Training in progress, step 200, 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 859942080
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:70b40b4ea852faa0dabd88f7ddbc8095331dcaf982cf3b34cc7272211b022508
|
| 3 |
size 859942080
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 90187222
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6973271476cd76cc32089427e31b042ab3cf370ff220d960bc069ac0e8b7e1d7
|
| 3 |
size 90187222
|
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 sha256:092d05c850cba14a4e1067d9540d36872c05fc71b3eeadb4562ba802384222c9
|
| 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 2080
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e75c704c71c99deef040fe407c6f53cc8d33f4439273d19c1681b1ebdfb69672
|
| 3 |
size 2080
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": 1.802311658859253,
|
| 3 |
"best_model_checkpoint": "miner_id_24/checkpoint-100",
|
| 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,
|
|
@@ -723,6 +723,714 @@
|
|
| 723 |
"eval_samples_per_second": 5.695,
|
| 724 |
"eval_steps_per_second": 1.898,
|
| 725 |
"step": 100
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 726 |
}
|
| 727 |
],
|
| 728 |
"logging_steps": 1,
|
|
@@ -737,7 +1445,7 @@
|
|
| 737 |
"early_stopping_threshold": 0.0
|
| 738 |
},
|
| 739 |
"attributes": {
|
| 740 |
-
"early_stopping_patience_counter":
|
| 741 |
}
|
| 742 |
},
|
| 743 |
"TrainerControl": {
|
|
@@ -751,7 +1459,7 @@
|
|
| 751 |
"attributes": {}
|
| 752 |
}
|
| 753 |
},
|
| 754 |
-
"total_flos":
|
| 755 |
"train_batch_size": 3,
|
| 756 |
"trial_name": null,
|
| 757 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": 1.802311658859253,
|
| 3 |
"best_model_checkpoint": "miner_id_24/checkpoint-100",
|
| 4 |
+
"epoch": 0.0014563354231746657,
|
| 5 |
"eval_steps": 100,
|
| 6 |
+
"global_step": 200,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 723 |
"eval_samples_per_second": 5.695,
|
| 724 |
"eval_steps_per_second": 1.898,
|
| 725 |
"step": 100
|
| 726 |
+
},
|
| 727 |
+
{
|
| 728 |
+
"epoch": 0.0007354493887032061,
|
| 729 |
+
"grad_norm": 30.542020797729492,
|
| 730 |
+
"learning_rate": 0.0001999999803956598,
|
| 731 |
+
"loss": 2.7235,
|
| 732 |
+
"step": 101
|
| 733 |
+
},
|
| 734 |
+
{
|
| 735 |
+
"epoch": 0.0007427310658190794,
|
| 736 |
+
"grad_norm": 12.602702140808105,
|
| 737 |
+
"learning_rate": 0.0001999999803956598,
|
| 738 |
+
"loss": 1.4984,
|
| 739 |
+
"step": 102
|
| 740 |
+
},
|
| 741 |
+
{
|
| 742 |
+
"epoch": 0.0007500127429349528,
|
| 743 |
+
"grad_norm": 13.053272247314453,
|
| 744 |
+
"learning_rate": 0.0001999999803956598,
|
| 745 |
+
"loss": 2.4308,
|
| 746 |
+
"step": 103
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 0.0007572944200508261,
|
| 750 |
+
"grad_norm": 11.506697654724121,
|
| 751 |
+
"learning_rate": 0.0001999999803956598,
|
| 752 |
+
"loss": 2.5812,
|
| 753 |
+
"step": 104
|
| 754 |
+
},
|
| 755 |
+
{
|
| 756 |
+
"epoch": 0.0007645760971666995,
|
| 757 |
+
"grad_norm": 10.1392822265625,
|
| 758 |
+
"learning_rate": 0.0001999999803956598,
|
| 759 |
+
"loss": 1.6152,
|
| 760 |
+
"step": 105
|
| 761 |
+
},
|
| 762 |
+
{
|
| 763 |
+
"epoch": 0.0007718577742825728,
|
| 764 |
+
"grad_norm": 14.92361068725586,
|
| 765 |
+
"learning_rate": 0.0001999999803956598,
|
| 766 |
+
"loss": 2.2026,
|
| 767 |
+
"step": 106
|
| 768 |
+
},
|
| 769 |
+
{
|
| 770 |
+
"epoch": 0.0007791394513984461,
|
| 771 |
+
"grad_norm": 8.74295425415039,
|
| 772 |
+
"learning_rate": 0.0001999999803956598,
|
| 773 |
+
"loss": 1.7257,
|
| 774 |
+
"step": 107
|
| 775 |
+
},
|
| 776 |
+
{
|
| 777 |
+
"epoch": 0.0007864211285143195,
|
| 778 |
+
"grad_norm": 9.7953519821167,
|
| 779 |
+
"learning_rate": 0.0001999999803956598,
|
| 780 |
+
"loss": 1.9238,
|
| 781 |
+
"step": 108
|
| 782 |
+
},
|
| 783 |
+
{
|
| 784 |
+
"epoch": 0.0007937028056301927,
|
| 785 |
+
"grad_norm": 11.444181442260742,
|
| 786 |
+
"learning_rate": 0.00019999996584374458,
|
| 787 |
+
"loss": 2.2776,
|
| 788 |
+
"step": 109
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 0.000800984482746066,
|
| 792 |
+
"grad_norm": 12.844846725463867,
|
| 793 |
+
"learning_rate": 0.00019999996584374458,
|
| 794 |
+
"loss": 1.9025,
|
| 795 |
+
"step": 110
|
| 796 |
+
},
|
| 797 |
+
{
|
| 798 |
+
"epoch": 0.0008082661598619394,
|
| 799 |
+
"grad_norm": 7.37370491027832,
|
| 800 |
+
"learning_rate": 0.00019999996584374458,
|
| 801 |
+
"loss": 1.0972,
|
| 802 |
+
"step": 111
|
| 803 |
+
},
|
| 804 |
+
{
|
| 805 |
+
"epoch": 0.0008155478369778127,
|
| 806 |
+
"grad_norm": 6.512350559234619,
|
| 807 |
+
"learning_rate": 0.00019999996584374458,
|
| 808 |
+
"loss": 1.4349,
|
| 809 |
+
"step": 112
|
| 810 |
+
},
|
| 811 |
+
{
|
| 812 |
+
"epoch": 0.0008228295140936861,
|
| 813 |
+
"grad_norm": 14.362659454345703,
|
| 814 |
+
"learning_rate": 0.00019999996584374458,
|
| 815 |
+
"loss": 2.0781,
|
| 816 |
+
"step": 113
|
| 817 |
+
},
|
| 818 |
+
{
|
| 819 |
+
"epoch": 0.0008301111912095594,
|
| 820 |
+
"grad_norm": 10.504934310913086,
|
| 821 |
+
"learning_rate": 0.00019999996584374458,
|
| 822 |
+
"loss": 1.8158,
|
| 823 |
+
"step": 114
|
| 824 |
+
},
|
| 825 |
+
{
|
| 826 |
+
"epoch": 0.0008373928683254327,
|
| 827 |
+
"grad_norm": 9.051297187805176,
|
| 828 |
+
"learning_rate": 0.00019999996584374458,
|
| 829 |
+
"loss": 1.393,
|
| 830 |
+
"step": 115
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 0.0008446745454413061,
|
| 834 |
+
"grad_norm": 8.922006607055664,
|
| 835 |
+
"learning_rate": 0.00019999996584374458,
|
| 836 |
+
"loss": 1.5453,
|
| 837 |
+
"step": 116
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 0.0008519562225571794,
|
| 841 |
+
"grad_norm": 10.427721977233887,
|
| 842 |
+
"learning_rate": 0.00019999996584374458,
|
| 843 |
+
"loss": 1.8282,
|
| 844 |
+
"step": 117
|
| 845 |
+
},
|
| 846 |
+
{
|
| 847 |
+
"epoch": 0.0008592378996730527,
|
| 848 |
+
"grad_norm": 7.2870917320251465,
|
| 849 |
+
"learning_rate": 0.00019999996584374458,
|
| 850 |
+
"loss": 1.0126,
|
| 851 |
+
"step": 118
|
| 852 |
+
},
|
| 853 |
+
{
|
| 854 |
+
"epoch": 0.000866519576788926,
|
| 855 |
+
"grad_norm": 11.430087089538574,
|
| 856 |
+
"learning_rate": 0.00019999996584374458,
|
| 857 |
+
"loss": 1.6173,
|
| 858 |
+
"step": 119
|
| 859 |
+
},
|
| 860 |
+
{
|
| 861 |
+
"epoch": 0.0008738012539047993,
|
| 862 |
+
"grad_norm": 8.79759407043457,
|
| 863 |
+
"learning_rate": 0.00019999996584374458,
|
| 864 |
+
"loss": 1.2973,
|
| 865 |
+
"step": 120
|
| 866 |
+
},
|
| 867 |
+
{
|
| 868 |
+
"epoch": 0.0008810829310206727,
|
| 869 |
+
"grad_norm": 27.648391723632812,
|
| 870 |
+
"learning_rate": 0.00019999996584374458,
|
| 871 |
+
"loss": 1.5452,
|
| 872 |
+
"step": 121
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 0.000888364608136546,
|
| 876 |
+
"grad_norm": 6.860217094421387,
|
| 877 |
+
"learning_rate": 0.00019999996584374458,
|
| 878 |
+
"loss": 1.1589,
|
| 879 |
+
"step": 122
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"epoch": 0.0008956462852524193,
|
| 883 |
+
"grad_norm": 9.904220581054688,
|
| 884 |
+
"learning_rate": 0.00019999996584374458,
|
| 885 |
+
"loss": 1.7574,
|
| 886 |
+
"step": 123
|
| 887 |
+
},
|
| 888 |
+
{
|
| 889 |
+
"epoch": 0.0009029279623682927,
|
| 890 |
+
"grad_norm": 14.817357063293457,
|
| 891 |
+
"learning_rate": 0.00019999996584374458,
|
| 892 |
+
"loss": 2.2985,
|
| 893 |
+
"step": 124
|
| 894 |
+
},
|
| 895 |
+
{
|
| 896 |
+
"epoch": 0.000910209639484166,
|
| 897 |
+
"grad_norm": 9.02469253540039,
|
| 898 |
+
"learning_rate": 0.00019999996584374458,
|
| 899 |
+
"loss": 2.0967,
|
| 900 |
+
"step": 125
|
| 901 |
+
},
|
| 902 |
+
{
|
| 903 |
+
"epoch": 0.0009174913166000394,
|
| 904 |
+
"grad_norm": 9.622922897338867,
|
| 905 |
+
"learning_rate": 0.00019999996584374458,
|
| 906 |
+
"loss": 2.4315,
|
| 907 |
+
"step": 126
|
| 908 |
+
},
|
| 909 |
+
{
|
| 910 |
+
"epoch": 0.0009247729937159126,
|
| 911 |
+
"grad_norm": 7.905317306518555,
|
| 912 |
+
"learning_rate": 0.00019999996584374458,
|
| 913 |
+
"loss": 1.3262,
|
| 914 |
+
"step": 127
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 0.0009320546708317859,
|
| 918 |
+
"grad_norm": 8.914352416992188,
|
| 919 |
+
"learning_rate": 0.00019999996584374458,
|
| 920 |
+
"loss": 1.7299,
|
| 921 |
+
"step": 128
|
| 922 |
+
},
|
| 923 |
+
{
|
| 924 |
+
"epoch": 0.0009393363479476593,
|
| 925 |
+
"grad_norm": 9.581490516662598,
|
| 926 |
+
"learning_rate": 0.00019999996584374458,
|
| 927 |
+
"loss": 1.8589,
|
| 928 |
+
"step": 129
|
| 929 |
+
},
|
| 930 |
+
{
|
| 931 |
+
"epoch": 0.0009466180250635326,
|
| 932 |
+
"grad_norm": 6.935079097747803,
|
| 933 |
+
"learning_rate": 0.00019999996584374458,
|
| 934 |
+
"loss": 1.1753,
|
| 935 |
+
"step": 130
|
| 936 |
+
},
|
| 937 |
+
{
|
| 938 |
+
"epoch": 0.000953899702179406,
|
| 939 |
+
"grad_norm": 6.804558753967285,
|
| 940 |
+
"learning_rate": 0.00019999996584374458,
|
| 941 |
+
"loss": 1.62,
|
| 942 |
+
"step": 131
|
| 943 |
+
},
|
| 944 |
+
{
|
| 945 |
+
"epoch": 0.0009611813792952793,
|
| 946 |
+
"grad_norm": 11.983720779418945,
|
| 947 |
+
"learning_rate": 0.00019999996584374458,
|
| 948 |
+
"loss": 1.5915,
|
| 949 |
+
"step": 132
|
| 950 |
+
},
|
| 951 |
+
{
|
| 952 |
+
"epoch": 0.0009684630564111526,
|
| 953 |
+
"grad_norm": 25.60173988342285,
|
| 954 |
+
"learning_rate": 0.00019999996584374458,
|
| 955 |
+
"loss": 2.0053,
|
| 956 |
+
"step": 133
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 0.000975744733527026,
|
| 960 |
+
"grad_norm": 8.45429515838623,
|
| 961 |
+
"learning_rate": 0.00019999996584374458,
|
| 962 |
+
"loss": 2.0073,
|
| 963 |
+
"step": 134
|
| 964 |
+
},
|
| 965 |
+
{
|
| 966 |
+
"epoch": 0.0009830264106428994,
|
| 967 |
+
"grad_norm": 6.321603775024414,
|
| 968 |
+
"learning_rate": 0.00019999996584374458,
|
| 969 |
+
"loss": 0.9744,
|
| 970 |
+
"step": 135
|
| 971 |
+
},
|
| 972 |
+
{
|
| 973 |
+
"epoch": 0.0009903080877587725,
|
| 974 |
+
"grad_norm": 9.055106163024902,
|
| 975 |
+
"learning_rate": 0.00019999996584374458,
|
| 976 |
+
"loss": 1.6416,
|
| 977 |
+
"step": 136
|
| 978 |
+
},
|
| 979 |
+
{
|
| 980 |
+
"epoch": 0.000997589764874646,
|
| 981 |
+
"grad_norm": 8.92955207824707,
|
| 982 |
+
"learning_rate": 0.00019999996584374458,
|
| 983 |
+
"loss": 2.0005,
|
| 984 |
+
"step": 137
|
| 985 |
+
},
|
| 986 |
+
{
|
| 987 |
+
"epoch": 0.0010048714419905193,
|
| 988 |
+
"grad_norm": 8.891891479492188,
|
| 989 |
+
"learning_rate": 0.00019999996584374458,
|
| 990 |
+
"loss": 1.6211,
|
| 991 |
+
"step": 138
|
| 992 |
+
},
|
| 993 |
+
{
|
| 994 |
+
"epoch": 0.0010121531191063925,
|
| 995 |
+
"grad_norm": 13.83609676361084,
|
| 996 |
+
"learning_rate": 0.00019999996584374458,
|
| 997 |
+
"loss": 1.6726,
|
| 998 |
+
"step": 139
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 0.001019434796222266,
|
| 1002 |
+
"grad_norm": 16.78993034362793,
|
| 1003 |
+
"learning_rate": 0.00019999996584374458,
|
| 1004 |
+
"loss": 1.86,
|
| 1005 |
+
"step": 140
|
| 1006 |
+
},
|
| 1007 |
+
{
|
| 1008 |
+
"epoch": 0.0010267164733381393,
|
| 1009 |
+
"grad_norm": 10.312461853027344,
|
| 1010 |
+
"learning_rate": 0.00019999996584374458,
|
| 1011 |
+
"loss": 1.4386,
|
| 1012 |
+
"step": 141
|
| 1013 |
+
},
|
| 1014 |
+
{
|
| 1015 |
+
"epoch": 0.0010339981504540125,
|
| 1016 |
+
"grad_norm": 8.104599952697754,
|
| 1017 |
+
"learning_rate": 0.00019999996584374458,
|
| 1018 |
+
"loss": 2.0064,
|
| 1019 |
+
"step": 142
|
| 1020 |
+
},
|
| 1021 |
+
{
|
| 1022 |
+
"epoch": 0.0010412798275698859,
|
| 1023 |
+
"grad_norm": 9.345172882080078,
|
| 1024 |
+
"learning_rate": 0.00019999996584374458,
|
| 1025 |
+
"loss": 1.9574,
|
| 1026 |
+
"step": 143
|
| 1027 |
+
},
|
| 1028 |
+
{
|
| 1029 |
+
"epoch": 0.0010485615046857593,
|
| 1030 |
+
"grad_norm": 8.887224197387695,
|
| 1031 |
+
"learning_rate": 0.00019999996584374458,
|
| 1032 |
+
"loss": 1.4813,
|
| 1033 |
+
"step": 144
|
| 1034 |
+
},
|
| 1035 |
+
{
|
| 1036 |
+
"epoch": 0.0010558431818016324,
|
| 1037 |
+
"grad_norm": 10.085160255432129,
|
| 1038 |
+
"learning_rate": 0.00019999996584374458,
|
| 1039 |
+
"loss": 1.4148,
|
| 1040 |
+
"step": 145
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 0.0010631248589175058,
|
| 1044 |
+
"grad_norm": 15.231815338134766,
|
| 1045 |
+
"learning_rate": 0.00019999996584374458,
|
| 1046 |
+
"loss": 2.1894,
|
| 1047 |
+
"step": 146
|
| 1048 |
+
},
|
| 1049 |
+
{
|
| 1050 |
+
"epoch": 0.0010704065360333792,
|
| 1051 |
+
"grad_norm": 9.214299201965332,
|
| 1052 |
+
"learning_rate": 0.00019999996584374458,
|
| 1053 |
+
"loss": 1.2747,
|
| 1054 |
+
"step": 147
|
| 1055 |
+
},
|
| 1056 |
+
{
|
| 1057 |
+
"epoch": 0.0010776882131492526,
|
| 1058 |
+
"grad_norm": 9.717480659484863,
|
| 1059 |
+
"learning_rate": 0.00019999996584374458,
|
| 1060 |
+
"loss": 2.3933,
|
| 1061 |
+
"step": 148
|
| 1062 |
+
},
|
| 1063 |
+
{
|
| 1064 |
+
"epoch": 0.0010849698902651258,
|
| 1065 |
+
"grad_norm": 12.608345031738281,
|
| 1066 |
+
"learning_rate": 0.00019999996584374458,
|
| 1067 |
+
"loss": 2.2024,
|
| 1068 |
+
"step": 149
|
| 1069 |
+
},
|
| 1070 |
+
{
|
| 1071 |
+
"epoch": 0.0010922515673809992,
|
| 1072 |
+
"grad_norm": 11.714061737060547,
|
| 1073 |
+
"learning_rate": 0.00019999995129182935,
|
| 1074 |
+
"loss": 2.1489,
|
| 1075 |
+
"step": 150
|
| 1076 |
+
},
|
| 1077 |
+
{
|
| 1078 |
+
"epoch": 0.0010995332444968726,
|
| 1079 |
+
"grad_norm": 6.733028411865234,
|
| 1080 |
+
"learning_rate": 0.00019999995129182935,
|
| 1081 |
+
"loss": 0.9061,
|
| 1082 |
+
"step": 151
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 0.0011068149216127458,
|
| 1086 |
+
"grad_norm": 9.173693656921387,
|
| 1087 |
+
"learning_rate": 0.00019999995129182935,
|
| 1088 |
+
"loss": 1.2983,
|
| 1089 |
+
"step": 152
|
| 1090 |
+
},
|
| 1091 |
+
{
|
| 1092 |
+
"epoch": 0.0011140965987286192,
|
| 1093 |
+
"grad_norm": 10.551212310791016,
|
| 1094 |
+
"learning_rate": 0.00019999995129182935,
|
| 1095 |
+
"loss": 2.1741,
|
| 1096 |
+
"step": 153
|
| 1097 |
+
},
|
| 1098 |
+
{
|
| 1099 |
+
"epoch": 0.0011213782758444926,
|
| 1100 |
+
"grad_norm": 9.643767356872559,
|
| 1101 |
+
"learning_rate": 0.00019999995129182935,
|
| 1102 |
+
"loss": 1.8785,
|
| 1103 |
+
"step": 154
|
| 1104 |
+
},
|
| 1105 |
+
{
|
| 1106 |
+
"epoch": 0.0011286599529603657,
|
| 1107 |
+
"grad_norm": 8.361176490783691,
|
| 1108 |
+
"learning_rate": 0.00019999995129182935,
|
| 1109 |
+
"loss": 1.7363,
|
| 1110 |
+
"step": 155
|
| 1111 |
+
},
|
| 1112 |
+
{
|
| 1113 |
+
"epoch": 0.0011359416300762391,
|
| 1114 |
+
"grad_norm": 7.118173122406006,
|
| 1115 |
+
"learning_rate": 0.00019999995129182935,
|
| 1116 |
+
"loss": 1.9851,
|
| 1117 |
+
"step": 156
|
| 1118 |
+
},
|
| 1119 |
+
{
|
| 1120 |
+
"epoch": 0.0011432233071921125,
|
| 1121 |
+
"grad_norm": 11.809035301208496,
|
| 1122 |
+
"learning_rate": 0.00019999995129182935,
|
| 1123 |
+
"loss": 2.0862,
|
| 1124 |
+
"step": 157
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 0.001150504984307986,
|
| 1128 |
+
"grad_norm": 8.845592498779297,
|
| 1129 |
+
"learning_rate": 0.00019999995129182935,
|
| 1130 |
+
"loss": 1.2204,
|
| 1131 |
+
"step": 158
|
| 1132 |
+
},
|
| 1133 |
+
{
|
| 1134 |
+
"epoch": 0.001157786661423859,
|
| 1135 |
+
"grad_norm": 8.715829849243164,
|
| 1136 |
+
"learning_rate": 0.00019999995129182935,
|
| 1137 |
+
"loss": 1.9076,
|
| 1138 |
+
"step": 159
|
| 1139 |
+
},
|
| 1140 |
+
{
|
| 1141 |
+
"epoch": 0.0011650683385397325,
|
| 1142 |
+
"grad_norm": 8.711019515991211,
|
| 1143 |
+
"learning_rate": 0.00019999995129182935,
|
| 1144 |
+
"loss": 1.4475,
|
| 1145 |
+
"step": 160
|
| 1146 |
+
},
|
| 1147 |
+
{
|
| 1148 |
+
"epoch": 0.0011723500156556059,
|
| 1149 |
+
"grad_norm": 9.015527725219727,
|
| 1150 |
+
"learning_rate": 0.00019999995129182935,
|
| 1151 |
+
"loss": 1.9736,
|
| 1152 |
+
"step": 161
|
| 1153 |
+
},
|
| 1154 |
+
{
|
| 1155 |
+
"epoch": 0.001179631692771479,
|
| 1156 |
+
"grad_norm": 13.795394897460938,
|
| 1157 |
+
"learning_rate": 0.00019999995129182935,
|
| 1158 |
+
"loss": 2.1726,
|
| 1159 |
+
"step": 162
|
| 1160 |
+
},
|
| 1161 |
+
{
|
| 1162 |
+
"epoch": 0.0011869133698873525,
|
| 1163 |
+
"grad_norm": 8.399027824401855,
|
| 1164 |
+
"learning_rate": 0.00019999995129182935,
|
| 1165 |
+
"loss": 1.5726,
|
| 1166 |
+
"step": 163
|
| 1167 |
+
},
|
| 1168 |
+
{
|
| 1169 |
+
"epoch": 0.0011941950470032259,
|
| 1170 |
+
"grad_norm": 9.823261260986328,
|
| 1171 |
+
"learning_rate": 0.00019999995129182935,
|
| 1172 |
+
"loss": 1.7958,
|
| 1173 |
+
"step": 164
|
| 1174 |
+
},
|
| 1175 |
+
{
|
| 1176 |
+
"epoch": 0.001201476724119099,
|
| 1177 |
+
"grad_norm": 7.4197516441345215,
|
| 1178 |
+
"learning_rate": 0.00019999995129182935,
|
| 1179 |
+
"loss": 1.1734,
|
| 1180 |
+
"step": 165
|
| 1181 |
+
},
|
| 1182 |
+
{
|
| 1183 |
+
"epoch": 0.0012087584012349724,
|
| 1184 |
+
"grad_norm": 9.796939849853516,
|
| 1185 |
+
"learning_rate": 0.00019999993673991412,
|
| 1186 |
+
"loss": 1.5177,
|
| 1187 |
+
"step": 166
|
| 1188 |
+
},
|
| 1189 |
+
{
|
| 1190 |
+
"epoch": 0.0012160400783508458,
|
| 1191 |
+
"grad_norm": 10.06812858581543,
|
| 1192 |
+
"learning_rate": 0.00019999993673991412,
|
| 1193 |
+
"loss": 2.1493,
|
| 1194 |
+
"step": 167
|
| 1195 |
+
},
|
| 1196 |
+
{
|
| 1197 |
+
"epoch": 0.001223321755466719,
|
| 1198 |
+
"grad_norm": 12.96749210357666,
|
| 1199 |
+
"learning_rate": 0.00019999993673991412,
|
| 1200 |
+
"loss": 2.4498,
|
| 1201 |
+
"step": 168
|
| 1202 |
+
},
|
| 1203 |
+
{
|
| 1204 |
+
"epoch": 0.0012306034325825924,
|
| 1205 |
+
"grad_norm": 10.271540641784668,
|
| 1206 |
+
"learning_rate": 0.00019999993673991412,
|
| 1207 |
+
"loss": 2.1217,
|
| 1208 |
+
"step": 169
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 0.0012378851096984658,
|
| 1212 |
+
"grad_norm": 10.417543411254883,
|
| 1213 |
+
"learning_rate": 0.00019999993673991412,
|
| 1214 |
+
"loss": 1.5065,
|
| 1215 |
+
"step": 170
|
| 1216 |
+
},
|
| 1217 |
+
{
|
| 1218 |
+
"epoch": 0.0012451667868143392,
|
| 1219 |
+
"grad_norm": 14.934460639953613,
|
| 1220 |
+
"learning_rate": 0.00019999993673991412,
|
| 1221 |
+
"loss": 1.8384,
|
| 1222 |
+
"step": 171
|
| 1223 |
+
},
|
| 1224 |
+
{
|
| 1225 |
+
"epoch": 0.0012524484639302124,
|
| 1226 |
+
"grad_norm": 29.06182098388672,
|
| 1227 |
+
"learning_rate": 0.00019999993673991412,
|
| 1228 |
+
"loss": 2.2379,
|
| 1229 |
+
"step": 172
|
| 1230 |
+
},
|
| 1231 |
+
{
|
| 1232 |
+
"epoch": 0.0012597301410460858,
|
| 1233 |
+
"grad_norm": 179.1177215576172,
|
| 1234 |
+
"learning_rate": 0.00019999993673991412,
|
| 1235 |
+
"loss": 2.4171,
|
| 1236 |
+
"step": 173
|
| 1237 |
+
},
|
| 1238 |
+
{
|
| 1239 |
+
"epoch": 0.0012670118181619592,
|
| 1240 |
+
"grad_norm": 17.668655395507812,
|
| 1241 |
+
"learning_rate": 0.00019999993673991412,
|
| 1242 |
+
"loss": 2.3388,
|
| 1243 |
+
"step": 174
|
| 1244 |
+
},
|
| 1245 |
+
{
|
| 1246 |
+
"epoch": 0.0012742934952778323,
|
| 1247 |
+
"grad_norm": 8.100573539733887,
|
| 1248 |
+
"learning_rate": 0.00019999993673991412,
|
| 1249 |
+
"loss": 1.3806,
|
| 1250 |
+
"step": 175
|
| 1251 |
+
},
|
| 1252 |
+
{
|
| 1253 |
+
"epoch": 0.0012815751723937057,
|
| 1254 |
+
"grad_norm": 9.43472671508789,
|
| 1255 |
+
"learning_rate": 0.00019999993673991412,
|
| 1256 |
+
"loss": 1.2791,
|
| 1257 |
+
"step": 176
|
| 1258 |
+
},
|
| 1259 |
+
{
|
| 1260 |
+
"epoch": 0.0012888568495095791,
|
| 1261 |
+
"grad_norm": 11.577341079711914,
|
| 1262 |
+
"learning_rate": 0.00019999993673991412,
|
| 1263 |
+
"loss": 1.7206,
|
| 1264 |
+
"step": 177
|
| 1265 |
+
},
|
| 1266 |
+
{
|
| 1267 |
+
"epoch": 0.0012961385266254523,
|
| 1268 |
+
"grad_norm": 12.485160827636719,
|
| 1269 |
+
"learning_rate": 0.00019999993673991412,
|
| 1270 |
+
"loss": 2.4225,
|
| 1271 |
+
"step": 178
|
| 1272 |
+
},
|
| 1273 |
+
{
|
| 1274 |
+
"epoch": 0.0013034202037413257,
|
| 1275 |
+
"grad_norm": 11.283897399902344,
|
| 1276 |
+
"learning_rate": 0.00019999993673991412,
|
| 1277 |
+
"loss": 1.535,
|
| 1278 |
+
"step": 179
|
| 1279 |
+
},
|
| 1280 |
+
{
|
| 1281 |
+
"epoch": 0.001310701880857199,
|
| 1282 |
+
"grad_norm": 9.801483154296875,
|
| 1283 |
+
"learning_rate": 0.00019999993673991412,
|
| 1284 |
+
"loss": 2.1996,
|
| 1285 |
+
"step": 180
|
| 1286 |
+
},
|
| 1287 |
+
{
|
| 1288 |
+
"epoch": 0.0013179835579730723,
|
| 1289 |
+
"grad_norm": 11.930302619934082,
|
| 1290 |
+
"learning_rate": 0.0001999999221879989,
|
| 1291 |
+
"loss": 1.4919,
|
| 1292 |
+
"step": 181
|
| 1293 |
+
},
|
| 1294 |
+
{
|
| 1295 |
+
"epoch": 0.0013252652350889457,
|
| 1296 |
+
"grad_norm": 10.425878524780273,
|
| 1297 |
+
"learning_rate": 0.0001999999221879989,
|
| 1298 |
+
"loss": 1.7397,
|
| 1299 |
+
"step": 182
|
| 1300 |
+
},
|
| 1301 |
+
{
|
| 1302 |
+
"epoch": 0.001332546912204819,
|
| 1303 |
+
"grad_norm": 9.171979904174805,
|
| 1304 |
+
"learning_rate": 0.0001999999221879989,
|
| 1305 |
+
"loss": 1.8727,
|
| 1306 |
+
"step": 183
|
| 1307 |
+
},
|
| 1308 |
+
{
|
| 1309 |
+
"epoch": 0.0013398285893206924,
|
| 1310 |
+
"grad_norm": 10.64809799194336,
|
| 1311 |
+
"learning_rate": 0.0001999999221879989,
|
| 1312 |
+
"loss": 1.742,
|
| 1313 |
+
"step": 184
|
| 1314 |
+
},
|
| 1315 |
+
{
|
| 1316 |
+
"epoch": 0.0013471102664365656,
|
| 1317 |
+
"grad_norm": 13.533452033996582,
|
| 1318 |
+
"learning_rate": 0.0001999999221879989,
|
| 1319 |
+
"loss": 1.7024,
|
| 1320 |
+
"step": 185
|
| 1321 |
+
},
|
| 1322 |
+
{
|
| 1323 |
+
"epoch": 0.001354391943552439,
|
| 1324 |
+
"grad_norm": 9.46273422241211,
|
| 1325 |
+
"learning_rate": 0.0001999999221879989,
|
| 1326 |
+
"loss": 1.8248,
|
| 1327 |
+
"step": 186
|
| 1328 |
+
},
|
| 1329 |
+
{
|
| 1330 |
+
"epoch": 0.0013616736206683124,
|
| 1331 |
+
"grad_norm": 20.131072998046875,
|
| 1332 |
+
"learning_rate": 0.0001999999221879989,
|
| 1333 |
+
"loss": 1.7554,
|
| 1334 |
+
"step": 187
|
| 1335 |
+
},
|
| 1336 |
+
{
|
| 1337 |
+
"epoch": 0.0013689552977841856,
|
| 1338 |
+
"grad_norm": 6.414751052856445,
|
| 1339 |
+
"learning_rate": 0.0001999999221879989,
|
| 1340 |
+
"loss": 0.8933,
|
| 1341 |
+
"step": 188
|
| 1342 |
+
},
|
| 1343 |
+
{
|
| 1344 |
+
"epoch": 0.001376236974900059,
|
| 1345 |
+
"grad_norm": 7.662179946899414,
|
| 1346 |
+
"learning_rate": 0.0001999999221879989,
|
| 1347 |
+
"loss": 1.4111,
|
| 1348 |
+
"step": 189
|
| 1349 |
+
},
|
| 1350 |
+
{
|
| 1351 |
+
"epoch": 0.0013835186520159324,
|
| 1352 |
+
"grad_norm": 10.391175270080566,
|
| 1353 |
+
"learning_rate": 0.0001999999221879989,
|
| 1354 |
+
"loss": 2.0454,
|
| 1355 |
+
"step": 190
|
| 1356 |
+
},
|
| 1357 |
+
{
|
| 1358 |
+
"epoch": 0.0013908003291318056,
|
| 1359 |
+
"grad_norm": 9.538223266601562,
|
| 1360 |
+
"learning_rate": 0.0001999999221879989,
|
| 1361 |
+
"loss": 2.2686,
|
| 1362 |
+
"step": 191
|
| 1363 |
+
},
|
| 1364 |
+
{
|
| 1365 |
+
"epoch": 0.001398082006247679,
|
| 1366 |
+
"grad_norm": 13.454922676086426,
|
| 1367 |
+
"learning_rate": 0.0001999999221879989,
|
| 1368 |
+
"loss": 2.0085,
|
| 1369 |
+
"step": 192
|
| 1370 |
+
},
|
| 1371 |
+
{
|
| 1372 |
+
"epoch": 0.0014053636833635523,
|
| 1373 |
+
"grad_norm": 8.68910026550293,
|
| 1374 |
+
"learning_rate": 0.0001999999221879989,
|
| 1375 |
+
"loss": 1.2429,
|
| 1376 |
+
"step": 193
|
| 1377 |
+
},
|
| 1378 |
+
{
|
| 1379 |
+
"epoch": 0.0014126453604794255,
|
| 1380 |
+
"grad_norm": 11.174205780029297,
|
| 1381 |
+
"learning_rate": 0.00019999990763608366,
|
| 1382 |
+
"loss": 1.5733,
|
| 1383 |
+
"step": 194
|
| 1384 |
+
},
|
| 1385 |
+
{
|
| 1386 |
+
"epoch": 0.001419927037595299,
|
| 1387 |
+
"grad_norm": 6.9463791847229,
|
| 1388 |
+
"learning_rate": 0.00019999990763608366,
|
| 1389 |
+
"loss": 1.1096,
|
| 1390 |
+
"step": 195
|
| 1391 |
+
},
|
| 1392 |
+
{
|
| 1393 |
+
"epoch": 0.0014272087147111723,
|
| 1394 |
+
"grad_norm": 10.863914489746094,
|
| 1395 |
+
"learning_rate": 0.00019999990763608366,
|
| 1396 |
+
"loss": 1.374,
|
| 1397 |
+
"step": 196
|
| 1398 |
+
},
|
| 1399 |
+
{
|
| 1400 |
+
"epoch": 0.0014344903918270457,
|
| 1401 |
+
"grad_norm": 7.314742565155029,
|
| 1402 |
+
"learning_rate": 0.00019999990763608366,
|
| 1403 |
+
"loss": 1.8196,
|
| 1404 |
+
"step": 197
|
| 1405 |
+
},
|
| 1406 |
+
{
|
| 1407 |
+
"epoch": 0.0014417720689429189,
|
| 1408 |
+
"grad_norm": 13.76647663116455,
|
| 1409 |
+
"learning_rate": 0.00019999990763608366,
|
| 1410 |
+
"loss": 1.8654,
|
| 1411 |
+
"step": 198
|
| 1412 |
+
},
|
| 1413 |
+
{
|
| 1414 |
+
"epoch": 0.0014490537460587923,
|
| 1415 |
+
"grad_norm": 18.60045051574707,
|
| 1416 |
+
"learning_rate": 0.00019999990763608366,
|
| 1417 |
+
"loss": 1.8438,
|
| 1418 |
+
"step": 199
|
| 1419 |
+
},
|
| 1420 |
+
{
|
| 1421 |
+
"epoch": 0.0014563354231746657,
|
| 1422 |
+
"grad_norm": 6.76107931137085,
|
| 1423 |
+
"learning_rate": 0.00019999990763608366,
|
| 1424 |
+
"loss": 1.1957,
|
| 1425 |
+
"step": 200
|
| 1426 |
+
},
|
| 1427 |
+
{
|
| 1428 |
+
"epoch": 0.0014563354231746657,
|
| 1429 |
+
"eval_loss": 1.8587133884429932,
|
| 1430 |
+
"eval_runtime": 36.3018,
|
| 1431 |
+
"eval_samples_per_second": 5.702,
|
| 1432 |
+
"eval_steps_per_second": 1.901,
|
| 1433 |
+
"step": 200
|
| 1434 |
}
|
| 1435 |
],
|
| 1436 |
"logging_steps": 1,
|
|
|
|
| 1445 |
"early_stopping_threshold": 0.0
|
| 1446 |
},
|
| 1447 |
"attributes": {
|
| 1448 |
+
"early_stopping_patience_counter": 1
|
| 1449 |
}
|
| 1450 |
},
|
| 1451 |
"TrainerControl": {
|
|
|
|
| 1459 |
"attributes": {}
|
| 1460 |
}
|
| 1461 |
},
|
| 1462 |
+
"total_flos": 1.32610101608448e+16,
|
| 1463 |
"train_batch_size": 3,
|
| 1464 |
"trial_name": null,
|
| 1465 |
"trial_params": null
|