857,
|
| 3548 |
+
"grad_norm": 0.12821297347545624,
|
| 3549 |
+
"learning_rate": 6.095890410958905e-06,
|
| 3550 |
+
"loss": 1.1515,
|
| 3551 |
+
"step": 506
|
| 3552 |
+
},
|
| 3553 |
+
{
|
| 3554 |
+
"epoch": 0.7407542690165282,
|
| 3555 |
+
"grad_norm": 0.11655119061470032,
|
| 3556 |
+
"learning_rate": 6.061643835616439e-06,
|
| 3557 |
+
"loss": 1.172,
|
| 3558 |
+
"step": 507
|
| 3559 |
+
},
|
| 3560 |
+
{
|
| 3561 |
+
"epoch": 0.7422153228015707,
|
| 3562 |
+
"grad_norm": 0.11455903202295303,
|
| 3563 |
+
"learning_rate": 6.027397260273973e-06,
|
| 3564 |
+
"loss": 1.1762,
|
| 3565 |
+
"step": 508
|
| 3566 |
+
},
|
| 3567 |
+
{
|
| 3568 |
+
"epoch": 0.7436763765866131,
|
| 3569 |
+
"grad_norm": 0.10701651871204376,
|
| 3570 |
+
"learning_rate": 5.993150684931507e-06,
|
| 3571 |
+
"loss": 1.1685,
|
| 3572 |
+
"step": 509
|
| 3573 |
+
},
|
| 3574 |
+
{
|
| 3575 |
+
"epoch": 0.7451374303716556,
|
| 3576 |
+
"grad_norm": 0.11655491590499878,
|
| 3577 |
+
"learning_rate": 5.958904109589042e-06,
|
| 3578 |
+
"loss": 1.1975,
|
| 3579 |
+
"step": 510
|
| 3580 |
+
},
|
| 3581 |
+
{
|
| 3582 |
+
"epoch": 0.746598484156698,
|
| 3583 |
+
"grad_norm": 0.11159254610538483,
|
| 3584 |
+
"learning_rate": 5.924657534246576e-06,
|
| 3585 |
+
"loss": 1.2226,
|
| 3586 |
+
"step": 511
|
| 3587 |
+
},
|
| 3588 |
+
{
|
| 3589 |
+
"epoch": 0.7480595379417405,
|
| 3590 |
+
"grad_norm": 0.11702670156955719,
|
| 3591 |
+
"learning_rate": 5.89041095890411e-06,
|
| 3592 |
+
"loss": 1.1552,
|
| 3593 |
+
"step": 512
|
| 3594 |
+
},
|
| 3595 |
+
{
|
| 3596 |
+
"epoch": 0.749520591726783,
|
| 3597 |
+
"grad_norm": 0.11294779181480408,
|
| 3598 |
+
"learning_rate": 5.856164383561645e-06,
|
| 3599 |
+
"loss": 1.189,
|
| 3600 |
+
"step": 513
|
| 3601 |
+
},
|
| 3602 |
+
{
|
| 3603 |
+
"epoch": 0.7509816455118254,
|
| 3604 |
+
"grad_norm": 0.10862728208303452,
|
| 3605 |
+
"learning_rate": 5.821917808219179e-06,
|
| 3606 |
+
"loss": 1.1997,
|
| 3607 |
+
"step": 514
|
| 3608 |
+
},
|
| 3609 |
+
{
|
| 3610 |
+
"epoch": 0.7524426992968679,
|
| 3611 |
+
"grad_norm": 0.12142271548509598,
|
| 3612 |
+
"learning_rate": 5.7876712328767125e-06,
|
| 3613 |
+
"loss": 1.1985,
|
| 3614 |
+
"step": 515
|
| 3615 |
+
},
|
| 3616 |
+
{
|
| 3617 |
+
"epoch": 0.7539037530819104,
|
| 3618 |
+
"grad_norm": 0.11388342082500458,
|
| 3619 |
+
"learning_rate": 5.753424657534246e-06,
|
| 3620 |
+
"loss": 1.1195,
|
| 3621 |
+
"step": 516
|
| 3622 |
+
},
|
| 3623 |
+
{
|
| 3624 |
+
"epoch": 0.7553648068669528,
|
| 3625 |
+
"grad_norm": 0.11917892098426819,
|
| 3626 |
+
"learning_rate": 5.719178082191781e-06,
|
| 3627 |
+
"loss": 1.1685,
|
| 3628 |
+
"step": 517
|
| 3629 |
+
},
|
| 3630 |
+
{
|
| 3631 |
+
"epoch": 0.7568258606519952,
|
| 3632 |
+
"grad_norm": 0.11466323584318161,
|
| 3633 |
+
"learning_rate": 5.684931506849316e-06,
|
| 3634 |
+
"loss": 1.1837,
|
| 3635 |
+
"step": 518
|
| 3636 |
+
},
|
| 3637 |
+
{
|
| 3638 |
+
"epoch": 0.7582869144370377,
|
| 3639 |
+
"grad_norm": 0.11477669328451157,
|
| 3640 |
+
"learning_rate": 5.6506849315068506e-06,
|
| 3641 |
+
"loss": 1.2336,
|
| 3642 |
+
"step": 519
|
| 3643 |
+
},
|
| 3644 |
+
{
|
| 3645 |
+
"epoch": 0.7597479682220801,
|
| 3646 |
+
"grad_norm": 0.12040074169635773,
|
| 3647 |
+
"learning_rate": 5.6164383561643845e-06,
|
| 3648 |
+
"loss": 1.2553,
|
| 3649 |
+
"step": 520
|
| 3650 |
+
},
|
| 3651 |
+
{
|
| 3652 |
+
"epoch": 0.7612090220071226,
|
| 3653 |
+
"grad_norm": 0.11461540311574936,
|
| 3654 |
+
"learning_rate": 5.582191780821918e-06,
|
| 3655 |
+
"loss": 1.1205,
|
| 3656 |
+
"step": 521
|
| 3657 |
+
},
|
| 3658 |
+
{
|
| 3659 |
+
"epoch": 0.7626700757921651,
|
| 3660 |
+
"grad_norm": 0.10867593437433243,
|
| 3661 |
+
"learning_rate": 5.547945205479452e-06,
|
| 3662 |
+
"loss": 1.1629,
|
| 3663 |
+
"step": 522
|
| 3664 |
+
},
|
| 3665 |
+
{
|
| 3666 |
+
"epoch": 0.7641311295772075,
|
| 3667 |
+
"grad_norm": 0.12366941571235657,
|
| 3668 |
+
"learning_rate": 5.513698630136987e-06,
|
| 3669 |
+
"loss": 1.2059,
|
| 3670 |
+
"step": 523
|
| 3671 |
+
},
|
| 3672 |
+
{
|
| 3673 |
+
"epoch": 0.76559218336225,
|
| 3674 |
+
"grad_norm": 0.11807534843683243,
|
| 3675 |
+
"learning_rate": 5.479452054794521e-06,
|
| 3676 |
+
"loss": 1.2863,
|
| 3677 |
+
"step": 524
|
| 3678 |
+
},
|
| 3679 |
+
{
|
| 3680 |
+
"epoch": 0.7670532371472925,
|
| 3681 |
+
"grad_norm": 0.11750409007072449,
|
| 3682 |
+
"learning_rate": 5.445205479452055e-06,
|
| 3683 |
+
"loss": 1.2293,
|
| 3684 |
+
"step": 525
|
| 3685 |
+
},
|
| 3686 |
+
{
|
| 3687 |
+
"epoch": 0.7685142909323349,
|
| 3688 |
+
"grad_norm": 0.11874507367610931,
|
| 3689 |
+
"learning_rate": 5.41095890410959e-06,
|
| 3690 |
+
"loss": 1.2275,
|
| 3691 |
+
"step": 526
|
| 3692 |
+
},
|
| 3693 |
+
{
|
| 3694 |
+
"epoch": 0.7699753447173774,
|
| 3695 |
+
"grad_norm": 0.12096529453992844,
|
| 3696 |
+
"learning_rate": 5.376712328767124e-06,
|
| 3697 |
+
"loss": 1.1525,
|
| 3698 |
+
"step": 527
|
| 3699 |
+
},
|
| 3700 |
+
{
|
| 3701 |
+
"epoch": 0.7714363985024199,
|
| 3702 |
+
"grad_norm": 0.10740137845277786,
|
| 3703 |
+
"learning_rate": 5.342465753424658e-06,
|
| 3704 |
+
"loss": 1.2267,
|
| 3705 |
+
"step": 528
|
| 3706 |
+
},
|
| 3707 |
+
{
|
| 3708 |
+
"epoch": 0.7728974522874623,
|
| 3709 |
+
"grad_norm": 0.12556499242782593,
|
| 3710 |
+
"learning_rate": 5.308219178082192e-06,
|
| 3711 |
+
"loss": 1.2042,
|
| 3712 |
+
"step": 529
|
| 3713 |
+
},
|
| 3714 |
+
{
|
| 3715 |
+
"epoch": 0.7743585060725048,
|
| 3716 |
+
"grad_norm": 0.1109650582075119,
|
| 3717 |
+
"learning_rate": 5.273972602739727e-06,
|
| 3718 |
+
"loss": 1.1813,
|
| 3719 |
+
"step": 530
|
| 3720 |
+
},
|
| 3721 |
+
{
|
| 3722 |
+
"epoch": 0.7758195598575472,
|
| 3723 |
+
"grad_norm": 0.10629246383905411,
|
| 3724 |
+
"learning_rate": 5.239726027397261e-06,
|
| 3725 |
+
"loss": 1.1959,
|
| 3726 |
+
"step": 531
|
| 3727 |
+
},
|
| 3728 |
+
{
|
| 3729 |
+
"epoch": 0.7772806136425897,
|
| 3730 |
+
"grad_norm": 0.1113128662109375,
|
| 3731 |
+
"learning_rate": 5.2054794520547945e-06,
|
| 3732 |
+
"loss": 1.1441,
|
| 3733 |
+
"step": 532
|
| 3734 |
+
},
|
| 3735 |
+
{
|
| 3736 |
+
"epoch": 0.7787416674276322,
|
| 3737 |
+
"grad_norm": 0.11163881421089172,
|
| 3738 |
+
"learning_rate": 5.171232876712328e-06,
|
| 3739 |
+
"loss": 1.1991,
|
| 3740 |
+
"step": 533
|
| 3741 |
+
},
|
| 3742 |
+
{
|
| 3743 |
+
"epoch": 0.7802027212126746,
|
| 3744 |
+
"grad_norm": 0.12462608516216278,
|
| 3745 |
+
"learning_rate": 5.136986301369864e-06,
|
| 3746 |
+
"loss": 1.1397,
|
| 3747 |
+
"step": 534
|
| 3748 |
+
},
|
| 3749 |
+
{
|
| 3750 |
+
"epoch": 0.7816637749977171,
|
| 3751 |
+
"grad_norm": 0.11631737649440765,
|
| 3752 |
+
"learning_rate": 5.102739726027398e-06,
|
| 3753 |
+
"loss": 1.1253,
|
| 3754 |
+
"step": 535
|
| 3755 |
+
},
|
| 3756 |
+
{
|
| 3757 |
+
"epoch": 0.7831248287827596,
|
| 3758 |
+
"grad_norm": 0.12405448406934738,
|
| 3759 |
+
"learning_rate": 5.068493150684932e-06,
|
| 3760 |
+
"loss": 1.1595,
|
| 3761 |
+
"step": 536
|
| 3762 |
+
},
|
| 3763 |
+
{
|
| 3764 |
+
"epoch": 0.784585882567802,
|
| 3765 |
+
"grad_norm": 0.1106800064444542,
|
| 3766 |
+
"learning_rate": 5.0342465753424665e-06,
|
| 3767 |
+
"loss": 1.2755,
|
| 3768 |
+
"step": 537
|
| 3769 |
+
},
|
| 3770 |
+
{
|
| 3771 |
+
"epoch": 0.7860469363528445,
|
| 3772 |
+
"grad_norm": 0.10687270015478134,
|
| 3773 |
+
"learning_rate": 5e-06,
|
| 3774 |
+
"loss": 1.1443,
|
| 3775 |
+
"step": 538
|
| 3776 |
+
},
|
| 3777 |
+
{
|
| 3778 |
+
"epoch": 0.787507990137887,
|
| 3779 |
+
"grad_norm": 0.10897688567638397,
|
| 3780 |
+
"learning_rate": 4.965753424657534e-06,
|
| 3781 |
+
"loss": 1.1917,
|
| 3782 |
+
"step": 539
|
| 3783 |
+
},
|
| 3784 |
+
{
|
| 3785 |
+
"epoch": 0.7889690439229294,
|
| 3786 |
+
"grad_norm": 0.12325593084096909,
|
| 3787 |
+
"learning_rate": 4.931506849315069e-06,
|
| 3788 |
+
"loss": 1.165,
|
| 3789 |
+
"step": 540
|
| 3790 |
+
},
|
| 3791 |
+
{
|
| 3792 |
+
"epoch": 0.7904300977079719,
|
| 3793 |
+
"grad_norm": 0.12180227786302567,
|
| 3794 |
+
"learning_rate": 4.897260273972603e-06,
|
| 3795 |
+
"loss": 1.15,
|
| 3796 |
+
"step": 541
|
| 3797 |
+
},
|
| 3798 |
+
{
|
| 3799 |
+
"epoch": 0.7918911514930144,
|
| 3800 |
+
"grad_norm": 0.1084585040807724,
|
| 3801 |
+
"learning_rate": 4.863013698630138e-06,
|
| 3802 |
+
"loss": 1.2431,
|
| 3803 |
+
"step": 542
|
| 3804 |
+
},
|
| 3805 |
+
{
|
| 3806 |
+
"epoch": 0.7933522052780568,
|
| 3807 |
+
"grad_norm": 0.11849282681941986,
|
| 3808 |
+
"learning_rate": 4.8287671232876716e-06,
|
| 3809 |
+
"loss": 1.2494,
|
| 3810 |
+
"step": 543
|
| 3811 |
+
},
|
| 3812 |
+
{
|
| 3813 |
+
"epoch": 0.7948132590630993,
|
| 3814 |
+
"grad_norm": 0.1109924465417862,
|
| 3815 |
+
"learning_rate": 4.7945205479452054e-06,
|
| 3816 |
+
"loss": 1.1517,
|
| 3817 |
+
"step": 544
|
| 3818 |
+
},
|
| 3819 |
+
{
|
| 3820 |
+
"epoch": 0.7962743128481418,
|
| 3821 |
+
"grad_norm": 0.1313486099243164,
|
| 3822 |
+
"learning_rate": 4.76027397260274e-06,
|
| 3823 |
+
"loss": 1.2044,
|
| 3824 |
+
"step": 545
|
| 3825 |
+
},
|
| 3826 |
+
{
|
| 3827 |
+
"epoch": 0.7977353666331842,
|
| 3828 |
+
"grad_norm": 0.12946535646915436,
|
| 3829 |
+
"learning_rate": 4.726027397260274e-06,
|
| 3830 |
+
"loss": 1.101,
|
| 3831 |
+
"step": 546
|
| 3832 |
+
},
|
| 3833 |
+
{
|
| 3834 |
+
"epoch": 0.7991964204182267,
|
| 3835 |
+
"grad_norm": 0.1083068773150444,
|
| 3836 |
+
"learning_rate": 4.691780821917809e-06,
|
| 3837 |
+
"loss": 1.1933,
|
| 3838 |
+
"step": 547
|
| 3839 |
+
},
|
| 3840 |
+
{
|
| 3841 |
+
"epoch": 0.8006574742032692,
|
| 3842 |
+
"grad_norm": 0.11590442061424255,
|
| 3843 |
+
"learning_rate": 4.657534246575343e-06,
|
| 3844 |
+
"loss": 1.198,
|
| 3845 |
+
"step": 548
|
| 3846 |
+
},
|
| 3847 |
+
{
|
| 3848 |
+
"epoch": 0.8021185279883116,
|
| 3849 |
+
"grad_norm": 0.11123711615800858,
|
| 3850 |
+
"learning_rate": 4.6232876712328774e-06,
|
| 3851 |
+
"loss": 1.1259,
|
| 3852 |
+
"step": 549
|
| 3853 |
+
},
|
| 3854 |
+
{
|
| 3855 |
+
"epoch": 0.803579581773354,
|
| 3856 |
+
"grad_norm": 0.1170891597867012,
|
| 3857 |
+
"learning_rate": 4.589041095890411e-06,
|
| 3858 |
+
"loss": 1.1977,
|
| 3859 |
+
"step": 550
|
| 3860 |
+
},
|
| 3861 |
+
{
|
| 3862 |
+
"epoch": 0.8050406355583964,
|
| 3863 |
+
"grad_norm": 0.12325557321310043,
|
| 3864 |
+
"learning_rate": 4.554794520547945e-06,
|
| 3865 |
+
"loss": 1.1549,
|
| 3866 |
+
"step": 551
|
| 3867 |
+
},
|
| 3868 |
+
{
|
| 3869 |
+
"epoch": 0.8065016893434389,
|
| 3870 |
+
"grad_norm": 0.10925264656543732,
|
| 3871 |
+
"learning_rate": 4.52054794520548e-06,
|
| 3872 |
+
"loss": 1.1376,
|
| 3873 |
+
"step": 552
|
| 3874 |
+
},
|
| 3875 |
+
{
|
| 3876 |
+
"epoch": 0.8079627431284814,
|
| 3877 |
+
"grad_norm": 0.10057859122753143,
|
| 3878 |
+
"learning_rate": 4.486301369863014e-06,
|
| 3879 |
+
"loss": 1.1237,
|
| 3880 |
+
"step": 553
|
| 3881 |
}
|
| 3882 |
],
|
| 3883 |
"logging_steps": 1,
|
|
|
|
| 3897 |
"attributes": {}
|
| 3898 |
}
|
| 3899 |
},
|
| 3900 |
+
"total_flos": 6.248047407796224e+17,
|
| 3901 |
"train_batch_size": 4,
|
| 3902 |
"trial_name": null,
|
| 3903 |
"trial_params": null
|