Training in progress, step 600, 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 50358592
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ec52b6dac11e4d703e461b57dbd96f63fdad826c54925c92c4a9ed3940c7f925
|
| 3 |
size 50358592
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 25785082
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8f992e9bf2526250ac48863a69bc14c754ff3750575d47e4dfdac1bf35fa0d94
|
| 3 |
size 25785082
|
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 sha256:82f393434236f666cbb4f691051257e303661800a0f9787f5942457107e1ccaa
|
| 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:beb02bcc76a1d18125c30bb2c994848252fc0f0f039db286b9a31c46c82cab52
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": 2.0830085277557373,
|
| 3 |
"best_model_checkpoint": "miner_id_24/checkpoint-400",
|
| 4 |
-
"epoch": 0.
|
| 5 |
"eval_steps": 100,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -3555,6 +3555,714 @@
|
|
| 3555 |
"eval_samples_per_second": 40.76,
|
| 3556 |
"eval_steps_per_second": 10.19,
|
| 3557 |
"step": 500
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3558 |
}
|
| 3559 |
],
|
| 3560 |
"logging_steps": 1,
|
|
@@ -3569,7 +4277,7 @@
|
|
| 3569 |
"early_stopping_threshold": 0.0
|
| 3570 |
},
|
| 3571 |
"attributes": {
|
| 3572 |
-
"early_stopping_patience_counter":
|
| 3573 |
}
|
| 3574 |
},
|
| 3575 |
"TrainerControl": {
|
|
@@ -3578,12 +4286,12 @@
|
|
| 3578 |
"should_evaluate": false,
|
| 3579 |
"should_log": false,
|
| 3580 |
"should_save": true,
|
| 3581 |
-
"should_training_stop":
|
| 3582 |
},
|
| 3583 |
"attributes": {}
|
| 3584 |
}
|
| 3585 |
},
|
| 3586 |
-
"total_flos": 3.
|
| 3587 |
"train_batch_size": 4,
|
| 3588 |
"trial_name": null,
|
| 3589 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": 2.0830085277557373,
|
| 3 |
"best_model_checkpoint": "miner_id_24/checkpoint-400",
|
| 4 |
+
"epoch": 0.07171671896010758,
|
| 5 |
"eval_steps": 100,
|
| 6 |
+
"global_step": 600,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 3555 |
"eval_samples_per_second": 40.76,
|
| 3556 |
"eval_steps_per_second": 10.19,
|
| 3557 |
"step": 500
|
| 3558 |
+
},
|
| 3559 |
+
{
|
| 3560 |
+
"epoch": 0.05988346033168983,
|
| 3561 |
+
"grad_norm": 6.650758743286133,
|
| 3562 |
+
"learning_rate": 0.0002484064591485923,
|
| 3563 |
+
"loss": 16.1606,
|
| 3564 |
+
"step": 501
|
| 3565 |
+
},
|
| 3566 |
+
{
|
| 3567 |
+
"epoch": 0.060002988196623336,
|
| 3568 |
+
"grad_norm": 6.113708972930908,
|
| 3569 |
+
"learning_rate": 0.0002483999754345083,
|
| 3570 |
+
"loss": 17.3134,
|
| 3571 |
+
"step": 502
|
| 3572 |
+
},
|
| 3573 |
+
{
|
| 3574 |
+
"epoch": 0.06012251606155685,
|
| 3575 |
+
"grad_norm": 6.330297946929932,
|
| 3576 |
+
"learning_rate": 0.0002483934786418565,
|
| 3577 |
+
"loss": 16.6002,
|
| 3578 |
+
"step": 503
|
| 3579 |
+
},
|
| 3580 |
+
{
|
| 3581 |
+
"epoch": 0.060242043926490366,
|
| 3582 |
+
"grad_norm": 6.722366809844971,
|
| 3583 |
+
"learning_rate": 0.0002483869687713254,
|
| 3584 |
+
"loss": 16.2453,
|
| 3585 |
+
"step": 504
|
| 3586 |
+
},
|
| 3587 |
+
{
|
| 3588 |
+
"epoch": 0.060361571791423874,
|
| 3589 |
+
"grad_norm": 6.331783771514893,
|
| 3590 |
+
"learning_rate": 0.000248380445823605,
|
| 3591 |
+
"loss": 17.3577,
|
| 3592 |
+
"step": 505
|
| 3593 |
+
},
|
| 3594 |
+
{
|
| 3595 |
+
"epoch": 0.06048109965635739,
|
| 3596 |
+
"grad_norm": 5.92828369140625,
|
| 3597 |
+
"learning_rate": 0.00024837390979938674,
|
| 3598 |
+
"loss": 16.2485,
|
| 3599 |
+
"step": 506
|
| 3600 |
+
},
|
| 3601 |
+
{
|
| 3602 |
+
"epoch": 0.0606006275212909,
|
| 3603 |
+
"grad_norm": 13.758573532104492,
|
| 3604 |
+
"learning_rate": 0.0002483673606993632,
|
| 3605 |
+
"loss": 16.9102,
|
| 3606 |
+
"step": 507
|
| 3607 |
+
},
|
| 3608 |
+
{
|
| 3609 |
+
"epoch": 0.06072015538622441,
|
| 3610 |
+
"grad_norm": 5.667209625244141,
|
| 3611 |
+
"learning_rate": 0.00024836079852422855,
|
| 3612 |
+
"loss": 15.9809,
|
| 3613 |
+
"step": 508
|
| 3614 |
+
},
|
| 3615 |
+
{
|
| 3616 |
+
"epoch": 0.060839683251157926,
|
| 3617 |
+
"grad_norm": 6.230560779571533,
|
| 3618 |
+
"learning_rate": 0.00024835422327467826,
|
| 3619 |
+
"loss": 16.8764,
|
| 3620 |
+
"step": 509
|
| 3621 |
+
},
|
| 3622 |
+
{
|
| 3623 |
+
"epoch": 0.06095921111609144,
|
| 3624 |
+
"grad_norm": 6.252189636230469,
|
| 3625 |
+
"learning_rate": 0.00024834763495140927,
|
| 3626 |
+
"loss": 16.3699,
|
| 3627 |
+
"step": 510
|
| 3628 |
+
},
|
| 3629 |
+
{
|
| 3630 |
+
"epoch": 0.06107873898102495,
|
| 3631 |
+
"grad_norm": 6.850605010986328,
|
| 3632 |
+
"learning_rate": 0.00024834103355511974,
|
| 3633 |
+
"loss": 18.5964,
|
| 3634 |
+
"step": 511
|
| 3635 |
+
},
|
| 3636 |
+
{
|
| 3637 |
+
"epoch": 0.061198266845958464,
|
| 3638 |
+
"grad_norm": 6.297305107116699,
|
| 3639 |
+
"learning_rate": 0.0002483344190865094,
|
| 3640 |
+
"loss": 16.974,
|
| 3641 |
+
"step": 512
|
| 3642 |
+
},
|
| 3643 |
+
{
|
| 3644 |
+
"epoch": 0.06131779471089198,
|
| 3645 |
+
"grad_norm": 6.262433052062988,
|
| 3646 |
+
"learning_rate": 0.00024832779154627927,
|
| 3647 |
+
"loss": 17.2464,
|
| 3648 |
+
"step": 513
|
| 3649 |
+
},
|
| 3650 |
+
{
|
| 3651 |
+
"epoch": 0.06143732257582549,
|
| 3652 |
+
"grad_norm": 6.210182189941406,
|
| 3653 |
+
"learning_rate": 0.00024832115093513177,
|
| 3654 |
+
"loss": 17.2708,
|
| 3655 |
+
"step": 514
|
| 3656 |
+
},
|
| 3657 |
+
{
|
| 3658 |
+
"epoch": 0.061556850440759,
|
| 3659 |
+
"grad_norm": 6.332021236419678,
|
| 3660 |
+
"learning_rate": 0.0002483144972537706,
|
| 3661 |
+
"loss": 17.5522,
|
| 3662 |
+
"step": 515
|
| 3663 |
+
},
|
| 3664 |
+
{
|
| 3665 |
+
"epoch": 0.06167637830569252,
|
| 3666 |
+
"grad_norm": 6.204654216766357,
|
| 3667 |
+
"learning_rate": 0.00024830783050290117,
|
| 3668 |
+
"loss": 17.1688,
|
| 3669 |
+
"step": 516
|
| 3670 |
+
},
|
| 3671 |
+
{
|
| 3672 |
+
"epoch": 0.061795906170626025,
|
| 3673 |
+
"grad_norm": 6.059911251068115,
|
| 3674 |
+
"learning_rate": 0.00024830115068322987,
|
| 3675 |
+
"loss": 16.7284,
|
| 3676 |
+
"step": 517
|
| 3677 |
+
},
|
| 3678 |
+
{
|
| 3679 |
+
"epoch": 0.06191543403555954,
|
| 3680 |
+
"grad_norm": 6.756778240203857,
|
| 3681 |
+
"learning_rate": 0.00024829445779546476,
|
| 3682 |
+
"loss": 17.4355,
|
| 3683 |
+
"step": 518
|
| 3684 |
+
},
|
| 3685 |
+
{
|
| 3686 |
+
"epoch": 0.062034961900493055,
|
| 3687 |
+
"grad_norm": 6.644315719604492,
|
| 3688 |
+
"learning_rate": 0.0002482877518403152,
|
| 3689 |
+
"loss": 17.4694,
|
| 3690 |
+
"step": 519
|
| 3691 |
+
},
|
| 3692 |
+
{
|
| 3693 |
+
"epoch": 0.06215448976542656,
|
| 3694 |
+
"grad_norm": 6.000011920928955,
|
| 3695 |
+
"learning_rate": 0.00024828103281849184,
|
| 3696 |
+
"loss": 16.1155,
|
| 3697 |
+
"step": 520
|
| 3698 |
+
},
|
| 3699 |
+
{
|
| 3700 |
+
"epoch": 0.06227401763036008,
|
| 3701 |
+
"grad_norm": 6.423489570617676,
|
| 3702 |
+
"learning_rate": 0.0002482743007307068,
|
| 3703 |
+
"loss": 17.8717,
|
| 3704 |
+
"step": 521
|
| 3705 |
+
},
|
| 3706 |
+
{
|
| 3707 |
+
"epoch": 0.06239354549529359,
|
| 3708 |
+
"grad_norm": 6.263865947723389,
|
| 3709 |
+
"learning_rate": 0.00024826755557767364,
|
| 3710 |
+
"loss": 16.6036,
|
| 3711 |
+
"step": 522
|
| 3712 |
+
},
|
| 3713 |
+
{
|
| 3714 |
+
"epoch": 0.0625130733602271,
|
| 3715 |
+
"grad_norm": 6.165550231933594,
|
| 3716 |
+
"learning_rate": 0.0002482607973601072,
|
| 3717 |
+
"loss": 16.4604,
|
| 3718 |
+
"step": 523
|
| 3719 |
+
},
|
| 3720 |
+
{
|
| 3721 |
+
"epoch": 0.06263260122516062,
|
| 3722 |
+
"grad_norm": 6.165558338165283,
|
| 3723 |
+
"learning_rate": 0.0002482540260787238,
|
| 3724 |
+
"loss": 17.6127,
|
| 3725 |
+
"step": 524
|
| 3726 |
+
},
|
| 3727 |
+
{
|
| 3728 |
+
"epoch": 0.06275212909009413,
|
| 3729 |
+
"grad_norm": 6.105308532714844,
|
| 3730 |
+
"learning_rate": 0.0002482472417342411,
|
| 3731 |
+
"loss": 16.3942,
|
| 3732 |
+
"step": 525
|
| 3733 |
+
},
|
| 3734 |
+
{
|
| 3735 |
+
"epoch": 0.06287165695502764,
|
| 3736 |
+
"grad_norm": 6.943295955657959,
|
| 3737 |
+
"learning_rate": 0.00024824044432737805,
|
| 3738 |
+
"loss": 16.9056,
|
| 3739 |
+
"step": 526
|
| 3740 |
+
},
|
| 3741 |
+
{
|
| 3742 |
+
"epoch": 0.06299118481996116,
|
| 3743 |
+
"grad_norm": 6.2281365394592285,
|
| 3744 |
+
"learning_rate": 0.00024823363385885515,
|
| 3745 |
+
"loss": 16.3102,
|
| 3746 |
+
"step": 527
|
| 3747 |
+
},
|
| 3748 |
+
{
|
| 3749 |
+
"epoch": 0.06311071268489467,
|
| 3750 |
+
"grad_norm": 6.561606407165527,
|
| 3751 |
+
"learning_rate": 0.0002482268103293942,
|
| 3752 |
+
"loss": 17.055,
|
| 3753 |
+
"step": 528
|
| 3754 |
+
},
|
| 3755 |
+
{
|
| 3756 |
+
"epoch": 0.06323024054982818,
|
| 3757 |
+
"grad_norm": 6.460758209228516,
|
| 3758 |
+
"learning_rate": 0.0002482199737397184,
|
| 3759 |
+
"loss": 15.1712,
|
| 3760 |
+
"step": 529
|
| 3761 |
+
},
|
| 3762 |
+
{
|
| 3763 |
+
"epoch": 0.0633497684147617,
|
| 3764 |
+
"grad_norm": 6.111406326293945,
|
| 3765 |
+
"learning_rate": 0.0002482131240905523,
|
| 3766 |
+
"loss": 15.0887,
|
| 3767 |
+
"step": 530
|
| 3768 |
+
},
|
| 3769 |
+
{
|
| 3770 |
+
"epoch": 0.0634692962796952,
|
| 3771 |
+
"grad_norm": 6.305261135101318,
|
| 3772 |
+
"learning_rate": 0.0002482062613826219,
|
| 3773 |
+
"loss": 16.2783,
|
| 3774 |
+
"step": 531
|
| 3775 |
+
},
|
| 3776 |
+
{
|
| 3777 |
+
"epoch": 0.06358882414462871,
|
| 3778 |
+
"grad_norm": 6.447376728057861,
|
| 3779 |
+
"learning_rate": 0.00024819938561665444,
|
| 3780 |
+
"loss": 15.3546,
|
| 3781 |
+
"step": 532
|
| 3782 |
+
},
|
| 3783 |
+
{
|
| 3784 |
+
"epoch": 0.06370835200956224,
|
| 3785 |
+
"grad_norm": 6.626522064208984,
|
| 3786 |
+
"learning_rate": 0.0002481924967933788,
|
| 3787 |
+
"loss": 16.1421,
|
| 3788 |
+
"step": 533
|
| 3789 |
+
},
|
| 3790 |
+
{
|
| 3791 |
+
"epoch": 0.06382787987449574,
|
| 3792 |
+
"grad_norm": 6.10511589050293,
|
| 3793 |
+
"learning_rate": 0.00024818559491352496,
|
| 3794 |
+
"loss": 17.1937,
|
| 3795 |
+
"step": 534
|
| 3796 |
+
},
|
| 3797 |
+
{
|
| 3798 |
+
"epoch": 0.06394740773942925,
|
| 3799 |
+
"grad_norm": 6.697993278503418,
|
| 3800 |
+
"learning_rate": 0.00024817867997782453,
|
| 3801 |
+
"loss": 17.1077,
|
| 3802 |
+
"step": 535
|
| 3803 |
+
},
|
| 3804 |
+
{
|
| 3805 |
+
"epoch": 0.06406693560436277,
|
| 3806 |
+
"grad_norm": 6.638538837432861,
|
| 3807 |
+
"learning_rate": 0.0002481717519870103,
|
| 3808 |
+
"loss": 16.2386,
|
| 3809 |
+
"step": 536
|
| 3810 |
+
},
|
| 3811 |
+
{
|
| 3812 |
+
"epoch": 0.06418646346929628,
|
| 3813 |
+
"grad_norm": 8.689882278442383,
|
| 3814 |
+
"learning_rate": 0.00024816481094181656,
|
| 3815 |
+
"loss": 15.8653,
|
| 3816 |
+
"step": 537
|
| 3817 |
+
},
|
| 3818 |
+
{
|
| 3819 |
+
"epoch": 0.06430599133422979,
|
| 3820 |
+
"grad_norm": 5.992562294006348,
|
| 3821 |
+
"learning_rate": 0.00024815785684297905,
|
| 3822 |
+
"loss": 16.4454,
|
| 3823 |
+
"step": 538
|
| 3824 |
+
},
|
| 3825 |
+
{
|
| 3826 |
+
"epoch": 0.06442551919916331,
|
| 3827 |
+
"grad_norm": 6.0935139656066895,
|
| 3828 |
+
"learning_rate": 0.0002481508896912346,
|
| 3829 |
+
"loss": 15.9363,
|
| 3830 |
+
"step": 539
|
| 3831 |
+
},
|
| 3832 |
+
{
|
| 3833 |
+
"epoch": 0.06454504706409682,
|
| 3834 |
+
"grad_norm": 6.106266021728516,
|
| 3835 |
+
"learning_rate": 0.00024814390948732187,
|
| 3836 |
+
"loss": 16.1408,
|
| 3837 |
+
"step": 540
|
| 3838 |
+
},
|
| 3839 |
+
{
|
| 3840 |
+
"epoch": 0.06466457492903033,
|
| 3841 |
+
"grad_norm": 5.9442830085754395,
|
| 3842 |
+
"learning_rate": 0.00024813691623198046,
|
| 3843 |
+
"loss": 16.451,
|
| 3844 |
+
"step": 541
|
| 3845 |
+
},
|
| 3846 |
+
{
|
| 3847 |
+
"epoch": 0.06478410279396385,
|
| 3848 |
+
"grad_norm": 6.229999542236328,
|
| 3849 |
+
"learning_rate": 0.0002481299099259517,
|
| 3850 |
+
"loss": 15.3861,
|
| 3851 |
+
"step": 542
|
| 3852 |
+
},
|
| 3853 |
+
{
|
| 3854 |
+
"epoch": 0.06490363065889736,
|
| 3855 |
+
"grad_norm": 6.283480167388916,
|
| 3856 |
+
"learning_rate": 0.000248122890569978,
|
| 3857 |
+
"loss": 16.3622,
|
| 3858 |
+
"step": 543
|
| 3859 |
+
},
|
| 3860 |
+
{
|
| 3861 |
+
"epoch": 0.06502315852383087,
|
| 3862 |
+
"grad_norm": 5.996384620666504,
|
| 3863 |
+
"learning_rate": 0.0002481158581648034,
|
| 3864 |
+
"loss": 15.9675,
|
| 3865 |
+
"step": 544
|
| 3866 |
+
},
|
| 3867 |
+
{
|
| 3868 |
+
"epoch": 0.06514268638876439,
|
| 3869 |
+
"grad_norm": 5.98213529586792,
|
| 3870 |
+
"learning_rate": 0.00024810881271117324,
|
| 3871 |
+
"loss": 15.637,
|
| 3872 |
+
"step": 545
|
| 3873 |
+
},
|
| 3874 |
+
{
|
| 3875 |
+
"epoch": 0.0652622142536979,
|
| 3876 |
+
"grad_norm": 6.469746112823486,
|
| 3877 |
+
"learning_rate": 0.0002481017542098342,
|
| 3878 |
+
"loss": 16.3619,
|
| 3879 |
+
"step": 546
|
| 3880 |
+
},
|
| 3881 |
+
{
|
| 3882 |
+
"epoch": 0.0653817421186314,
|
| 3883 |
+
"grad_norm": 6.57551383972168,
|
| 3884 |
+
"learning_rate": 0.0002480946826615344,
|
| 3885 |
+
"loss": 17.0118,
|
| 3886 |
+
"step": 547
|
| 3887 |
+
},
|
| 3888 |
+
{
|
| 3889 |
+
"epoch": 0.06550126998356492,
|
| 3890 |
+
"grad_norm": 6.5282182693481445,
|
| 3891 |
+
"learning_rate": 0.0002480875980670233,
|
| 3892 |
+
"loss": 18.4883,
|
| 3893 |
+
"step": 548
|
| 3894 |
+
},
|
| 3895 |
+
{
|
| 3896 |
+
"epoch": 0.06562079784849843,
|
| 3897 |
+
"grad_norm": 6.027569770812988,
|
| 3898 |
+
"learning_rate": 0.0002480805004270518,
|
| 3899 |
+
"loss": 15.2508,
|
| 3900 |
+
"step": 549
|
| 3901 |
+
},
|
| 3902 |
+
{
|
| 3903 |
+
"epoch": 0.06574032571343194,
|
| 3904 |
+
"grad_norm": 6.519893646240234,
|
| 3905 |
+
"learning_rate": 0.0002480733897423721,
|
| 3906 |
+
"loss": 17.2495,
|
| 3907 |
+
"step": 550
|
| 3908 |
+
},
|
| 3909 |
+
{
|
| 3910 |
+
"epoch": 0.06585985357836546,
|
| 3911 |
+
"grad_norm": 5.971823215484619,
|
| 3912 |
+
"learning_rate": 0.0002480662660137378,
|
| 3913 |
+
"loss": 15.5504,
|
| 3914 |
+
"step": 551
|
| 3915 |
+
},
|
| 3916 |
+
{
|
| 3917 |
+
"epoch": 0.06597938144329897,
|
| 3918 |
+
"grad_norm": 6.29182243347168,
|
| 3919 |
+
"learning_rate": 0.000248059129241904,
|
| 3920 |
+
"loss": 15.4693,
|
| 3921 |
+
"step": 552
|
| 3922 |
+
},
|
| 3923 |
+
{
|
| 3924 |
+
"epoch": 0.06609890930823248,
|
| 3925 |
+
"grad_norm": 6.490065574645996,
|
| 3926 |
+
"learning_rate": 0.000248051979427627,
|
| 3927 |
+
"loss": 15.931,
|
| 3928 |
+
"step": 553
|
| 3929 |
+
},
|
| 3930 |
+
{
|
| 3931 |
+
"epoch": 0.066218437173166,
|
| 3932 |
+
"grad_norm": 6.518825531005859,
|
| 3933 |
+
"learning_rate": 0.00024804481657166467,
|
| 3934 |
+
"loss": 16.3236,
|
| 3935 |
+
"step": 554
|
| 3936 |
+
},
|
| 3937 |
+
{
|
| 3938 |
+
"epoch": 0.06633796503809951,
|
| 3939 |
+
"grad_norm": 6.617817401885986,
|
| 3940 |
+
"learning_rate": 0.0002480376406747761,
|
| 3941 |
+
"loss": 15.8417,
|
| 3942 |
+
"step": 555
|
| 3943 |
+
},
|
| 3944 |
+
{
|
| 3945 |
+
"epoch": 0.06645749290303302,
|
| 3946 |
+
"grad_norm": 6.263620376586914,
|
| 3947 |
+
"learning_rate": 0.0002480304517377218,
|
| 3948 |
+
"loss": 16.4035,
|
| 3949 |
+
"step": 556
|
| 3950 |
+
},
|
| 3951 |
+
{
|
| 3952 |
+
"epoch": 0.06657702076796654,
|
| 3953 |
+
"grad_norm": 6.821305274963379,
|
| 3954 |
+
"learning_rate": 0.0002480232497612638,
|
| 3955 |
+
"loss": 16.9507,
|
| 3956 |
+
"step": 557
|
| 3957 |
+
},
|
| 3958 |
+
{
|
| 3959 |
+
"epoch": 0.06669654863290005,
|
| 3960 |
+
"grad_norm": 6.247030735015869,
|
| 3961 |
+
"learning_rate": 0.0002480160347461653,
|
| 3962 |
+
"loss": 16.5576,
|
| 3963 |
+
"step": 558
|
| 3964 |
+
},
|
| 3965 |
+
{
|
| 3966 |
+
"epoch": 0.06681607649783355,
|
| 3967 |
+
"grad_norm": 6.491495132446289,
|
| 3968 |
+
"learning_rate": 0.00024800880669319106,
|
| 3969 |
+
"loss": 18.7603,
|
| 3970 |
+
"step": 559
|
| 3971 |
+
},
|
| 3972 |
+
{
|
| 3973 |
+
"epoch": 0.06693560436276708,
|
| 3974 |
+
"grad_norm": 6.250596523284912,
|
| 3975 |
+
"learning_rate": 0.0002480015656031071,
|
| 3976 |
+
"loss": 15.7767,
|
| 3977 |
+
"step": 560
|
| 3978 |
+
},
|
| 3979 |
+
{
|
| 3980 |
+
"epoch": 0.06705513222770058,
|
| 3981 |
+
"grad_norm": 6.2819390296936035,
|
| 3982 |
+
"learning_rate": 0.0002479943114766808,
|
| 3983 |
+
"loss": 15.8,
|
| 3984 |
+
"step": 561
|
| 3985 |
+
},
|
| 3986 |
+
{
|
| 3987 |
+
"epoch": 0.06717466009263409,
|
| 3988 |
+
"grad_norm": 5.973819732666016,
|
| 3989 |
+
"learning_rate": 0.0002479870443146811,
|
| 3990 |
+
"loss": 15.6882,
|
| 3991 |
+
"step": 562
|
| 3992 |
+
},
|
| 3993 |
+
{
|
| 3994 |
+
"epoch": 0.06729418795756761,
|
| 3995 |
+
"grad_norm": 6.051755428314209,
|
| 3996 |
+
"learning_rate": 0.0002479797641178782,
|
| 3997 |
+
"loss": 16.4719,
|
| 3998 |
+
"step": 563
|
| 3999 |
+
},
|
| 4000 |
+
{
|
| 4001 |
+
"epoch": 0.06741371582250112,
|
| 4002 |
+
"grad_norm": 6.508502960205078,
|
| 4003 |
+
"learning_rate": 0.00024797247088704366,
|
| 4004 |
+
"loss": 16.9064,
|
| 4005 |
+
"step": 564
|
| 4006 |
+
},
|
| 4007 |
+
{
|
| 4008 |
+
"epoch": 0.06753324368743463,
|
| 4009 |
+
"grad_norm": 6.121264934539795,
|
| 4010 |
+
"learning_rate": 0.0002479651646229505,
|
| 4011 |
+
"loss": 16.2867,
|
| 4012 |
+
"step": 565
|
| 4013 |
+
},
|
| 4014 |
+
{
|
| 4015 |
+
"epoch": 0.06765277155236815,
|
| 4016 |
+
"grad_norm": 6.16292667388916,
|
| 4017 |
+
"learning_rate": 0.00024795784532637296,
|
| 4018 |
+
"loss": 15.8647,
|
| 4019 |
+
"step": 566
|
| 4020 |
+
},
|
| 4021 |
+
{
|
| 4022 |
+
"epoch": 0.06777229941730166,
|
| 4023 |
+
"grad_norm": 6.289575099945068,
|
| 4024 |
+
"learning_rate": 0.0002479505129980869,
|
| 4025 |
+
"loss": 16.5442,
|
| 4026 |
+
"step": 567
|
| 4027 |
+
},
|
| 4028 |
+
{
|
| 4029 |
+
"epoch": 0.06789182728223517,
|
| 4030 |
+
"grad_norm": 6.375883102416992,
|
| 4031 |
+
"learning_rate": 0.0002479431676388694,
|
| 4032 |
+
"loss": 16.143,
|
| 4033 |
+
"step": 568
|
| 4034 |
+
},
|
| 4035 |
+
{
|
| 4036 |
+
"epoch": 0.06801135514716869,
|
| 4037 |
+
"grad_norm": 6.734157562255859,
|
| 4038 |
+
"learning_rate": 0.00024793580924949897,
|
| 4039 |
+
"loss": 16.5816,
|
| 4040 |
+
"step": 569
|
| 4041 |
+
},
|
| 4042 |
+
{
|
| 4043 |
+
"epoch": 0.0681308830121022,
|
| 4044 |
+
"grad_norm": 5.916708469390869,
|
| 4045 |
+
"learning_rate": 0.00024792843783075544,
|
| 4046 |
+
"loss": 15.6337,
|
| 4047 |
+
"step": 570
|
| 4048 |
+
},
|
| 4049 |
+
{
|
| 4050 |
+
"epoch": 0.0682504108770357,
|
| 4051 |
+
"grad_norm": 5.987206935882568,
|
| 4052 |
+
"learning_rate": 0.00024792105338342015,
|
| 4053 |
+
"loss": 16.1654,
|
| 4054 |
+
"step": 571
|
| 4055 |
+
},
|
| 4056 |
+
{
|
| 4057 |
+
"epoch": 0.06836993874196923,
|
| 4058 |
+
"grad_norm": 7.18080997467041,
|
| 4059 |
+
"learning_rate": 0.00024791365590827566,
|
| 4060 |
+
"loss": 17.0761,
|
| 4061 |
+
"step": 572
|
| 4062 |
+
},
|
| 4063 |
+
{
|
| 4064 |
+
"epoch": 0.06848946660690274,
|
| 4065 |
+
"grad_norm": 5.998250961303711,
|
| 4066 |
+
"learning_rate": 0.000247906245406106,
|
| 4067 |
+
"loss": 15.8114,
|
| 4068 |
+
"step": 573
|
| 4069 |
+
},
|
| 4070 |
+
{
|
| 4071 |
+
"epoch": 0.06860899447183624,
|
| 4072 |
+
"grad_norm": 5.987746715545654,
|
| 4073 |
+
"learning_rate": 0.0002478988218776967,
|
| 4074 |
+
"loss": 16.9754,
|
| 4075 |
+
"step": 574
|
| 4076 |
+
},
|
| 4077 |
+
{
|
| 4078 |
+
"epoch": 0.06872852233676977,
|
| 4079 |
+
"grad_norm": 6.771132469177246,
|
| 4080 |
+
"learning_rate": 0.0002478913853238344,
|
| 4081 |
+
"loss": 16.317,
|
| 4082 |
+
"step": 575
|
| 4083 |
+
},
|
| 4084 |
+
{
|
| 4085 |
+
"epoch": 0.06884805020170327,
|
| 4086 |
+
"grad_norm": 6.411704063415527,
|
| 4087 |
+
"learning_rate": 0.00024788393574530726,
|
| 4088 |
+
"loss": 17.1471,
|
| 4089 |
+
"step": 576
|
| 4090 |
+
},
|
| 4091 |
+
{
|
| 4092 |
+
"epoch": 0.06896757806663678,
|
| 4093 |
+
"grad_norm": 6.628751754760742,
|
| 4094 |
+
"learning_rate": 0.0002478764731429049,
|
| 4095 |
+
"loss": 16.4329,
|
| 4096 |
+
"step": 577
|
| 4097 |
+
},
|
| 4098 |
+
{
|
| 4099 |
+
"epoch": 0.0690871059315703,
|
| 4100 |
+
"grad_norm": 7.8661651611328125,
|
| 4101 |
+
"learning_rate": 0.00024786899751741827,
|
| 4102 |
+
"loss": 15.7631,
|
| 4103 |
+
"step": 578
|
| 4104 |
+
},
|
| 4105 |
+
{
|
| 4106 |
+
"epoch": 0.06920663379650381,
|
| 4107 |
+
"grad_norm": 6.639915943145752,
|
| 4108 |
+
"learning_rate": 0.0002478615088696396,
|
| 4109 |
+
"loss": 16.3586,
|
| 4110 |
+
"step": 579
|
| 4111 |
+
},
|
| 4112 |
+
{
|
| 4113 |
+
"epoch": 0.06932616166143732,
|
| 4114 |
+
"grad_norm": 6.35560941696167,
|
| 4115 |
+
"learning_rate": 0.0002478540072003626,
|
| 4116 |
+
"loss": 17.8284,
|
| 4117 |
+
"step": 580
|
| 4118 |
+
},
|
| 4119 |
+
{
|
| 4120 |
+
"epoch": 0.06944568952637084,
|
| 4121 |
+
"grad_norm": 6.45347261428833,
|
| 4122 |
+
"learning_rate": 0.00024784649251038233,
|
| 4123 |
+
"loss": 16.4598,
|
| 4124 |
+
"step": 581
|
| 4125 |
+
},
|
| 4126 |
+
{
|
| 4127 |
+
"epoch": 0.06956521739130435,
|
| 4128 |
+
"grad_norm": 6.136906147003174,
|
| 4129 |
+
"learning_rate": 0.00024783896480049525,
|
| 4130 |
+
"loss": 16.1433,
|
| 4131 |
+
"step": 582
|
| 4132 |
+
},
|
| 4133 |
+
{
|
| 4134 |
+
"epoch": 0.06968474525623786,
|
| 4135 |
+
"grad_norm": 6.355082035064697,
|
| 4136 |
+
"learning_rate": 0.00024783142407149917,
|
| 4137 |
+
"loss": 17.7579,
|
| 4138 |
+
"step": 583
|
| 4139 |
+
},
|
| 4140 |
+
{
|
| 4141 |
+
"epoch": 0.06980427312117138,
|
| 4142 |
+
"grad_norm": 6.28627872467041,
|
| 4143 |
+
"learning_rate": 0.00024782387032419334,
|
| 4144 |
+
"loss": 16.5747,
|
| 4145 |
+
"step": 584
|
| 4146 |
+
},
|
| 4147 |
+
{
|
| 4148 |
+
"epoch": 0.06992380098610489,
|
| 4149 |
+
"grad_norm": 6.3861188888549805,
|
| 4150 |
+
"learning_rate": 0.0002478163035593783,
|
| 4151 |
+
"loss": 16.4684,
|
| 4152 |
+
"step": 585
|
| 4153 |
+
},
|
| 4154 |
+
{
|
| 4155 |
+
"epoch": 0.0700433288510384,
|
| 4156 |
+
"grad_norm": 8.057326316833496,
|
| 4157 |
+
"learning_rate": 0.00024780872377785603,
|
| 4158 |
+
"loss": 16.7268,
|
| 4159 |
+
"step": 586
|
| 4160 |
+
},
|
| 4161 |
+
{
|
| 4162 |
+
"epoch": 0.07016285671597192,
|
| 4163 |
+
"grad_norm": 6.8107805252075195,
|
| 4164 |
+
"learning_rate": 0.0002478011309804298,
|
| 4165 |
+
"loss": 14.9781,
|
| 4166 |
+
"step": 587
|
| 4167 |
+
},
|
| 4168 |
+
{
|
| 4169 |
+
"epoch": 0.07028238458090542,
|
| 4170 |
+
"grad_norm": 7.4561076164245605,
|
| 4171 |
+
"learning_rate": 0.0002477935251679045,
|
| 4172 |
+
"loss": 15.9972,
|
| 4173 |
+
"step": 588
|
| 4174 |
+
},
|
| 4175 |
+
{
|
| 4176 |
+
"epoch": 0.07040191244583893,
|
| 4177 |
+
"grad_norm": 6.438401222229004,
|
| 4178 |
+
"learning_rate": 0.0002477859063410861,
|
| 4179 |
+
"loss": 16.3148,
|
| 4180 |
+
"step": 589
|
| 4181 |
+
},
|
| 4182 |
+
{
|
| 4183 |
+
"epoch": 0.07052144031077245,
|
| 4184 |
+
"grad_norm": 6.0730299949646,
|
| 4185 |
+
"learning_rate": 0.0002477782745007821,
|
| 4186 |
+
"loss": 15.5541,
|
| 4187 |
+
"step": 590
|
| 4188 |
+
},
|
| 4189 |
+
{
|
| 4190 |
+
"epoch": 0.07064096817570596,
|
| 4191 |
+
"grad_norm": 6.609666347503662,
|
| 4192 |
+
"learning_rate": 0.00024777062964780137,
|
| 4193 |
+
"loss": 16.2029,
|
| 4194 |
+
"step": 591
|
| 4195 |
+
},
|
| 4196 |
+
{
|
| 4197 |
+
"epoch": 0.07076049604063947,
|
| 4198 |
+
"grad_norm": 6.8198652267456055,
|
| 4199 |
+
"learning_rate": 0.00024776297178295424,
|
| 4200 |
+
"loss": 17.0991,
|
| 4201 |
+
"step": 592
|
| 4202 |
+
},
|
| 4203 |
+
{
|
| 4204 |
+
"epoch": 0.07088002390557299,
|
| 4205 |
+
"grad_norm": 7.101873874664307,
|
| 4206 |
+
"learning_rate": 0.0002477553009070522,
|
| 4207 |
+
"loss": 15.2356,
|
| 4208 |
+
"step": 593
|
| 4209 |
+
},
|
| 4210 |
+
{
|
| 4211 |
+
"epoch": 0.0709995517705065,
|
| 4212 |
+
"grad_norm": 6.614659786224365,
|
| 4213 |
+
"learning_rate": 0.0002477476170209083,
|
| 4214 |
+
"loss": 16.8781,
|
| 4215 |
+
"step": 594
|
| 4216 |
+
},
|
| 4217 |
+
{
|
| 4218 |
+
"epoch": 0.07111907963544001,
|
| 4219 |
+
"grad_norm": 9.546778678894043,
|
| 4220 |
+
"learning_rate": 0.0002477399201253369,
|
| 4221 |
+
"loss": 18.754,
|
| 4222 |
+
"step": 595
|
| 4223 |
+
},
|
| 4224 |
+
{
|
| 4225 |
+
"epoch": 0.07123860750037353,
|
| 4226 |
+
"grad_norm": 5.819904327392578,
|
| 4227 |
+
"learning_rate": 0.0002477322102211538,
|
| 4228 |
+
"loss": 15.7819,
|
| 4229 |
+
"step": 596
|
| 4230 |
+
},
|
| 4231 |
+
{
|
| 4232 |
+
"epoch": 0.07135813536530704,
|
| 4233 |
+
"grad_norm": 6.660562515258789,
|
| 4234 |
+
"learning_rate": 0.00024772448730917614,
|
| 4235 |
+
"loss": 17.3348,
|
| 4236 |
+
"step": 597
|
| 4237 |
+
},
|
| 4238 |
+
{
|
| 4239 |
+
"epoch": 0.07147766323024055,
|
| 4240 |
+
"grad_norm": 6.330270290374756,
|
| 4241 |
+
"learning_rate": 0.0002477167513902224,
|
| 4242 |
+
"loss": 16.4988,
|
| 4243 |
+
"step": 598
|
| 4244 |
+
},
|
| 4245 |
+
{
|
| 4246 |
+
"epoch": 0.07159719109517407,
|
| 4247 |
+
"grad_norm": 5.982934951782227,
|
| 4248 |
+
"learning_rate": 0.00024770900246511246,
|
| 4249 |
+
"loss": 14.5084,
|
| 4250 |
+
"step": 599
|
| 4251 |
+
},
|
| 4252 |
+
{
|
| 4253 |
+
"epoch": 0.07171671896010758,
|
| 4254 |
+
"grad_norm": 6.04694128036499,
|
| 4255 |
+
"learning_rate": 0.0002477012405346676,
|
| 4256 |
+
"loss": 16.4352,
|
| 4257 |
+
"step": 600
|
| 4258 |
+
},
|
| 4259 |
+
{
|
| 4260 |
+
"epoch": 0.07171671896010758,
|
| 4261 |
+
"eval_loss": 2.0841848850250244,
|
| 4262 |
+
"eval_runtime": 122.6125,
|
| 4263 |
+
"eval_samples_per_second": 40.779,
|
| 4264 |
+
"eval_steps_per_second": 10.195,
|
| 4265 |
+
"step": 600
|
| 4266 |
}
|
| 4267 |
],
|
| 4268 |
"logging_steps": 1,
|
|
|
|
| 4277 |
"early_stopping_threshold": 0.0
|
| 4278 |
},
|
| 4279 |
"attributes": {
|
| 4280 |
+
"early_stopping_patience_counter": 2
|
| 4281 |
}
|
| 4282 |
},
|
| 4283 |
"TrainerControl": {
|
|
|
|
| 4286 |
"should_evaluate": false,
|
| 4287 |
"should_log": false,
|
| 4288 |
"should_save": true,
|
| 4289 |
+
"should_training_stop": true
|
| 4290 |
},
|
| 4291 |
"attributes": {}
|
| 4292 |
}
|
| 4293 |
},
|
| 4294 |
+
"total_flos": 3.716990563988275e+16,
|
| 4295 |
"train_batch_size": 4,
|
| 4296 |
"trial_name": null,
|
| 4297 |
"trial_params": null
|