Training in progress, step 8000, 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 328277848
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:87f17ddb27c78df3df9ebbdcf34ba0e534bed6b159e38ed164a359bf939c519b
|
| 3 |
size 328277848
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 318646859
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2d926ca500ef14515d2d612de8ff61253060acb30bb272798b045d8da75c1e72
|
| 3 |
size 318646859
|
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 sha256:329a377c90ca49d3bcb8c01bcb7bdf9bc769af05915d36720b3201a9c222f867
|
| 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 sha256:578bef6269d270c9ba7be042609ff28604e2fee3538e234c365c9aa652e62f33
|
| 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": 1.
|
| 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,
|
|
@@ -5385,6 +5385,364 @@
|
|
| 5385 |
"eval_samples_per_second": 270.066,
|
| 5386 |
"eval_steps_per_second": 5.671,
|
| 5387 |
"step": 7500
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 5388 |
}
|
| 5389 |
],
|
| 5390 |
"logging_steps": 10,
|
|
@@ -5404,7 +5762,7 @@
|
|
| 5404 |
"attributes": {}
|
| 5405 |
}
|
| 5406 |
},
|
| 5407 |
-
"total_flos": 2.
|
| 5408 |
"train_batch_size": 48,
|
| 5409 |
"trial_name": null,
|
| 5410 |
"trial_params": null
|
|
|
|
| 2 |
"best_global_step": null,
|
| 3 |
"best_metric": null,
|
| 4 |
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 1.3515796587261362,
|
| 6 |
"eval_steps": 500,
|
| 7 |
+
"global_step": 8000,
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
|
|
|
| 5385 |
"eval_samples_per_second": 270.066,
|
| 5386 |
"eval_steps_per_second": 5.671,
|
| 5387 |
"step": 7500
|
| 5388 |
+
},
|
| 5389 |
+
{
|
| 5390 |
+
"epoch": 1.2687954046291603,
|
| 5391 |
+
"grad_norm": 0.4743061661720276,
|
| 5392 |
+
"learning_rate": 0.00012190597732468595,
|
| 5393 |
+
"loss": 4.406225204467773,
|
| 5394 |
+
"step": 7510
|
| 5395 |
+
},
|
| 5396 |
+
{
|
| 5397 |
+
"epoch": 1.270484879202568,
|
| 5398 |
+
"grad_norm": 0.47182679176330566,
|
| 5399 |
+
"learning_rate": 0.00012143559904956533,
|
| 5400 |
+
"loss": 4.388427734375,
|
| 5401 |
+
"step": 7520
|
| 5402 |
+
},
|
| 5403 |
+
{
|
| 5404 |
+
"epoch": 1.2721743537759758,
|
| 5405 |
+
"grad_norm": 0.4888850152492523,
|
| 5406 |
+
"learning_rate": 0.00012096551205457511,
|
| 5407 |
+
"loss": 4.383931350708008,
|
| 5408 |
+
"step": 7530
|
| 5409 |
+
},
|
| 5410 |
+
{
|
| 5411 |
+
"epoch": 1.2738638283493833,
|
| 5412 |
+
"grad_norm": 0.49569153785705566,
|
| 5413 |
+
"learning_rate": 0.00012049572113333949,
|
| 5414 |
+
"loss": 4.372357177734375,
|
| 5415 |
+
"step": 7540
|
| 5416 |
+
},
|
| 5417 |
+
{
|
| 5418 |
+
"epoch": 1.275553302922791,
|
| 5419 |
+
"grad_norm": 0.47200775146484375,
|
| 5420 |
+
"learning_rate": 0.00012002623107646327,
|
| 5421 |
+
"loss": 4.393280410766602,
|
| 5422 |
+
"step": 7550
|
| 5423 |
+
},
|
| 5424 |
+
{
|
| 5425 |
+
"epoch": 1.2772427774961987,
|
| 5426 |
+
"grad_norm": 0.4849415123462677,
|
| 5427 |
+
"learning_rate": 0.00011955704667148361,
|
| 5428 |
+
"loss": 4.397673416137695,
|
| 5429 |
+
"step": 7560
|
| 5430 |
+
},
|
| 5431 |
+
{
|
| 5432 |
+
"epoch": 1.2789322520696063,
|
| 5433 |
+
"grad_norm": 0.5245839357376099,
|
| 5434 |
+
"learning_rate": 0.00011908817270282048,
|
| 5435 |
+
"loss": 4.423194122314453,
|
| 5436 |
+
"step": 7570
|
| 5437 |
+
},
|
| 5438 |
+
{
|
| 5439 |
+
"epoch": 1.280621726643014,
|
| 5440 |
+
"grad_norm": 0.47324326634407043,
|
| 5441 |
+
"learning_rate": 0.00011861961395172844,
|
| 5442 |
+
"loss": 4.417614364624024,
|
| 5443 |
+
"step": 7580
|
| 5444 |
+
},
|
| 5445 |
+
{
|
| 5446 |
+
"epoch": 1.2823112012164217,
|
| 5447 |
+
"grad_norm": 0.46404823660850525,
|
| 5448 |
+
"learning_rate": 0.00011815137519624767,
|
| 5449 |
+
"loss": 4.41317138671875,
|
| 5450 |
+
"step": 7590
|
| 5451 |
+
},
|
| 5452 |
+
{
|
| 5453 |
+
"epoch": 1.2840006757898292,
|
| 5454 |
+
"grad_norm": 0.46742913126945496,
|
| 5455 |
+
"learning_rate": 0.0001176834612111551,
|
| 5456 |
+
"loss": 4.3837333679199215,
|
| 5457 |
+
"step": 7600
|
| 5458 |
+
},
|
| 5459 |
+
{
|
| 5460 |
+
"epoch": 1.285690150363237,
|
| 5461 |
+
"grad_norm": 0.49143141508102417,
|
| 5462 |
+
"learning_rate": 0.0001172158767679161,
|
| 5463 |
+
"loss": 4.37153205871582,
|
| 5464 |
+
"step": 7610
|
| 5465 |
+
},
|
| 5466 |
+
{
|
| 5467 |
+
"epoch": 1.2873796249366447,
|
| 5468 |
+
"grad_norm": 0.5067402720451355,
|
| 5469 |
+
"learning_rate": 0.00011674862663463538,
|
| 5470 |
+
"loss": 4.377300262451172,
|
| 5471 |
+
"step": 7620
|
| 5472 |
+
},
|
| 5473 |
+
{
|
| 5474 |
+
"epoch": 1.2890690995100524,
|
| 5475 |
+
"grad_norm": 0.4705940783023834,
|
| 5476 |
+
"learning_rate": 0.00011628171557600869,
|
| 5477 |
+
"loss": 4.387155914306641,
|
| 5478 |
+
"step": 7630
|
| 5479 |
+
},
|
| 5480 |
+
{
|
| 5481 |
+
"epoch": 1.2907585740834602,
|
| 5482 |
+
"grad_norm": 0.5150594711303711,
|
| 5483 |
+
"learning_rate": 0.0001158151483532742,
|
| 5484 |
+
"loss": 4.366217041015625,
|
| 5485 |
+
"step": 7640
|
| 5486 |
+
},
|
| 5487 |
+
{
|
| 5488 |
+
"epoch": 1.2924480486568677,
|
| 5489 |
+
"grad_norm": 0.4702792465686798,
|
| 5490 |
+
"learning_rate": 0.00011534892972416382,
|
| 5491 |
+
"loss": 4.3991741180419925,
|
| 5492 |
+
"step": 7650
|
| 5493 |
+
},
|
| 5494 |
+
{
|
| 5495 |
+
"epoch": 1.2941375232302754,
|
| 5496 |
+
"grad_norm": 0.5021132230758667,
|
| 5497 |
+
"learning_rate": 0.00011488306444285465,
|
| 5498 |
+
"loss": 4.408302307128906,
|
| 5499 |
+
"step": 7660
|
| 5500 |
+
},
|
| 5501 |
+
{
|
| 5502 |
+
"epoch": 1.2958269978036832,
|
| 5503 |
+
"grad_norm": 0.46044430136680603,
|
| 5504 |
+
"learning_rate": 0.0001144175572599207,
|
| 5505 |
+
"loss": 4.390703582763672,
|
| 5506 |
+
"step": 7670
|
| 5507 |
+
},
|
| 5508 |
+
{
|
| 5509 |
+
"epoch": 1.2975164723770907,
|
| 5510 |
+
"grad_norm": 0.501556396484375,
|
| 5511 |
+
"learning_rate": 0.00011395241292228435,
|
| 5512 |
+
"loss": 4.350027847290039,
|
| 5513 |
+
"step": 7680
|
| 5514 |
+
},
|
| 5515 |
+
{
|
| 5516 |
+
"epoch": 1.2992059469504984,
|
| 5517 |
+
"grad_norm": 0.4919809103012085,
|
| 5518 |
+
"learning_rate": 0.00011348763617316781,
|
| 5519 |
+
"loss": 4.407309341430664,
|
| 5520 |
+
"step": 7690
|
| 5521 |
+
},
|
| 5522 |
+
{
|
| 5523 |
+
"epoch": 1.3008954215239061,
|
| 5524 |
+
"grad_norm": 0.4790572226047516,
|
| 5525 |
+
"learning_rate": 0.00011302323175204497,
|
| 5526 |
+
"loss": 4.382944107055664,
|
| 5527 |
+
"step": 7700
|
| 5528 |
+
},
|
| 5529 |
+
{
|
| 5530 |
+
"epoch": 1.3025848960973136,
|
| 5531 |
+
"grad_norm": 0.4744700491428375,
|
| 5532 |
+
"learning_rate": 0.00011255920439459302,
|
| 5533 |
+
"loss": 4.376468658447266,
|
| 5534 |
+
"step": 7710
|
| 5535 |
+
},
|
| 5536 |
+
{
|
| 5537 |
+
"epoch": 1.3042743706707214,
|
| 5538 |
+
"grad_norm": 0.4606315791606903,
|
| 5539 |
+
"learning_rate": 0.00011209555883264406,
|
| 5540 |
+
"loss": 4.396706771850586,
|
| 5541 |
+
"step": 7720
|
| 5542 |
+
},
|
| 5543 |
+
{
|
| 5544 |
+
"epoch": 1.3059638452441291,
|
| 5545 |
+
"grad_norm": 0.46377789974212646,
|
| 5546 |
+
"learning_rate": 0.00011163229979413685,
|
| 5547 |
+
"loss": 4.378550720214844,
|
| 5548 |
+
"step": 7730
|
| 5549 |
+
},
|
| 5550 |
+
{
|
| 5551 |
+
"epoch": 1.3076533198175366,
|
| 5552 |
+
"grad_norm": 0.4852358102798462,
|
| 5553 |
+
"learning_rate": 0.00011116943200306871,
|
| 5554 |
+
"loss": 4.369587326049805,
|
| 5555 |
+
"step": 7740
|
| 5556 |
+
},
|
| 5557 |
+
{
|
| 5558 |
+
"epoch": 1.3093427943909444,
|
| 5559 |
+
"grad_norm": 0.47700658440589905,
|
| 5560 |
+
"learning_rate": 0.00011070696017944728,
|
| 5561 |
+
"loss": 4.420570755004883,
|
| 5562 |
+
"step": 7750
|
| 5563 |
+
},
|
| 5564 |
+
{
|
| 5565 |
+
"epoch": 1.311032268964352,
|
| 5566 |
+
"grad_norm": 0.49495404958724976,
|
| 5567 |
+
"learning_rate": 0.00011024488903924235,
|
| 5568 |
+
"loss": 4.396437072753907,
|
| 5569 |
+
"step": 7760
|
| 5570 |
+
},
|
| 5571 |
+
{
|
| 5572 |
+
"epoch": 1.3127217435377598,
|
| 5573 |
+
"grad_norm": 0.4793596863746643,
|
| 5574 |
+
"learning_rate": 0.00010978322329433796,
|
| 5575 |
+
"loss": 4.368120574951172,
|
| 5576 |
+
"step": 7770
|
| 5577 |
+
},
|
| 5578 |
+
{
|
| 5579 |
+
"epoch": 1.3144112181111673,
|
| 5580 |
+
"grad_norm": 0.47642168402671814,
|
| 5581 |
+
"learning_rate": 0.00010932196765248396,
|
| 5582 |
+
"loss": 4.349945449829102,
|
| 5583 |
+
"step": 7780
|
| 5584 |
+
},
|
| 5585 |
+
{
|
| 5586 |
+
"epoch": 1.316100692684575,
|
| 5587 |
+
"grad_norm": 0.470968633890152,
|
| 5588 |
+
"learning_rate": 0.0001088611268172485,
|
| 5589 |
+
"loss": 4.380016326904297,
|
| 5590 |
+
"step": 7790
|
| 5591 |
+
},
|
| 5592 |
+
{
|
| 5593 |
+
"epoch": 1.3177901672579828,
|
| 5594 |
+
"grad_norm": 0.4649656116962433,
|
| 5595 |
+
"learning_rate": 0.00010840070548796967,
|
| 5596 |
+
"loss": 4.385102844238281,
|
| 5597 |
+
"step": 7800
|
| 5598 |
+
},
|
| 5599 |
+
{
|
| 5600 |
+
"epoch": 1.3194796418313905,
|
| 5601 |
+
"grad_norm": 0.4923093318939209,
|
| 5602 |
+
"learning_rate": 0.00010794070835970782,
|
| 5603 |
+
"loss": 4.3910682678222654,
|
| 5604 |
+
"step": 7810
|
| 5605 |
+
},
|
| 5606 |
+
{
|
| 5607 |
+
"epoch": 1.321169116404798,
|
| 5608 |
+
"grad_norm": 0.4957950711250305,
|
| 5609 |
+
"learning_rate": 0.00010748114012319747,
|
| 5610 |
+
"loss": 4.3693492889404295,
|
| 5611 |
+
"step": 7820
|
| 5612 |
+
},
|
| 5613 |
+
{
|
| 5614 |
+
"epoch": 1.3228585909782058,
|
| 5615 |
+
"grad_norm": 0.4974667727947235,
|
| 5616 |
+
"learning_rate": 0.0001070220054647997,
|
| 5617 |
+
"loss": 4.351894760131836,
|
| 5618 |
+
"step": 7830
|
| 5619 |
+
},
|
| 5620 |
+
{
|
| 5621 |
+
"epoch": 1.3245480655516135,
|
| 5622 |
+
"grad_norm": 0.47996777296066284,
|
| 5623 |
+
"learning_rate": 0.00010656330906645422,
|
| 5624 |
+
"loss": 4.377753067016601,
|
| 5625 |
+
"step": 7840
|
| 5626 |
+
},
|
| 5627 |
+
{
|
| 5628 |
+
"epoch": 1.326237540125021,
|
| 5629 |
+
"grad_norm": 0.496206134557724,
|
| 5630 |
+
"learning_rate": 0.00010610505560563163,
|
| 5631 |
+
"loss": 4.364116668701172,
|
| 5632 |
+
"step": 7850
|
| 5633 |
+
},
|
| 5634 |
+
{
|
| 5635 |
+
"epoch": 1.3279270146984288,
|
| 5636 |
+
"grad_norm": 0.4971146881580353,
|
| 5637 |
+
"learning_rate": 0.00010564724975528584,
|
| 5638 |
+
"loss": 4.382867813110352,
|
| 5639 |
+
"step": 7860
|
| 5640 |
+
},
|
| 5641 |
+
{
|
| 5642 |
+
"epoch": 1.3296164892718365,
|
| 5643 |
+
"grad_norm": 0.500912606716156,
|
| 5644 |
+
"learning_rate": 0.00010518989618380632,
|
| 5645 |
+
"loss": 4.386323165893555,
|
| 5646 |
+
"step": 7870
|
| 5647 |
+
},
|
| 5648 |
+
{
|
| 5649 |
+
"epoch": 1.331305963845244,
|
| 5650 |
+
"grad_norm": 0.48914024233818054,
|
| 5651 |
+
"learning_rate": 0.00010473299955497044,
|
| 5652 |
+
"loss": 4.394392013549805,
|
| 5653 |
+
"step": 7880
|
| 5654 |
+
},
|
| 5655 |
+
{
|
| 5656 |
+
"epoch": 1.3329954384186518,
|
| 5657 |
+
"grad_norm": 0.5237129926681519,
|
| 5658 |
+
"learning_rate": 0.000104276564527896,
|
| 5659 |
+
"loss": 4.340487289428711,
|
| 5660 |
+
"step": 7890
|
| 5661 |
+
},
|
| 5662 |
+
{
|
| 5663 |
+
"epoch": 1.3346849129920595,
|
| 5664 |
+
"grad_norm": 0.5235133171081543,
|
| 5665 |
+
"learning_rate": 0.0001038205957569938,
|
| 5666 |
+
"loss": 4.360867309570312,
|
| 5667 |
+
"step": 7900
|
| 5668 |
+
},
|
| 5669 |
+
{
|
| 5670 |
+
"epoch": 1.336374387565467,
|
| 5671 |
+
"grad_norm": 0.506583571434021,
|
| 5672 |
+
"learning_rate": 0.00010336509789191994,
|
| 5673 |
+
"loss": 4.368827056884766,
|
| 5674 |
+
"step": 7910
|
| 5675 |
+
},
|
| 5676 |
+
{
|
| 5677 |
+
"epoch": 1.3380638621388747,
|
| 5678 |
+
"grad_norm": 0.4910382330417633,
|
| 5679 |
+
"learning_rate": 0.00010291007557752861,
|
| 5680 |
+
"loss": 4.372630310058594,
|
| 5681 |
+
"step": 7920
|
| 5682 |
+
},
|
| 5683 |
+
{
|
| 5684 |
+
"epoch": 1.3397533367122825,
|
| 5685 |
+
"grad_norm": 0.4972977340221405,
|
| 5686 |
+
"learning_rate": 0.00010245553345382467,
|
| 5687 |
+
"loss": 4.360449981689453,
|
| 5688 |
+
"step": 7930
|
| 5689 |
+
},
|
| 5690 |
+
{
|
| 5691 |
+
"epoch": 1.3414428112856902,
|
| 5692 |
+
"grad_norm": 0.4779084324836731,
|
| 5693 |
+
"learning_rate": 0.00010200147615591643,
|
| 5694 |
+
"loss": 4.333792877197266,
|
| 5695 |
+
"step": 7940
|
| 5696 |
+
},
|
| 5697 |
+
{
|
| 5698 |
+
"epoch": 1.343132285859098,
|
| 5699 |
+
"grad_norm": 0.48390597105026245,
|
| 5700 |
+
"learning_rate": 0.00010154790831396805,
|
| 5701 |
+
"loss": 4.4007080078125,
|
| 5702 |
+
"step": 7950
|
| 5703 |
+
},
|
| 5704 |
+
{
|
| 5705 |
+
"epoch": 1.3448217604325055,
|
| 5706 |
+
"grad_norm": 0.47715461254119873,
|
| 5707 |
+
"learning_rate": 0.00010109483455315269,
|
| 5708 |
+
"loss": 4.379761123657227,
|
| 5709 |
+
"step": 7960
|
| 5710 |
+
},
|
| 5711 |
+
{
|
| 5712 |
+
"epoch": 1.3465112350059132,
|
| 5713 |
+
"grad_norm": 0.4725069999694824,
|
| 5714 |
+
"learning_rate": 0.00010064225949360525,
|
| 5715 |
+
"loss": 4.364334487915039,
|
| 5716 |
+
"step": 7970
|
| 5717 |
+
},
|
| 5718 |
+
{
|
| 5719 |
+
"epoch": 1.348200709579321,
|
| 5720 |
+
"grad_norm": 0.4733550250530243,
|
| 5721 |
+
"learning_rate": 0.00010019018775037509,
|
| 5722 |
+
"loss": 4.398410034179688,
|
| 5723 |
+
"step": 7980
|
| 5724 |
+
},
|
| 5725 |
+
{
|
| 5726 |
+
"epoch": 1.3498901841527284,
|
| 5727 |
+
"grad_norm": 0.49187690019607544,
|
| 5728 |
+
"learning_rate": 9.973862393337925e-05,
|
| 5729 |
+
"loss": 4.387223815917968,
|
| 5730 |
+
"step": 7990
|
| 5731 |
+
},
|
| 5732 |
+
{
|
| 5733 |
+
"epoch": 1.3515796587261362,
|
| 5734 |
+
"grad_norm": 0.4789150357246399,
|
| 5735 |
+
"learning_rate": 9.928757264735506e-05,
|
| 5736 |
+
"loss": 4.403899383544922,
|
| 5737 |
+
"step": 8000
|
| 5738 |
+
},
|
| 5739 |
+
{
|
| 5740 |
+
"epoch": 1.3515796587261362,
|
| 5741 |
+
"eval_loss": 4.3366875648498535,
|
| 5742 |
+
"eval_runtime": 3.6684,
|
| 5743 |
+
"eval_samples_per_second": 272.602,
|
| 5744 |
+
"eval_steps_per_second": 5.725,
|
| 5745 |
+
"step": 8000
|
| 5746 |
}
|
| 5747 |
],
|
| 5748 |
"logging_steps": 10,
|
|
|
|
| 5762 |
"attributes": {}
|
| 5763 |
}
|
| 5764 |
},
|
| 5765 |
+
"total_flos": 2.6756336633197363e+17,
|
| 5766 |
"train_batch_size": 48,
|
| 5767 |
"trial_name": null,
|
| 5768 |
"trial_params": null
|