Training in progress, step 588, 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 639691872
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8fe0fed13538d60e29cd4c40617a904be0d495d20dce3d817763c4605a48f2e2
|
| 3 |
size 639691872
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 325340244
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4a67808c9d84c5e1501e5ef8f2a4e8792e9097f61b1ccda1274e89044232073a
|
| 3 |
size 325340244
|
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:d7732b0a21b71e9a4fdd4bf7a6f9d6298b39c6392652af04122d065c214b3065
|
| 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 sha256:7f9668780ea7cd0efd62652380c15b597f4f52be588b53462349e9a7707101b3
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": 0.0009602023055776954,
|
| 3 |
"best_model_checkpoint": "miner_id_24/checkpoint-500",
|
| 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 +3555,622 @@
|
|
| 3555 |
"eval_samples_per_second": 7.949,
|
| 3556 |
"eval_steps_per_second": 1.99,
|
| 3557 |
"step": 500
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3558 |
}
|
| 3559 |
],
|
| 3560 |
"logging_steps": 1,
|
|
@@ -3578,12 +4194,12 @@
|
|
| 3578 |
"should_evaluate": false,
|
| 3579 |
"should_log": false,
|
| 3580 |
"should_save": true,
|
| 3581 |
-
"should_training_stop":
|
| 3582 |
},
|
| 3583 |
"attributes": {}
|
| 3584 |
}
|
| 3585 |
},
|
| 3586 |
-
"total_flos": 1.
|
| 3587 |
"train_batch_size": 4,
|
| 3588 |
"trial_name": null,
|
| 3589 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": 0.0009602023055776954,
|
| 3 |
"best_model_checkpoint": "miner_id_24/checkpoint-500",
|
| 4 |
+
"epoch": 0.33913294242045566,
|
| 5 |
"eval_steps": 100,
|
| 6 |
+
"global_step": 588,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 3555 |
"eval_samples_per_second": 7.949,
|
| 3556 |
"eval_steps_per_second": 1.99,
|
| 3557 |
"step": 500
|
| 3558 |
+
},
|
| 3559 |
+
{
|
| 3560 |
+
"epoch": 0.2889551091031433,
|
| 3561 |
+
"grad_norm": 0.21408432722091675,
|
| 3562 |
+
"learning_rate": 1.0973492982661793e-05,
|
| 3563 |
+
"loss": 0.0032,
|
| 3564 |
+
"step": 501
|
| 3565 |
+
},
|
| 3566 |
+
{
|
| 3567 |
+
"epoch": 0.28953186580794,
|
| 3568 |
+
"grad_norm": 0.07948876172304153,
|
| 3569 |
+
"learning_rate": 1.0727263266554011e-05,
|
| 3570 |
+
"loss": 0.0017,
|
| 3571 |
+
"step": 502
|
| 3572 |
+
},
|
| 3573 |
+
{
|
| 3574 |
+
"epoch": 0.2901086225127367,
|
| 3575 |
+
"grad_norm": 0.16862964630126953,
|
| 3576 |
+
"learning_rate": 1.0483670864493778e-05,
|
| 3577 |
+
"loss": 0.0019,
|
| 3578 |
+
"step": 503
|
| 3579 |
+
},
|
| 3580 |
+
{
|
| 3581 |
+
"epoch": 0.2906853792175334,
|
| 3582 |
+
"grad_norm": 0.01340674702078104,
|
| 3583 |
+
"learning_rate": 1.024272297273925e-05,
|
| 3584 |
+
"loss": 0.0002,
|
| 3585 |
+
"step": 504
|
| 3586 |
+
},
|
| 3587 |
+
{
|
| 3588 |
+
"epoch": 0.2912621359223301,
|
| 3589 |
+
"grad_norm": 0.0023153014481067657,
|
| 3590 |
+
"learning_rate": 1.0004426709423974e-05,
|
| 3591 |
+
"loss": 0.0001,
|
| 3592 |
+
"step": 505
|
| 3593 |
+
},
|
| 3594 |
+
{
|
| 3595 |
+
"epoch": 0.2918388926271268,
|
| 3596 |
+
"grad_norm": 0.005769818089902401,
|
| 3597 |
+
"learning_rate": 9.768789114346499e-06,
|
| 3598 |
+
"loss": 0.0002,
|
| 3599 |
+
"step": 506
|
| 3600 |
+
},
|
| 3601 |
+
{
|
| 3602 |
+
"epoch": 0.2924156493319235,
|
| 3603 |
+
"grad_norm": 0.019190780818462372,
|
| 3604 |
+
"learning_rate": 9.535817148762461e-06,
|
| 3605 |
+
"loss": 0.0002,
|
| 3606 |
+
"step": 507
|
| 3607 |
+
},
|
| 3608 |
+
{
|
| 3609 |
+
"epoch": 0.29299240603672017,
|
| 3610 |
+
"grad_norm": 1.366793155670166,
|
| 3611 |
+
"learning_rate": 9.305517695178833e-06,
|
| 3612 |
+
"loss": 0.005,
|
| 3613 |
+
"step": 508
|
| 3614 |
+
},
|
| 3615 |
+
{
|
| 3616 |
+
"epoch": 0.29356916274151684,
|
| 3617 |
+
"grad_norm": 0.017835253849625587,
|
| 3618 |
+
"learning_rate": 9.07789755715075e-06,
|
| 3619 |
+
"loss": 0.0003,
|
| 3620 |
+
"step": 509
|
| 3621 |
+
},
|
| 3622 |
+
{
|
| 3623 |
+
"epoch": 0.2941459194463136,
|
| 3624 |
+
"grad_norm": 0.05861745774745941,
|
| 3625 |
+
"learning_rate": 8.85296345908041e-06,
|
| 3626 |
+
"loss": 0.0005,
|
| 3627 |
+
"step": 510
|
| 3628 |
+
},
|
| 3629 |
+
{
|
| 3630 |
+
"epoch": 0.29472267615111025,
|
| 3631 |
+
"grad_norm": 0.4936632812023163,
|
| 3632 |
+
"learning_rate": 8.630722046018458e-06,
|
| 3633 |
+
"loss": 0.0008,
|
| 3634 |
+
"step": 511
|
| 3635 |
+
},
|
| 3636 |
+
{
|
| 3637 |
+
"epoch": 0.29529943285590693,
|
| 3638 |
+
"grad_norm": 0.008170605637133121,
|
| 3639 |
+
"learning_rate": 8.411179883467667e-06,
|
| 3640 |
+
"loss": 0.0003,
|
| 3641 |
+
"step": 512
|
| 3642 |
+
},
|
| 3643 |
+
{
|
| 3644 |
+
"epoch": 0.29587618956070366,
|
| 3645 |
+
"grad_norm": 0.002018228406086564,
|
| 3646 |
+
"learning_rate": 8.194343457188991e-06,
|
| 3647 |
+
"loss": 0.0001,
|
| 3648 |
+
"step": 513
|
| 3649 |
+
},
|
| 3650 |
+
{
|
| 3651 |
+
"epoch": 0.29645294626550034,
|
| 3652 |
+
"grad_norm": 0.1278875768184662,
|
| 3653 |
+
"learning_rate": 7.98021917300993e-06,
|
| 3654 |
+
"loss": 0.001,
|
| 3655 |
+
"step": 514
|
| 3656 |
+
},
|
| 3657 |
+
{
|
| 3658 |
+
"epoch": 0.297029702970297,
|
| 3659 |
+
"grad_norm": 0.18191634118556976,
|
| 3660 |
+
"learning_rate": 7.76881335663534e-06,
|
| 3661 |
+
"loss": 0.0046,
|
| 3662 |
+
"step": 515
|
| 3663 |
+
},
|
| 3664 |
+
{
|
| 3665 |
+
"epoch": 0.29760645967509375,
|
| 3666 |
+
"grad_norm": 0.00487172557041049,
|
| 3667 |
+
"learning_rate": 7.560132253460483e-06,
|
| 3668 |
+
"loss": 0.0002,
|
| 3669 |
+
"step": 516
|
| 3670 |
+
},
|
| 3671 |
+
{
|
| 3672 |
+
"epoch": 0.2981832163798904,
|
| 3673 |
+
"grad_norm": 0.003059436334297061,
|
| 3674 |
+
"learning_rate": 7.354182028386591e-06,
|
| 3675 |
+
"loss": 0.0002,
|
| 3676 |
+
"step": 517
|
| 3677 |
+
},
|
| 3678 |
+
{
|
| 3679 |
+
"epoch": 0.2987599730846871,
|
| 3680 |
+
"grad_norm": 0.4325723946094513,
|
| 3681 |
+
"learning_rate": 7.150968765638743e-06,
|
| 3682 |
+
"loss": 0.0108,
|
| 3683 |
+
"step": 518
|
| 3684 |
+
},
|
| 3685 |
+
{
|
| 3686 |
+
"epoch": 0.2993367297894838,
|
| 3687 |
+
"grad_norm": 0.6904163360595703,
|
| 3688 |
+
"learning_rate": 6.950498468586075e-06,
|
| 3689 |
+
"loss": 0.004,
|
| 3690 |
+
"step": 519
|
| 3691 |
+
},
|
| 3692 |
+
{
|
| 3693 |
+
"epoch": 0.2999134864942805,
|
| 3694 |
+
"grad_norm": 0.02728419378399849,
|
| 3695 |
+
"learning_rate": 6.75277705956443e-06,
|
| 3696 |
+
"loss": 0.0004,
|
| 3697 |
+
"step": 520
|
| 3698 |
+
},
|
| 3699 |
+
{
|
| 3700 |
+
"epoch": 0.3004902431990772,
|
| 3701 |
+
"grad_norm": 0.0645672082901001,
|
| 3702 |
+
"learning_rate": 6.5578103797014455e-06,
|
| 3703 |
+
"loss": 0.0012,
|
| 3704 |
+
"step": 521
|
| 3705 |
+
},
|
| 3706 |
+
{
|
| 3707 |
+
"epoch": 0.30106699990387387,
|
| 3708 |
+
"grad_norm": 0.13279329240322113,
|
| 3709 |
+
"learning_rate": 6.365604188743979e-06,
|
| 3710 |
+
"loss": 0.0008,
|
| 3711 |
+
"step": 522
|
| 3712 |
+
},
|
| 3713 |
+
{
|
| 3714 |
+
"epoch": 0.3016437566086706,
|
| 3715 |
+
"grad_norm": 0.020064158365130424,
|
| 3716 |
+
"learning_rate": 6.176164164887932e-06,
|
| 3717 |
+
"loss": 0.0005,
|
| 3718 |
+
"step": 523
|
| 3719 |
+
},
|
| 3720 |
+
{
|
| 3721 |
+
"epoch": 0.3022205133134673,
|
| 3722 |
+
"grad_norm": 0.02414029650390148,
|
| 3723 |
+
"learning_rate": 5.9894959046105095e-06,
|
| 3724 |
+
"loss": 0.0005,
|
| 3725 |
+
"step": 524
|
| 3726 |
+
},
|
| 3727 |
+
{
|
| 3728 |
+
"epoch": 0.30279727001826395,
|
| 3729 |
+
"grad_norm": 0.11476976424455643,
|
| 3730 |
+
"learning_rate": 5.805604922504859e-06,
|
| 3731 |
+
"loss": 0.0035,
|
| 3732 |
+
"step": 525
|
| 3733 |
+
},
|
| 3734 |
+
{
|
| 3735 |
+
"epoch": 0.30337402672306063,
|
| 3736 |
+
"grad_norm": 0.15488499402999878,
|
| 3737 |
+
"learning_rate": 5.6244966511172505e-06,
|
| 3738 |
+
"loss": 0.006,
|
| 3739 |
+
"step": 526
|
| 3740 |
+
},
|
| 3741 |
+
{
|
| 3742 |
+
"epoch": 0.30395078342785736,
|
| 3743 |
+
"grad_norm": 1.3745477199554443,
|
| 3744 |
+
"learning_rate": 5.446176440786488e-06,
|
| 3745 |
+
"loss": 0.0229,
|
| 3746 |
+
"step": 527
|
| 3747 |
+
},
|
| 3748 |
+
{
|
| 3749 |
+
"epoch": 0.30452754013265404,
|
| 3750 |
+
"grad_norm": 0.003079983638599515,
|
| 3751 |
+
"learning_rate": 5.270649559485907e-06,
|
| 3752 |
+
"loss": 0.0001,
|
| 3753 |
+
"step": 528
|
| 3754 |
+
},
|
| 3755 |
+
{
|
| 3756 |
+
"epoch": 0.3051042968374507,
|
| 3757 |
+
"grad_norm": 0.018386349081993103,
|
| 3758 |
+
"learning_rate": 5.097921192667687e-06,
|
| 3759 |
+
"loss": 0.0004,
|
| 3760 |
+
"step": 529
|
| 3761 |
+
},
|
| 3762 |
+
{
|
| 3763 |
+
"epoch": 0.30568105354224745,
|
| 3764 |
+
"grad_norm": 0.9645707011222839,
|
| 3765 |
+
"learning_rate": 4.92799644310975e-06,
|
| 3766 |
+
"loss": 0.0022,
|
| 3767 |
+
"step": 530
|
| 3768 |
+
},
|
| 3769 |
+
{
|
| 3770 |
+
"epoch": 0.3062578102470441,
|
| 3771 |
+
"grad_norm": 0.054794877767562866,
|
| 3772 |
+
"learning_rate": 4.7608803307649385e-06,
|
| 3773 |
+
"loss": 0.0052,
|
| 3774 |
+
"step": 531
|
| 3775 |
+
},
|
| 3776 |
+
{
|
| 3777 |
+
"epoch": 0.3068345669518408,
|
| 3778 |
+
"grad_norm": 0.18055804073810577,
|
| 3779 |
+
"learning_rate": 4.596577792612755e-06,
|
| 3780 |
+
"loss": 0.0008,
|
| 3781 |
+
"step": 532
|
| 3782 |
+
},
|
| 3783 |
+
{
|
| 3784 |
+
"epoch": 0.3074113236566375,
|
| 3785 |
+
"grad_norm": 0.009832300245761871,
|
| 3786 |
+
"learning_rate": 4.4350936825134805e-06,
|
| 3787 |
+
"loss": 0.0002,
|
| 3788 |
+
"step": 533
|
| 3789 |
+
},
|
| 3790 |
+
{
|
| 3791 |
+
"epoch": 0.3079880803614342,
|
| 3792 |
+
"grad_norm": 0.0028213104233145714,
|
| 3793 |
+
"learning_rate": 4.27643277106482e-06,
|
| 3794 |
+
"loss": 0.0002,
|
| 3795 |
+
"step": 534
|
| 3796 |
+
},
|
| 3797 |
+
{
|
| 3798 |
+
"epoch": 0.3085648370662309,
|
| 3799 |
+
"grad_norm": 0.004487840458750725,
|
| 3800 |
+
"learning_rate": 4.120599745460918e-06,
|
| 3801 |
+
"loss": 0.0002,
|
| 3802 |
+
"step": 535
|
| 3803 |
+
},
|
| 3804 |
+
{
|
| 3805 |
+
"epoch": 0.30914159377102757,
|
| 3806 |
+
"grad_norm": 0.0015976703725755215,
|
| 3807 |
+
"learning_rate": 3.967599209353967e-06,
|
| 3808 |
+
"loss": 0.0001,
|
| 3809 |
+
"step": 536
|
| 3810 |
+
},
|
| 3811 |
+
{
|
| 3812 |
+
"epoch": 0.3097183504758243,
|
| 3813 |
+
"grad_norm": 0.013129237107932568,
|
| 3814 |
+
"learning_rate": 3.817435682718096e-06,
|
| 3815 |
+
"loss": 0.0002,
|
| 3816 |
+
"step": 537
|
| 3817 |
+
},
|
| 3818 |
+
{
|
| 3819 |
+
"epoch": 0.310295107180621,
|
| 3820 |
+
"grad_norm": 0.3533538281917572,
|
| 3821 |
+
"learning_rate": 3.670113601715941e-06,
|
| 3822 |
+
"loss": 0.0187,
|
| 3823 |
+
"step": 538
|
| 3824 |
+
},
|
| 3825 |
+
{
|
| 3826 |
+
"epoch": 0.31087186388541765,
|
| 3827 |
+
"grad_norm": 0.1365794688463211,
|
| 3828 |
+
"learning_rate": 3.525637318567554e-06,
|
| 3829 |
+
"loss": 0.0024,
|
| 3830 |
+
"step": 539
|
| 3831 |
+
},
|
| 3832 |
+
{
|
| 3833 |
+
"epoch": 0.3114486205902144,
|
| 3834 |
+
"grad_norm": 0.0068528540432453156,
|
| 3835 |
+
"learning_rate": 3.384011101421802e-06,
|
| 3836 |
+
"loss": 0.0003,
|
| 3837 |
+
"step": 540
|
| 3838 |
+
},
|
| 3839 |
+
{
|
| 3840 |
+
"epoch": 0.31202537729501106,
|
| 3841 |
+
"grad_norm": 0.006609945558011532,
|
| 3842 |
+
"learning_rate": 3.2452391342303046e-06,
|
| 3843 |
+
"loss": 0.0002,
|
| 3844 |
+
"step": 541
|
| 3845 |
+
},
|
| 3846 |
+
{
|
| 3847 |
+
"epoch": 0.31260213399980774,
|
| 3848 |
+
"grad_norm": 4.070964813232422,
|
| 3849 |
+
"learning_rate": 3.1093255166238176e-06,
|
| 3850 |
+
"loss": 0.0221,
|
| 3851 |
+
"step": 542
|
| 3852 |
+
},
|
| 3853 |
+
{
|
| 3854 |
+
"epoch": 0.3131788907046044,
|
| 3855 |
+
"grad_norm": 0.1819545030593872,
|
| 3856 |
+
"learning_rate": 2.976274263791179e-06,
|
| 3857 |
+
"loss": 0.0043,
|
| 3858 |
+
"step": 543
|
| 3859 |
+
},
|
| 3860 |
+
{
|
| 3861 |
+
"epoch": 0.31375564740940115,
|
| 3862 |
+
"grad_norm": 0.004702473059296608,
|
| 3863 |
+
"learning_rate": 2.8460893063606e-06,
|
| 3864 |
+
"loss": 0.0001,
|
| 3865 |
+
"step": 544
|
| 3866 |
+
},
|
| 3867 |
+
{
|
| 3868 |
+
"epoch": 0.3143324041141978,
|
| 3869 |
+
"grad_norm": 0.1847194880247116,
|
| 3870 |
+
"learning_rate": 2.718774490283593e-06,
|
| 3871 |
+
"loss": 0.0014,
|
| 3872 |
+
"step": 545
|
| 3873 |
+
},
|
| 3874 |
+
{
|
| 3875 |
+
"epoch": 0.3149091608189945,
|
| 3876 |
+
"grad_norm": 0.06856244802474976,
|
| 3877 |
+
"learning_rate": 2.5943335767213304e-06,
|
| 3878 |
+
"loss": 0.0003,
|
| 3879 |
+
"step": 546
|
| 3880 |
+
},
|
| 3881 |
+
{
|
| 3882 |
+
"epoch": 0.31548591752379124,
|
| 3883 |
+
"grad_norm": 0.021140173077583313,
|
| 3884 |
+
"learning_rate": 2.4727702419335864e-06,
|
| 3885 |
+
"loss": 0.0004,
|
| 3886 |
+
"step": 547
|
| 3887 |
+
},
|
| 3888 |
+
{
|
| 3889 |
+
"epoch": 0.3160626742285879,
|
| 3890 |
+
"grad_norm": 0.30309152603149414,
|
| 3891 |
+
"learning_rate": 2.3540880771700803e-06,
|
| 3892 |
+
"loss": 0.0025,
|
| 3893 |
+
"step": 548
|
| 3894 |
+
},
|
| 3895 |
+
{
|
| 3896 |
+
"epoch": 0.3166394309333846,
|
| 3897 |
+
"grad_norm": 0.07376673817634583,
|
| 3898 |
+
"learning_rate": 2.2382905885643844e-06,
|
| 3899 |
+
"loss": 0.0007,
|
| 3900 |
+
"step": 549
|
| 3901 |
+
},
|
| 3902 |
+
{
|
| 3903 |
+
"epoch": 0.31721618763818127,
|
| 3904 |
+
"grad_norm": 0.012836321257054806,
|
| 3905 |
+
"learning_rate": 2.125381197030374e-06,
|
| 3906 |
+
"loss": 0.0001,
|
| 3907 |
+
"step": 550
|
| 3908 |
+
},
|
| 3909 |
+
{
|
| 3910 |
+
"epoch": 0.317792944342978,
|
| 3911 |
+
"grad_norm": 0.012486038729548454,
|
| 3912 |
+
"learning_rate": 2.0153632381611498e-06,
|
| 3913 |
+
"loss": 0.0004,
|
| 3914 |
+
"step": 551
|
| 3915 |
+
},
|
| 3916 |
+
{
|
| 3917 |
+
"epoch": 0.3183697010477747,
|
| 3918 |
+
"grad_norm": 2.240330696105957,
|
| 3919 |
+
"learning_rate": 1.908239962130476e-06,
|
| 3920 |
+
"loss": 0.0023,
|
| 3921 |
+
"step": 552
|
| 3922 |
+
},
|
| 3923 |
+
{
|
| 3924 |
+
"epoch": 0.31894645775257136,
|
| 3925 |
+
"grad_norm": 1.6802948713302612,
|
| 3926 |
+
"learning_rate": 1.8040145335968318e-06,
|
| 3927 |
+
"loss": 0.0176,
|
| 3928 |
+
"step": 553
|
| 3929 |
+
},
|
| 3930 |
+
{
|
| 3931 |
+
"epoch": 0.3195232144573681,
|
| 3932 |
+
"grad_norm": 0.055240605026483536,
|
| 3933 |
+
"learning_rate": 1.7026900316098215e-06,
|
| 3934 |
+
"loss": 0.0005,
|
| 3935 |
+
"step": 554
|
| 3936 |
+
},
|
| 3937 |
+
{
|
| 3938 |
+
"epoch": 0.32009997116216476,
|
| 3939 |
+
"grad_norm": 0.0030760413501411676,
|
| 3940 |
+
"learning_rate": 1.6042694495193022e-06,
|
| 3941 |
+
"loss": 0.0002,
|
| 3942 |
+
"step": 555
|
| 3943 |
+
},
|
| 3944 |
+
{
|
| 3945 |
+
"epoch": 0.32067672786696144,
|
| 3946 |
+
"grad_norm": 0.010371128097176552,
|
| 3947 |
+
"learning_rate": 1.5087556948868876e-06,
|
| 3948 |
+
"loss": 0.0001,
|
| 3949 |
+
"step": 556
|
| 3950 |
+
},
|
| 3951 |
+
{
|
| 3952 |
+
"epoch": 0.3212534845717582,
|
| 3953 |
+
"grad_norm": 0.21550880372524261,
|
| 3954 |
+
"learning_rate": 1.4161515894001165e-06,
|
| 3955 |
+
"loss": 0.0005,
|
| 3956 |
+
"step": 557
|
| 3957 |
+
},
|
| 3958 |
+
{
|
| 3959 |
+
"epoch": 0.32183024127655485,
|
| 3960 |
+
"grad_norm": 0.030834214761853218,
|
| 3961 |
+
"learning_rate": 1.3264598687890205e-06,
|
| 3962 |
+
"loss": 0.0004,
|
| 3963 |
+
"step": 558
|
| 3964 |
+
},
|
| 3965 |
+
{
|
| 3966 |
+
"epoch": 0.32240699798135153,
|
| 3967 |
+
"grad_norm": 0.04491027444601059,
|
| 3968 |
+
"learning_rate": 1.2396831827453436e-06,
|
| 3969 |
+
"loss": 0.0013,
|
| 3970 |
+
"step": 559
|
| 3971 |
+
},
|
| 3972 |
+
{
|
| 3973 |
+
"epoch": 0.3229837546861482,
|
| 3974 |
+
"grad_norm": 0.0054609524086117744,
|
| 3975 |
+
"learning_rate": 1.1558240948443044e-06,
|
| 3976 |
+
"loss": 0.0002,
|
| 3977 |
+
"step": 560
|
| 3978 |
+
},
|
| 3979 |
+
{
|
| 3980 |
+
"epoch": 0.32356051139094494,
|
| 3981 |
+
"grad_norm": 0.06006991118192673,
|
| 3982 |
+
"learning_rate": 1.0748850824687795e-06,
|
| 3983 |
+
"loss": 0.0011,
|
| 3984 |
+
"step": 561
|
| 3985 |
+
},
|
| 3986 |
+
{
|
| 3987 |
+
"epoch": 0.3241372680957416,
|
| 3988 |
+
"grad_norm": 0.09065622091293335,
|
| 3989 |
+
"learning_rate": 9.968685367361618e-07,
|
| 3990 |
+
"loss": 0.0012,
|
| 3991 |
+
"step": 562
|
| 3992 |
+
},
|
| 3993 |
+
{
|
| 3994 |
+
"epoch": 0.3247140248005383,
|
| 3995 |
+
"grad_norm": 0.11779052764177322,
|
| 3996 |
+
"learning_rate": 9.217767624277396e-07,
|
| 3997 |
+
"loss": 0.0009,
|
| 3998 |
+
"step": 563
|
| 3999 |
+
},
|
| 4000 |
+
{
|
| 4001 |
+
"epoch": 0.325290781505335,
|
| 4002 |
+
"grad_norm": 0.004247438628226519,
|
| 4003 |
+
"learning_rate": 8.496119779205725e-07,
|
| 4004 |
+
"loss": 0.0002,
|
| 4005 |
+
"step": 564
|
| 4006 |
+
},
|
| 4007 |
+
{
|
| 4008 |
+
"epoch": 0.3258675382101317,
|
| 4009 |
+
"grad_norm": 0.26266857981681824,
|
| 4010 |
+
"learning_rate": 7.803763151219779e-07,
|
| 4011 |
+
"loss": 0.0012,
|
| 4012 |
+
"step": 565
|
| 4013 |
+
},
|
| 4014 |
+
{
|
| 4015 |
+
"epoch": 0.3264442949149284,
|
| 4016 |
+
"grad_norm": 0.17352361977100372,
|
| 4017 |
+
"learning_rate": 7.140718194065032e-07,
|
| 4018 |
+
"loss": 0.0046,
|
| 4019 |
+
"step": 566
|
| 4020 |
+
},
|
| 4021 |
+
{
|
| 4022 |
+
"epoch": 0.32702105161972506,
|
| 4023 |
+
"grad_norm": 0.003937472123652697,
|
| 4024 |
+
"learning_rate": 6.507004495555969e-07,
|
| 4025 |
+
"loss": 0.0002,
|
| 4026 |
+
"step": 567
|
| 4027 |
+
},
|
| 4028 |
+
{
|
| 4029 |
+
"epoch": 0.3275978083245218,
|
| 4030 |
+
"grad_norm": 0.0046991510316729546,
|
| 4031 |
+
"learning_rate": 5.902640776996315e-07,
|
| 4032 |
+
"loss": 0.0002,
|
| 4033 |
+
"step": 568
|
| 4034 |
+
},
|
| 4035 |
+
{
|
| 4036 |
+
"epoch": 0.32817456502931847,
|
| 4037 |
+
"grad_norm": 0.007674611173570156,
|
| 4038 |
+
"learning_rate": 5.327644892626938e-07,
|
| 4039 |
+
"loss": 0.0001,
|
| 4040 |
+
"step": 569
|
| 4041 |
+
},
|
| 4042 |
+
{
|
| 4043 |
+
"epoch": 0.32875132173411514,
|
| 4044 |
+
"grad_norm": 0.011769871227443218,
|
| 4045 |
+
"learning_rate": 4.782033829097587e-07,
|
| 4046 |
+
"loss": 0.0002,
|
| 4047 |
+
"step": 570
|
| 4048 |
+
},
|
| 4049 |
+
{
|
| 4050 |
+
"epoch": 0.3293280784389119,
|
| 4051 |
+
"grad_norm": 0.3156275749206543,
|
| 4052 |
+
"learning_rate": 4.2658237049655323e-07,
|
| 4053 |
+
"loss": 0.0011,
|
| 4054 |
+
"step": 571
|
| 4055 |
+
},
|
| 4056 |
+
{
|
| 4057 |
+
"epoch": 0.32990483514370855,
|
| 4058 |
+
"grad_norm": 0.14030705392360687,
|
| 4059 |
+
"learning_rate": 3.779029770219378e-07,
|
| 4060 |
+
"loss": 0.0006,
|
| 4061 |
+
"step": 572
|
| 4062 |
+
},
|
| 4063 |
+
{
|
| 4064 |
+
"epoch": 0.33048159184850523,
|
| 4065 |
+
"grad_norm": 1.7659718990325928,
|
| 4066 |
+
"learning_rate": 3.3216664058283165e-07,
|
| 4067 |
+
"loss": 0.0134,
|
| 4068 |
+
"step": 573
|
| 4069 |
+
},
|
| 4070 |
+
{
|
| 4071 |
+
"epoch": 0.3310583485533019,
|
| 4072 |
+
"grad_norm": 0.0017080976394936442,
|
| 4073 |
+
"learning_rate": 2.893747123317581e-07,
|
| 4074 |
+
"loss": 0.0001,
|
| 4075 |
+
"step": 574
|
| 4076 |
+
},
|
| 4077 |
+
{
|
| 4078 |
+
"epoch": 0.33163510525809864,
|
| 4079 |
+
"grad_norm": 0.003221493447199464,
|
| 4080 |
+
"learning_rate": 2.4952845643689827e-07,
|
| 4081 |
+
"loss": 0.0001,
|
| 4082 |
+
"step": 575
|
| 4083 |
+
},
|
| 4084 |
+
{
|
| 4085 |
+
"epoch": 0.3322118619628953,
|
| 4086 |
+
"grad_norm": 0.006105966400355101,
|
| 4087 |
+
"learning_rate": 2.1262905004475475e-07,
|
| 4088 |
+
"loss": 0.0001,
|
| 4089 |
+
"step": 576
|
| 4090 |
+
},
|
| 4091 |
+
{
|
| 4092 |
+
"epoch": 0.332788618667692,
|
| 4093 |
+
"grad_norm": 0.9341827034950256,
|
| 4094 |
+
"learning_rate": 1.786775832454013e-07,
|
| 4095 |
+
"loss": 0.0124,
|
| 4096 |
+
"step": 577
|
| 4097 |
+
},
|
| 4098 |
+
{
|
| 4099 |
+
"epoch": 0.3333653753724887,
|
| 4100 |
+
"grad_norm": 0.0012107370421290398,
|
| 4101 |
+
"learning_rate": 1.4767505904021983e-07,
|
| 4102 |
+
"loss": 0.0001,
|
| 4103 |
+
"step": 578
|
| 4104 |
+
},
|
| 4105 |
+
{
|
| 4106 |
+
"epoch": 0.3339421320772854,
|
| 4107 |
+
"grad_norm": 0.004891383461654186,
|
| 4108 |
+
"learning_rate": 1.1962239331234637e-07,
|
| 4109 |
+
"loss": 0.0002,
|
| 4110 |
+
"step": 579
|
| 4111 |
+
},
|
| 4112 |
+
{
|
| 4113 |
+
"epoch": 0.3345188887820821,
|
| 4114 |
+
"grad_norm": 0.006321438588202,
|
| 4115 |
+
"learning_rate": 9.452041479954821e-08,
|
| 4116 |
+
"loss": 0.0003,
|
| 4117 |
+
"step": 580
|
| 4118 |
+
},
|
| 4119 |
+
{
|
| 4120 |
+
"epoch": 0.3350956454868788,
|
| 4121 |
+
"grad_norm": 0.16756032407283783,
|
| 4122 |
+
"learning_rate": 7.236986506978794e-08,
|
| 4123 |
+
"loss": 0.001,
|
| 4124 |
+
"step": 581
|
| 4125 |
+
},
|
| 4126 |
+
{
|
| 4127 |
+
"epoch": 0.3356724021916755,
|
| 4128 |
+
"grad_norm": 0.14708253741264343,
|
| 4129 |
+
"learning_rate": 5.317139849928543e-08,
|
| 4130 |
+
"loss": 0.0062,
|
| 4131 |
+
"step": 582
|
| 4132 |
+
},
|
| 4133 |
+
{
|
| 4134 |
+
"epoch": 0.33624915889647217,
|
| 4135 |
+
"grad_norm": 0.4846436679363251,
|
| 4136 |
+
"learning_rate": 3.692558225317777e-08,
|
| 4137 |
+
"loss": 0.0169,
|
| 4138 |
+
"step": 583
|
| 4139 |
+
},
|
| 4140 |
+
{
|
| 4141 |
+
"epoch": 0.33682591560126884,
|
| 4142 |
+
"grad_norm": 0.0019764131866395473,
|
| 4143 |
+
"learning_rate": 2.363289626882148e-08,
|
| 4144 |
+
"loss": 0.0001,
|
| 4145 |
+
"step": 584
|
| 4146 |
+
},
|
| 4147 |
+
{
|
| 4148 |
+
"epoch": 0.3374026723060656,
|
| 4149 |
+
"grad_norm": 0.016219504177570343,
|
| 4150 |
+
"learning_rate": 1.3293733241537266e-08,
|
| 4151 |
+
"loss": 0.0004,
|
| 4152 |
+
"step": 585
|
| 4153 |
+
},
|
| 4154 |
+
{
|
| 4155 |
+
"epoch": 0.33797942901086225,
|
| 4156 |
+
"grad_norm": 0.052470527589321136,
|
| 4157 |
+
"learning_rate": 5.908398613074795e-09,
|
| 4158 |
+
"loss": 0.0007,
|
| 4159 |
+
"step": 586
|
| 4160 |
+
},
|
| 4161 |
+
{
|
| 4162 |
+
"epoch": 0.33855618571565893,
|
| 4163 |
+
"grad_norm": 0.0014326622476801276,
|
| 4164 |
+
"learning_rate": 1.4771105625421834e-09,
|
| 4165 |
+
"loss": 0.0001,
|
| 4166 |
+
"step": 587
|
| 4167 |
+
},
|
| 4168 |
+
{
|
| 4169 |
+
"epoch": 0.33913294242045566,
|
| 4170 |
+
"grad_norm": 0.028719905763864517,
|
| 4171 |
+
"learning_rate": 0.0,
|
| 4172 |
+
"loss": 0.0004,
|
| 4173 |
+
"step": 588
|
| 4174 |
}
|
| 4175 |
],
|
| 4176 |
"logging_steps": 1,
|
|
|
|
| 4194 |
"should_evaluate": false,
|
| 4195 |
"should_log": false,
|
| 4196 |
"should_save": true,
|
| 4197 |
+
"should_training_stop": true
|
| 4198 |
},
|
| 4199 |
"attributes": {}
|
| 4200 |
}
|
| 4201 |
},
|
| 4202 |
+
"total_flos": 1.243015677103571e+18,
|
| 4203 |
"train_batch_size": 4,
|
| 4204 |
"trial_name": null,
|
| 4205 |
"trial_params": null
|