Training in progress, step 510000
Browse files
last-checkpoint/optimizer.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 995605445
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:68aa498d0e3983775adcecf2a5f78debf331313630e2436769b35a9399923c5c
|
3 |
size 995605445
|
last-checkpoint/rng_state.pth
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 14575
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:397a701b388141e6d39ca7be9b1469b741594cf39c18781ab0d884cb484ca723
|
3 |
size 14575
|
last-checkpoint/scheduler.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 627
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:fa99ca31aedd6429013044fa3d239b87534181d83f6f020dbd10161874bc75c9
|
3 |
size 627
|
last-checkpoint/trainer_state.json
CHANGED
@@ -1,8 +1,8 @@
|
|
1 |
{
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
-
"epoch":
|
5 |
-
"global_step":
|
6 |
"is_hyper_param_search": false,
|
7 |
"is_local_process_zero": true,
|
8 |
"is_world_process_zero": true,
|
@@ -6412,11 +6412,139 @@
|
|
6412 |
"eval_samples_per_second": 166.11,
|
6413 |
"eval_steps_per_second": 20.768,
|
6414 |
"step": 500000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
6415 |
}
|
6416 |
],
|
6417 |
"max_steps": 633540,
|
6418 |
"num_train_epochs": 15,
|
6419 |
-
"total_flos": 1.
|
6420 |
"trial_name": null,
|
6421 |
"trial_params": null
|
6422 |
}
|
|
|
1 |
{
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
+
"epoch": 12.075007102945355,
|
5 |
+
"global_step": 510000,
|
6 |
"is_hyper_param_search": false,
|
7 |
"is_local_process_zero": true,
|
8 |
"is_world_process_zero": true,
|
|
|
6412 |
"eval_samples_per_second": 166.11,
|
6413 |
"eval_steps_per_second": 20.768,
|
6414 |
"step": 500000
|
6415 |
+
},
|
6416 |
+
{
|
6417 |
+
"epoch": 11.85,
|
6418 |
+
"learning_rate": 0.0,
|
6419 |
+
"loss": 2.4292,
|
6420 |
+
"step": 500500
|
6421 |
+
},
|
6422 |
+
{
|
6423 |
+
"epoch": 11.86,
|
6424 |
+
"learning_rate": 0.0,
|
6425 |
+
"loss": 2.4061,
|
6426 |
+
"step": 501000
|
6427 |
+
},
|
6428 |
+
{
|
6429 |
+
"epoch": 11.87,
|
6430 |
+
"learning_rate": 0.0,
|
6431 |
+
"loss": 2.4113,
|
6432 |
+
"step": 501500
|
6433 |
+
},
|
6434 |
+
{
|
6435 |
+
"epoch": 11.89,
|
6436 |
+
"learning_rate": 0.0,
|
6437 |
+
"loss": 2.4295,
|
6438 |
+
"step": 502000
|
6439 |
+
},
|
6440 |
+
{
|
6441 |
+
"epoch": 11.9,
|
6442 |
+
"learning_rate": 0.0,
|
6443 |
+
"loss": 2.4258,
|
6444 |
+
"step": 502500
|
6445 |
+
},
|
6446 |
+
{
|
6447 |
+
"epoch": 11.91,
|
6448 |
+
"learning_rate": 0.0,
|
6449 |
+
"loss": 2.4297,
|
6450 |
+
"step": 503000
|
6451 |
+
},
|
6452 |
+
{
|
6453 |
+
"epoch": 11.92,
|
6454 |
+
"learning_rate": 0.0,
|
6455 |
+
"loss": 2.4361,
|
6456 |
+
"step": 503500
|
6457 |
+
},
|
6458 |
+
{
|
6459 |
+
"epoch": 11.93,
|
6460 |
+
"learning_rate": 0.0,
|
6461 |
+
"loss": 2.4306,
|
6462 |
+
"step": 504000
|
6463 |
+
},
|
6464 |
+
{
|
6465 |
+
"epoch": 11.94,
|
6466 |
+
"learning_rate": 0.0,
|
6467 |
+
"loss": 2.4255,
|
6468 |
+
"step": 504500
|
6469 |
+
},
|
6470 |
+
{
|
6471 |
+
"epoch": 11.96,
|
6472 |
+
"learning_rate": 0.0,
|
6473 |
+
"loss": 2.4262,
|
6474 |
+
"step": 505000
|
6475 |
+
},
|
6476 |
+
{
|
6477 |
+
"epoch": 11.97,
|
6478 |
+
"learning_rate": 0.0,
|
6479 |
+
"loss": 2.4273,
|
6480 |
+
"step": 505500
|
6481 |
+
},
|
6482 |
+
{
|
6483 |
+
"epoch": 11.98,
|
6484 |
+
"learning_rate": 0.0,
|
6485 |
+
"loss": 2.4112,
|
6486 |
+
"step": 506000
|
6487 |
+
},
|
6488 |
+
{
|
6489 |
+
"epoch": 11.99,
|
6490 |
+
"learning_rate": 0.0,
|
6491 |
+
"loss": 2.4283,
|
6492 |
+
"step": 506500
|
6493 |
+
},
|
6494 |
+
{
|
6495 |
+
"epoch": 12.0,
|
6496 |
+
"learning_rate": 0.0,
|
6497 |
+
"loss": 2.4236,
|
6498 |
+
"step": 507000
|
6499 |
+
},
|
6500 |
+
{
|
6501 |
+
"epoch": 12.02,
|
6502 |
+
"learning_rate": 0.0,
|
6503 |
+
"loss": 2.4315,
|
6504 |
+
"step": 507500
|
6505 |
+
},
|
6506 |
+
{
|
6507 |
+
"epoch": 12.03,
|
6508 |
+
"learning_rate": 0.0,
|
6509 |
+
"loss": 2.4321,
|
6510 |
+
"step": 508000
|
6511 |
+
},
|
6512 |
+
{
|
6513 |
+
"epoch": 12.04,
|
6514 |
+
"learning_rate": 0.0,
|
6515 |
+
"loss": 2.4189,
|
6516 |
+
"step": 508500
|
6517 |
+
},
|
6518 |
+
{
|
6519 |
+
"epoch": 12.05,
|
6520 |
+
"learning_rate": 0.0,
|
6521 |
+
"loss": 2.422,
|
6522 |
+
"step": 509000
|
6523 |
+
},
|
6524 |
+
{
|
6525 |
+
"epoch": 12.06,
|
6526 |
+
"learning_rate": 0.0,
|
6527 |
+
"loss": 2.4198,
|
6528 |
+
"step": 509500
|
6529 |
+
},
|
6530 |
+
{
|
6531 |
+
"epoch": 12.08,
|
6532 |
+
"learning_rate": 0.0,
|
6533 |
+
"loss": 2.4139,
|
6534 |
+
"step": 510000
|
6535 |
+
},
|
6536 |
+
{
|
6537 |
+
"epoch": 12.08,
|
6538 |
+
"eval_loss": 3.1522228717803955,
|
6539 |
+
"eval_runtime": 112.9202,
|
6540 |
+
"eval_samples_per_second": 166.241,
|
6541 |
+
"eval_steps_per_second": 20.785,
|
6542 |
+
"step": 510000
|
6543 |
}
|
6544 |
],
|
6545 |
"max_steps": 633540,
|
6546 |
"num_train_epochs": 15,
|
6547 |
+
"total_flos": 1.07592688505088e+17,
|
6548 |
"trial_name": null,
|
6549 |
"trial_params": null
|
6550 |
}
|