Training in progress, step 47000, 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 140637408
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7f0ee59efb89be6c8c2f2da1a15ad64aea84eb23e30ffc79e933c5c95c4b71ed
|
| 3 |
size 140637408
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 281362507
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d5613fd93a6eef63cc579895de48170bbf1a18d8c1d7e673913762c125aa187c
|
| 3 |
size 281362507
|
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:fb096c09a5ceae5902cef83e88874ae4114869833d2337c185f133bf18eb7ed2
|
| 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:e15e095d88bb49dabc1d3b5734aa85480dc84afb235f317f103ea6bed7007201
|
| 3 |
size 1465
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -2,9 +2,9 @@
|
|
| 2 |
"best_global_step": null,
|
| 3 |
"best_metric": null,
|
| 4 |
"best_model_checkpoint": null,
|
| 5 |
-
"epoch": 0.
|
| 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,
|
|
@@ -33395,6 +33395,365 @@
|
|
| 33395 |
"eval_samples_per_second": 32.9,
|
| 33396 |
"eval_steps_per_second": 0.526,
|
| 33397 |
"step": 46500
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 33398 |
}
|
| 33399 |
],
|
| 33400 |
"logging_steps": 10,
|
|
@@ -33414,7 +33773,7 @@
|
|
| 33414 |
"attributes": {}
|
| 33415 |
}
|
| 33416 |
},
|
| 33417 |
-
"total_flos": 8.
|
| 33418 |
"train_batch_size": 128,
|
| 33419 |
"trial_name": null,
|
| 33420 |
"trial_params": null
|
|
|
|
| 2 |
"best_global_step": null,
|
| 3 |
"best_metric": null,
|
| 4 |
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 0.6160379584239914,
|
| 6 |
"eval_steps": 500,
|
| 7 |
+
"global_step": 47000,
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
|
|
|
| 33395 |
"eval_samples_per_second": 32.9,
|
| 33396 |
"eval_steps_per_second": 0.526,
|
| 33397 |
"step": 46500
|
| 33398 |
+
},
|
| 33399 |
+
{
|
| 33400 |
+
"epoch": 0.6096154350276561,
|
| 33401 |
+
"grad_norm": 0.7248334288597107,
|
| 33402 |
+
"learning_rate": 0.00010050970265938755,
|
| 33403 |
+
"loss": 3.2191,
|
| 33404 |
+
"step": 46510
|
| 33405 |
+
},
|
| 33406 |
+
{
|
| 33407 |
+
"epoch": 0.6097465069337038,
|
| 33408 |
+
"grad_norm": 0.6939072608947754,
|
| 33409 |
+
"learning_rate": 0.0001004510147745937,
|
| 33410 |
+
"loss": 3.2081,
|
| 33411 |
+
"step": 46520
|
| 33412 |
+
},
|
| 33413 |
+
{
|
| 33414 |
+
"epoch": 0.6098775788397515,
|
| 33415 |
+
"grad_norm": 0.6874194145202637,
|
| 33416 |
+
"learning_rate": 0.00010039233540245125,
|
| 33417 |
+
"loss": 3.2095,
|
| 33418 |
+
"step": 46530
|
| 33419 |
+
},
|
| 33420 |
+
{
|
| 33421 |
+
"epoch": 0.6100086507457991,
|
| 33422 |
+
"grad_norm": 0.6661149859428406,
|
| 33423 |
+
"learning_rate": 0.00010033366455304145,
|
| 33424 |
+
"loss": 3.212,
|
| 33425 |
+
"step": 46540
|
| 33426 |
+
},
|
| 33427 |
+
{
|
| 33428 |
+
"epoch": 0.6101397226518468,
|
| 33429 |
+
"grad_norm": 0.6120688915252686,
|
| 33430 |
+
"learning_rate": 0.00010027500223644424,
|
| 33431 |
+
"loss": 3.2682,
|
| 33432 |
+
"step": 46550
|
| 33433 |
+
},
|
| 33434 |
+
{
|
| 33435 |
+
"epoch": 0.6102707945578945,
|
| 33436 |
+
"grad_norm": 0.6188989281654358,
|
| 33437 |
+
"learning_rate": 0.00010021634846273775,
|
| 33438 |
+
"loss": 3.2536,
|
| 33439 |
+
"step": 46560
|
| 33440 |
+
},
|
| 33441 |
+
{
|
| 33442 |
+
"epoch": 0.6104018664639421,
|
| 33443 |
+
"grad_norm": 0.677006721496582,
|
| 33444 |
+
"learning_rate": 0.00010015770324199903,
|
| 33445 |
+
"loss": 3.2352,
|
| 33446 |
+
"step": 46570
|
| 33447 |
+
},
|
| 33448 |
+
{
|
| 33449 |
+
"epoch": 0.6105329383699898,
|
| 33450 |
+
"grad_norm": 0.5996061563491821,
|
| 33451 |
+
"learning_rate": 0.00010009906658430341,
|
| 33452 |
+
"loss": 3.235,
|
| 33453 |
+
"step": 46580
|
| 33454 |
+
},
|
| 33455 |
+
{
|
| 33456 |
+
"epoch": 0.6106640102760374,
|
| 33457 |
+
"grad_norm": 0.7179942727088928,
|
| 33458 |
+
"learning_rate": 0.00010004043849972489,
|
| 33459 |
+
"loss": 3.1978,
|
| 33460 |
+
"step": 46590
|
| 33461 |
+
},
|
| 33462 |
+
{
|
| 33463 |
+
"epoch": 0.6107950821820851,
|
| 33464 |
+
"grad_norm": 0.6874227523803711,
|
| 33465 |
+
"learning_rate": 9.998181899833583e-05,
|
| 33466 |
+
"loss": 3.1724,
|
| 33467 |
+
"step": 46600
|
| 33468 |
+
},
|
| 33469 |
+
{
|
| 33470 |
+
"epoch": 0.6109261540881328,
|
| 33471 |
+
"grad_norm": 0.7165815830230713,
|
| 33472 |
+
"learning_rate": 9.99232080902073e-05,
|
| 33473 |
+
"loss": 3.1896,
|
| 33474 |
+
"step": 46610
|
| 33475 |
+
},
|
| 33476 |
+
{
|
| 33477 |
+
"epoch": 0.6110572259941804,
|
| 33478 |
+
"grad_norm": 0.6888381242752075,
|
| 33479 |
+
"learning_rate": 9.986460578540886e-05,
|
| 33480 |
+
"loss": 3.2641,
|
| 33481 |
+
"step": 46620
|
| 33482 |
+
},
|
| 33483 |
+
{
|
| 33484 |
+
"epoch": 0.6111882979002281,
|
| 33485 |
+
"grad_norm": 0.7022085785865784,
|
| 33486 |
+
"learning_rate": 9.980601209400843e-05,
|
| 33487 |
+
"loss": 3.2582,
|
| 33488 |
+
"step": 46630
|
| 33489 |
+
},
|
| 33490 |
+
{
|
| 33491 |
+
"epoch": 0.6113193698062758,
|
| 33492 |
+
"grad_norm": 0.6943024396896362,
|
| 33493 |
+
"learning_rate": 9.974742702607262e-05,
|
| 33494 |
+
"loss": 3.1946,
|
| 33495 |
+
"step": 46640
|
| 33496 |
+
},
|
| 33497 |
+
{
|
| 33498 |
+
"epoch": 0.6114504417123234,
|
| 33499 |
+
"grad_norm": 0.6320481896400452,
|
| 33500 |
+
"learning_rate": 9.968885059166653e-05,
|
| 33501 |
+
"loss": 3.2411,
|
| 33502 |
+
"step": 46650
|
| 33503 |
+
},
|
| 33504 |
+
{
|
| 33505 |
+
"epoch": 0.611581513618371,
|
| 33506 |
+
"grad_norm": 0.6720809936523438,
|
| 33507 |
+
"learning_rate": 9.963028280085373e-05,
|
| 33508 |
+
"loss": 3.2235,
|
| 33509 |
+
"step": 46660
|
| 33510 |
+
},
|
| 33511 |
+
{
|
| 33512 |
+
"epoch": 0.6117125855244187,
|
| 33513 |
+
"grad_norm": 0.6375182867050171,
|
| 33514 |
+
"learning_rate": 9.957172366369632e-05,
|
| 33515 |
+
"loss": 3.2466,
|
| 33516 |
+
"step": 46670
|
| 33517 |
+
},
|
| 33518 |
+
{
|
| 33519 |
+
"epoch": 0.6118436574304663,
|
| 33520 |
+
"grad_norm": 0.6096304655075073,
|
| 33521 |
+
"learning_rate": 9.951317319025491e-05,
|
| 33522 |
+
"loss": 3.228,
|
| 33523 |
+
"step": 46680
|
| 33524 |
+
},
|
| 33525 |
+
{
|
| 33526 |
+
"epoch": 0.611974729336514,
|
| 33527 |
+
"grad_norm": 0.6518574953079224,
|
| 33528 |
+
"learning_rate": 9.945463139058867e-05,
|
| 33529 |
+
"loss": 3.2361,
|
| 33530 |
+
"step": 46690
|
| 33531 |
+
},
|
| 33532 |
+
{
|
| 33533 |
+
"epoch": 0.6121058012425616,
|
| 33534 |
+
"grad_norm": 0.7310232520103455,
|
| 33535 |
+
"learning_rate": 9.939609827475524e-05,
|
| 33536 |
+
"loss": 3.2229,
|
| 33537 |
+
"step": 46700
|
| 33538 |
+
},
|
| 33539 |
+
{
|
| 33540 |
+
"epoch": 0.6122368731486093,
|
| 33541 |
+
"grad_norm": 0.5949954986572266,
|
| 33542 |
+
"learning_rate": 9.933757385281073e-05,
|
| 33543 |
+
"loss": 3.2449,
|
| 33544 |
+
"step": 46710
|
| 33545 |
+
},
|
| 33546 |
+
{
|
| 33547 |
+
"epoch": 0.612367945054657,
|
| 33548 |
+
"grad_norm": 0.6023111343383789,
|
| 33549 |
+
"learning_rate": 9.927905813480977e-05,
|
| 33550 |
+
"loss": 3.2233,
|
| 33551 |
+
"step": 46720
|
| 33552 |
+
},
|
| 33553 |
+
{
|
| 33554 |
+
"epoch": 0.6124990169607046,
|
| 33555 |
+
"grad_norm": 0.6577599048614502,
|
| 33556 |
+
"learning_rate": 9.922055113080565e-05,
|
| 33557 |
+
"loss": 3.2113,
|
| 33558 |
+
"step": 46730
|
| 33559 |
+
},
|
| 33560 |
+
{
|
| 33561 |
+
"epoch": 0.6126300888667523,
|
| 33562 |
+
"grad_norm": 0.6531931161880493,
|
| 33563 |
+
"learning_rate": 9.91620528508499e-05,
|
| 33564 |
+
"loss": 3.2246,
|
| 33565 |
+
"step": 46740
|
| 33566 |
+
},
|
| 33567 |
+
{
|
| 33568 |
+
"epoch": 0.6127611607728,
|
| 33569 |
+
"grad_norm": 0.7038547396659851,
|
| 33570 |
+
"learning_rate": 9.910356330499277e-05,
|
| 33571 |
+
"loss": 3.2321,
|
| 33572 |
+
"step": 46750
|
| 33573 |
+
},
|
| 33574 |
+
{
|
| 33575 |
+
"epoch": 0.6128922326788476,
|
| 33576 |
+
"grad_norm": 0.838652491569519,
|
| 33577 |
+
"learning_rate": 9.904508250328285e-05,
|
| 33578 |
+
"loss": 3.2985,
|
| 33579 |
+
"step": 46760
|
| 33580 |
+
},
|
| 33581 |
+
{
|
| 33582 |
+
"epoch": 0.6130233045848953,
|
| 33583 |
+
"grad_norm": 0.6685378551483154,
|
| 33584 |
+
"learning_rate": 9.89866104557674e-05,
|
| 33585 |
+
"loss": 3.2357,
|
| 33586 |
+
"step": 46770
|
| 33587 |
+
},
|
| 33588 |
+
{
|
| 33589 |
+
"epoch": 0.613154376490943,
|
| 33590 |
+
"grad_norm": 0.7856571078300476,
|
| 33591 |
+
"learning_rate": 9.892814717249197e-05,
|
| 33592 |
+
"loss": 3.2264,
|
| 33593 |
+
"step": 46780
|
| 33594 |
+
},
|
| 33595 |
+
{
|
| 33596 |
+
"epoch": 0.6132854483969906,
|
| 33597 |
+
"grad_norm": 0.6753517389297485,
|
| 33598 |
+
"learning_rate": 9.886969266350074e-05,
|
| 33599 |
+
"loss": 3.2112,
|
| 33600 |
+
"step": 46790
|
| 33601 |
+
},
|
| 33602 |
+
{
|
| 33603 |
+
"epoch": 0.6134165203030383,
|
| 33604 |
+
"grad_norm": 0.6568011045455933,
|
| 33605 |
+
"learning_rate": 9.881124693883639e-05,
|
| 33606 |
+
"loss": 3.2281,
|
| 33607 |
+
"step": 46800
|
| 33608 |
+
},
|
| 33609 |
+
{
|
| 33610 |
+
"epoch": 0.613547592209086,
|
| 33611 |
+
"grad_norm": 0.707349956035614,
|
| 33612 |
+
"learning_rate": 9.875281000854009e-05,
|
| 33613 |
+
"loss": 3.2872,
|
| 33614 |
+
"step": 46810
|
| 33615 |
+
},
|
| 33616 |
+
{
|
| 33617 |
+
"epoch": 0.6136786641151336,
|
| 33618 |
+
"grad_norm": 0.7462494373321533,
|
| 33619 |
+
"learning_rate": 9.869438188265137e-05,
|
| 33620 |
+
"loss": 3.2561,
|
| 33621 |
+
"step": 46820
|
| 33622 |
+
},
|
| 33623 |
+
{
|
| 33624 |
+
"epoch": 0.6138097360211813,
|
| 33625 |
+
"grad_norm": 0.6671332716941833,
|
| 33626 |
+
"learning_rate": 9.863596257120843e-05,
|
| 33627 |
+
"loss": 3.2462,
|
| 33628 |
+
"step": 46830
|
| 33629 |
+
},
|
| 33630 |
+
{
|
| 33631 |
+
"epoch": 0.6139408079272288,
|
| 33632 |
+
"grad_norm": 0.7925904393196106,
|
| 33633 |
+
"learning_rate": 9.857755208424777e-05,
|
| 33634 |
+
"loss": 3.2392,
|
| 33635 |
+
"step": 46840
|
| 33636 |
+
},
|
| 33637 |
+
{
|
| 33638 |
+
"epoch": 0.6140718798332765,
|
| 33639 |
+
"grad_norm": 0.6601688861846924,
|
| 33640 |
+
"learning_rate": 9.851915043180461e-05,
|
| 33641 |
+
"loss": 3.2442,
|
| 33642 |
+
"step": 46850
|
| 33643 |
+
},
|
| 33644 |
+
{
|
| 33645 |
+
"epoch": 0.6142029517393242,
|
| 33646 |
+
"grad_norm": 0.7439512610435486,
|
| 33647 |
+
"learning_rate": 9.846075762391234e-05,
|
| 33648 |
+
"loss": 3.2471,
|
| 33649 |
+
"step": 46860
|
| 33650 |
+
},
|
| 33651 |
+
{
|
| 33652 |
+
"epoch": 0.6143340236453718,
|
| 33653 |
+
"grad_norm": 0.5951361656188965,
|
| 33654 |
+
"learning_rate": 9.840237367060314e-05,
|
| 33655 |
+
"loss": 3.2039,
|
| 33656 |
+
"step": 46870
|
| 33657 |
+
},
|
| 33658 |
+
{
|
| 33659 |
+
"epoch": 0.6144650955514195,
|
| 33660 |
+
"grad_norm": 0.5695250034332275,
|
| 33661 |
+
"learning_rate": 9.834399858190753e-05,
|
| 33662 |
+
"loss": 3.2369,
|
| 33663 |
+
"step": 46880
|
| 33664 |
+
},
|
| 33665 |
+
{
|
| 33666 |
+
"epoch": 0.6145961674574671,
|
| 33667 |
+
"grad_norm": 0.6092811822891235,
|
| 33668 |
+
"learning_rate": 9.828563236785441e-05,
|
| 33669 |
+
"loss": 3.2454,
|
| 33670 |
+
"step": 46890
|
| 33671 |
+
},
|
| 33672 |
+
{
|
| 33673 |
+
"epoch": 0.6147272393635148,
|
| 33674 |
+
"grad_norm": 0.6412503719329834,
|
| 33675 |
+
"learning_rate": 9.822727503847131e-05,
|
| 33676 |
+
"loss": 3.2488,
|
| 33677 |
+
"step": 46900
|
| 33678 |
+
},
|
| 33679 |
+
{
|
| 33680 |
+
"epoch": 0.6148583112695625,
|
| 33681 |
+
"grad_norm": 0.6511025428771973,
|
| 33682 |
+
"learning_rate": 9.816892660378418e-05,
|
| 33683 |
+
"loss": 3.2716,
|
| 33684 |
+
"step": 46910
|
| 33685 |
+
},
|
| 33686 |
+
{
|
| 33687 |
+
"epoch": 0.6149893831756101,
|
| 33688 |
+
"grad_norm": 0.6001058220863342,
|
| 33689 |
+
"learning_rate": 9.81105870738175e-05,
|
| 33690 |
+
"loss": 3.2538,
|
| 33691 |
+
"step": 46920
|
| 33692 |
+
},
|
| 33693 |
+
{
|
| 33694 |
+
"epoch": 0.6151204550816578,
|
| 33695 |
+
"grad_norm": 0.6340818405151367,
|
| 33696 |
+
"learning_rate": 9.805225645859406e-05,
|
| 33697 |
+
"loss": 3.1722,
|
| 33698 |
+
"step": 46930
|
| 33699 |
+
},
|
| 33700 |
+
{
|
| 33701 |
+
"epoch": 0.6152515269877055,
|
| 33702 |
+
"grad_norm": 0.6574063897132874,
|
| 33703 |
+
"learning_rate": 9.79939347681353e-05,
|
| 33704 |
+
"loss": 3.2109,
|
| 33705 |
+
"step": 46940
|
| 33706 |
+
},
|
| 33707 |
+
{
|
| 33708 |
+
"epoch": 0.6153825988937531,
|
| 33709 |
+
"grad_norm": 0.6378859877586365,
|
| 33710 |
+
"learning_rate": 9.793562201246096e-05,
|
| 33711 |
+
"loss": 3.1795,
|
| 33712 |
+
"step": 46950
|
| 33713 |
+
},
|
| 33714 |
+
{
|
| 33715 |
+
"epoch": 0.6155136707998008,
|
| 33716 |
+
"grad_norm": 0.6317182779312134,
|
| 33717 |
+
"learning_rate": 9.787731820158948e-05,
|
| 33718 |
+
"loss": 3.2079,
|
| 33719 |
+
"step": 46960
|
| 33720 |
+
},
|
| 33721 |
+
{
|
| 33722 |
+
"epoch": 0.6156447427058485,
|
| 33723 |
+
"grad_norm": 0.6875243782997131,
|
| 33724 |
+
"learning_rate": 9.781902334553744e-05,
|
| 33725 |
+
"loss": 3.1947,
|
| 33726 |
+
"step": 46970
|
| 33727 |
+
},
|
| 33728 |
+
{
|
| 33729 |
+
"epoch": 0.6157758146118961,
|
| 33730 |
+
"grad_norm": 0.6557679772377014,
|
| 33731 |
+
"learning_rate": 9.776073745432013e-05,
|
| 33732 |
+
"loss": 3.2715,
|
| 33733 |
+
"step": 46980
|
| 33734 |
+
},
|
| 33735 |
+
{
|
| 33736 |
+
"epoch": 0.6159068865179438,
|
| 33737 |
+
"grad_norm": 0.7047168612480164,
|
| 33738 |
+
"learning_rate": 9.770246053795127e-05,
|
| 33739 |
+
"loss": 3.3205,
|
| 33740 |
+
"step": 46990
|
| 33741 |
+
},
|
| 33742 |
+
{
|
| 33743 |
+
"epoch": 0.6160379584239914,
|
| 33744 |
+
"grad_norm": 0.627339243888855,
|
| 33745 |
+
"learning_rate": 9.764419260644292e-05,
|
| 33746 |
+
"loss": 3.2086,
|
| 33747 |
+
"step": 47000
|
| 33748 |
+
},
|
| 33749 |
+
{
|
| 33750 |
+
"epoch": 0.6160379584239914,
|
| 33751 |
+
"eval_accuracy": 0.39909970674486805,
|
| 33752 |
+
"eval_loss": 3.2312915325164795,
|
| 33753 |
+
"eval_runtime": 60.0193,
|
| 33754 |
+
"eval_samples_per_second": 33.323,
|
| 33755 |
+
"eval_steps_per_second": 0.533,
|
| 33756 |
+
"step": 47000
|
| 33757 |
}
|
| 33758 |
],
|
| 33759 |
"logging_steps": 10,
|
|
|
|
| 33773 |
"attributes": {}
|
| 33774 |
}
|
| 33775 |
},
|
| 33776 |
+
"total_flos": 8.45238591553536e+17,
|
| 33777 |
"train_batch_size": 128,
|
| 33778 |
"trial_name": null,
|
| 33779 |
"trial_params": null
|