Training in progress, step 2000, 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 1573038792
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0eff2c7a72b1bd7a2154447eca5f146b8ca5e7366b2ee2b1e1b1ba795e79c381
|
| 3 |
size 1573038792
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 520574714
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9b578c2ecf931245e13d8454c6cfafeddde4216fa06f8f438fc30d0e9c43bfeb
|
| 3 |
size 520574714
|
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 sha256:28ffaac67b1e5975b97cbb1f1fd2ace18e322aa6d0cb09f44e7757b6245496bc
|
| 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 1000
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:84539daf1c6cd50735af6f3b31ba998dcf05d26804999780a028b860cd6bc73e
|
| 3 |
size 1000
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch": 0.
|
| 5 |
"eval_steps": 1000,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -715,6 +715,714 @@
|
|
| 715 |
"eval_samples_per_second": 6.534,
|
| 716 |
"eval_steps_per_second": 0.847,
|
| 717 |
"step": 1000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 718 |
}
|
| 719 |
],
|
| 720 |
"logging_steps": 10,
|
|
@@ -734,7 +1442,7 @@
|
|
| 734 |
"attributes": {}
|
| 735 |
}
|
| 736 |
},
|
| 737 |
-
"total_flos":
|
| 738 |
"train_batch_size": 4,
|
| 739 |
"trial_name": null,
|
| 740 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.2991660745671441,
|
| 5 |
"eval_steps": 1000,
|
| 6 |
+
"global_step": 2000,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 715 |
"eval_samples_per_second": 6.534,
|
| 716 |
"eval_steps_per_second": 0.847,
|
| 717 |
"step": 1000
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 0.15107886765640777,
|
| 721 |
+
"grad_norm": 0.4198203980922699,
|
| 722 |
+
"learning_rate": 0.0004995341055758453,
|
| 723 |
+
"loss": 7.6051,
|
| 724 |
+
"step": 1010
|
| 725 |
+
},
|
| 726 |
+
{
|
| 727 |
+
"epoch": 0.15257469802924348,
|
| 728 |
+
"grad_norm": 0.614424467086792,
|
| 729 |
+
"learning_rate": 0.0004995156620444583,
|
| 730 |
+
"loss": 7.5955,
|
| 731 |
+
"step": 1020
|
| 732 |
+
},
|
| 733 |
+
{
|
| 734 |
+
"epoch": 0.15407052840207922,
|
| 735 |
+
"grad_norm": 0.5094268918037415,
|
| 736 |
+
"learning_rate": 0.0004994968608535066,
|
| 737 |
+
"loss": 7.5807,
|
| 738 |
+
"step": 1030
|
| 739 |
+
},
|
| 740 |
+
{
|
| 741 |
+
"epoch": 0.15556635877491493,
|
| 742 |
+
"grad_norm": 0.4448883831501007,
|
| 743 |
+
"learning_rate": 0.0004994777020299404,
|
| 744 |
+
"loss": 7.6112,
|
| 745 |
+
"step": 1040
|
| 746 |
+
},
|
| 747 |
+
{
|
| 748 |
+
"epoch": 0.15706218914775064,
|
| 749 |
+
"grad_norm": 0.4672151803970337,
|
| 750 |
+
"learning_rate": 0.0004994581856012219,
|
| 751 |
+
"loss": 7.6151,
|
| 752 |
+
"step": 1050
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 0.15855801952058637,
|
| 756 |
+
"grad_norm": 0.4254607856273651,
|
| 757 |
+
"learning_rate": 0.0004994383115953266,
|
| 758 |
+
"loss": 7.5931,
|
| 759 |
+
"step": 1060
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"epoch": 0.16005384989342208,
|
| 763 |
+
"grad_norm": 0.5613588094711304,
|
| 764 |
+
"learning_rate": 0.000499418080040742,
|
| 765 |
+
"loss": 7.6144,
|
| 766 |
+
"step": 1070
|
| 767 |
+
},
|
| 768 |
+
{
|
| 769 |
+
"epoch": 0.1615496802662578,
|
| 770 |
+
"grad_norm": 0.5766307711601257,
|
| 771 |
+
"learning_rate": 0.0004993974909664683,
|
| 772 |
+
"loss": 7.595,
|
| 773 |
+
"step": 1080
|
| 774 |
+
},
|
| 775 |
+
{
|
| 776 |
+
"epoch": 0.16304551063909353,
|
| 777 |
+
"grad_norm": 0.44581693410873413,
|
| 778 |
+
"learning_rate": 0.0004993765444020183,
|
| 779 |
+
"loss": 7.6035,
|
| 780 |
+
"step": 1090
|
| 781 |
+
},
|
| 782 |
+
{
|
| 783 |
+
"epoch": 0.16454134101192924,
|
| 784 |
+
"grad_norm": 0.48525750637054443,
|
| 785 |
+
"learning_rate": 0.000499355240377417,
|
| 786 |
+
"loss": 7.622,
|
| 787 |
+
"step": 1100
|
| 788 |
+
},
|
| 789 |
+
{
|
| 790 |
+
"epoch": 0.16603717138476498,
|
| 791 |
+
"grad_norm": 0.44798675179481506,
|
| 792 |
+
"learning_rate": 0.000499333578923202,
|
| 793 |
+
"loss": 7.5642,
|
| 794 |
+
"step": 1110
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 0.1675330017576007,
|
| 798 |
+
"grad_norm": 0.45845216512680054,
|
| 799 |
+
"learning_rate": 0.0004993115600704233,
|
| 800 |
+
"loss": 7.6037,
|
| 801 |
+
"step": 1120
|
| 802 |
+
},
|
| 803 |
+
{
|
| 804 |
+
"epoch": 0.1690288321304364,
|
| 805 |
+
"grad_norm": 0.4126313030719757,
|
| 806 |
+
"learning_rate": 0.0004992891838506429,
|
| 807 |
+
"loss": 7.6047,
|
| 808 |
+
"step": 1130
|
| 809 |
+
},
|
| 810 |
+
{
|
| 811 |
+
"epoch": 0.17052466250327214,
|
| 812 |
+
"grad_norm": 0.42024609446525574,
|
| 813 |
+
"learning_rate": 0.0004992664502959351,
|
| 814 |
+
"loss": 7.576,
|
| 815 |
+
"step": 1140
|
| 816 |
+
},
|
| 817 |
+
{
|
| 818 |
+
"epoch": 0.17202049287610785,
|
| 819 |
+
"grad_norm": 0.40088099241256714,
|
| 820 |
+
"learning_rate": 0.0004992433594388868,
|
| 821 |
+
"loss": 7.6071,
|
| 822 |
+
"step": 1150
|
| 823 |
+
},
|
| 824 |
+
{
|
| 825 |
+
"epoch": 0.17351632324894356,
|
| 826 |
+
"grad_norm": 0.48331788182258606,
|
| 827 |
+
"learning_rate": 0.0004992199113125964,
|
| 828 |
+
"loss": 7.6145,
|
| 829 |
+
"step": 1160
|
| 830 |
+
},
|
| 831 |
+
{
|
| 832 |
+
"epoch": 0.1750121536217793,
|
| 833 |
+
"grad_norm": 0.40282952785491943,
|
| 834 |
+
"learning_rate": 0.0004991961059506754,
|
| 835 |
+
"loss": 7.6324,
|
| 836 |
+
"step": 1170
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 0.176507983994615,
|
| 840 |
+
"grad_norm": 0.4179229736328125,
|
| 841 |
+
"learning_rate": 0.0004991719433872461,
|
| 842 |
+
"loss": 7.5516,
|
| 843 |
+
"step": 1180
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 0.17800381436745072,
|
| 847 |
+
"grad_norm": 0.41357994079589844,
|
| 848 |
+
"learning_rate": 0.000499147423656944,
|
| 849 |
+
"loss": 7.5959,
|
| 850 |
+
"step": 1190
|
| 851 |
+
},
|
| 852 |
+
{
|
| 853 |
+
"epoch": 0.17949964474028646,
|
| 854 |
+
"grad_norm": 0.39742010831832886,
|
| 855 |
+
"learning_rate": 0.0004991225467949157,
|
| 856 |
+
"loss": 7.5794,
|
| 857 |
+
"step": 1200
|
| 858 |
+
},
|
| 859 |
+
{
|
| 860 |
+
"epoch": 0.18099547511312217,
|
| 861 |
+
"grad_norm": 0.415096253156662,
|
| 862 |
+
"learning_rate": 0.0004990973128368205,
|
| 863 |
+
"loss": 7.5593,
|
| 864 |
+
"step": 1210
|
| 865 |
+
},
|
| 866 |
+
{
|
| 867 |
+
"epoch": 0.1824913054859579,
|
| 868 |
+
"grad_norm": 0.4735938012599945,
|
| 869 |
+
"learning_rate": 0.0004990717218188286,
|
| 870 |
+
"loss": 7.6023,
|
| 871 |
+
"step": 1220
|
| 872 |
+
},
|
| 873 |
+
{
|
| 874 |
+
"epoch": 0.18398713585879362,
|
| 875 |
+
"grad_norm": 0.4850127696990967,
|
| 876 |
+
"learning_rate": 0.000499045773777623,
|
| 877 |
+
"loss": 7.596,
|
| 878 |
+
"step": 1230
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 0.18548296623162933,
|
| 882 |
+
"grad_norm": 0.3725271224975586,
|
| 883 |
+
"learning_rate": 0.0004990194687503977,
|
| 884 |
+
"loss": 7.646,
|
| 885 |
+
"step": 1240
|
| 886 |
+
},
|
| 887 |
+
{
|
| 888 |
+
"epoch": 0.18697879660446506,
|
| 889 |
+
"grad_norm": 0.42448315024375916,
|
| 890 |
+
"learning_rate": 0.000498992806774859,
|
| 891 |
+
"loss": 7.6134,
|
| 892 |
+
"step": 1250
|
| 893 |
+
},
|
| 894 |
+
{
|
| 895 |
+
"epoch": 0.18847462697730077,
|
| 896 |
+
"grad_norm": 0.4265682101249695,
|
| 897 |
+
"learning_rate": 0.0004989657878892244,
|
| 898 |
+
"loss": 7.5899,
|
| 899 |
+
"step": 1260
|
| 900 |
+
},
|
| 901 |
+
{
|
| 902 |
+
"epoch": 0.18997045735013648,
|
| 903 |
+
"grad_norm": 0.4955170154571533,
|
| 904 |
+
"learning_rate": 0.0004989384121322232,
|
| 905 |
+
"loss": 7.6075,
|
| 906 |
+
"step": 1270
|
| 907 |
+
},
|
| 908 |
+
{
|
| 909 |
+
"epoch": 0.19146628772297222,
|
| 910 |
+
"grad_norm": 0.37979021668434143,
|
| 911 |
+
"learning_rate": 0.0004989106795430965,
|
| 912 |
+
"loss": 7.6068,
|
| 913 |
+
"step": 1280
|
| 914 |
+
},
|
| 915 |
+
{
|
| 916 |
+
"epoch": 0.19296211809580793,
|
| 917 |
+
"grad_norm": 0.414298415184021,
|
| 918 |
+
"learning_rate": 0.0004988825901615962,
|
| 919 |
+
"loss": 7.6005,
|
| 920 |
+
"step": 1290
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 0.19445794846864364,
|
| 924 |
+
"grad_norm": 0.44306114315986633,
|
| 925 |
+
"learning_rate": 0.0004988541440279862,
|
| 926 |
+
"loss": 7.5973,
|
| 927 |
+
"step": 1300
|
| 928 |
+
},
|
| 929 |
+
{
|
| 930 |
+
"epoch": 0.19595377884147938,
|
| 931 |
+
"grad_norm": 0.3580173850059509,
|
| 932 |
+
"learning_rate": 0.0004988253411830418,
|
| 933 |
+
"loss": 7.595,
|
| 934 |
+
"step": 1310
|
| 935 |
+
},
|
| 936 |
+
{
|
| 937 |
+
"epoch": 0.1974496092143151,
|
| 938 |
+
"grad_norm": 0.4121604859828949,
|
| 939 |
+
"learning_rate": 0.0004987961816680492,
|
| 940 |
+
"loss": 7.5682,
|
| 941 |
+
"step": 1320
|
| 942 |
+
},
|
| 943 |
+
{
|
| 944 |
+
"epoch": 0.19894543958715083,
|
| 945 |
+
"grad_norm": 0.38618120551109314,
|
| 946 |
+
"learning_rate": 0.0004987666655248063,
|
| 947 |
+
"loss": 7.586,
|
| 948 |
+
"step": 1330
|
| 949 |
+
},
|
| 950 |
+
{
|
| 951 |
+
"epoch": 0.20044126995998654,
|
| 952 |
+
"grad_norm": 0.38373953104019165,
|
| 953 |
+
"learning_rate": 0.0004987367927956217,
|
| 954 |
+
"loss": 7.5604,
|
| 955 |
+
"step": 1340
|
| 956 |
+
},
|
| 957 |
+
{
|
| 958 |
+
"epoch": 0.20193710033282225,
|
| 959 |
+
"grad_norm": 0.420476496219635,
|
| 960 |
+
"learning_rate": 0.0004987065635233158,
|
| 961 |
+
"loss": 7.6216,
|
| 962 |
+
"step": 1350
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 0.203432930705658,
|
| 966 |
+
"grad_norm": 0.34908464550971985,
|
| 967 |
+
"learning_rate": 0.0004986759777512196,
|
| 968 |
+
"loss": 7.6102,
|
| 969 |
+
"step": 1360
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 0.2049287610784937,
|
| 973 |
+
"grad_norm": 0.5113145112991333,
|
| 974 |
+
"learning_rate": 0.0004986450355231748,
|
| 975 |
+
"loss": 7.6339,
|
| 976 |
+
"step": 1370
|
| 977 |
+
},
|
| 978 |
+
{
|
| 979 |
+
"epoch": 0.2064245914513294,
|
| 980 |
+
"grad_norm": 0.3865181803703308,
|
| 981 |
+
"learning_rate": 0.0004986137368835351,
|
| 982 |
+
"loss": 7.6221,
|
| 983 |
+
"step": 1380
|
| 984 |
+
},
|
| 985 |
+
{
|
| 986 |
+
"epoch": 0.20792042182416515,
|
| 987 |
+
"grad_norm": 0.3864174783229828,
|
| 988 |
+
"learning_rate": 0.0004985820818771639,
|
| 989 |
+
"loss": 7.5998,
|
| 990 |
+
"step": 1390
|
| 991 |
+
},
|
| 992 |
+
{
|
| 993 |
+
"epoch": 0.20941625219700086,
|
| 994 |
+
"grad_norm": 0.5763099789619446,
|
| 995 |
+
"learning_rate": 0.0004985500705494364,
|
| 996 |
+
"loss": 7.5634,
|
| 997 |
+
"step": 1400
|
| 998 |
+
},
|
| 999 |
+
{
|
| 1000 |
+
"epoch": 0.21091208256983657,
|
| 1001 |
+
"grad_norm": 0.3169045150279999,
|
| 1002 |
+
"learning_rate": 0.0004985177029462379,
|
| 1003 |
+
"loss": 7.5592,
|
| 1004 |
+
"step": 1410
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 0.2124079129426723,
|
| 1008 |
+
"grad_norm": 0.4025494456291199,
|
| 1009 |
+
"learning_rate": 0.0004984849791139646,
|
| 1010 |
+
"loss": 7.5902,
|
| 1011 |
+
"step": 1420
|
| 1012 |
+
},
|
| 1013 |
+
{
|
| 1014 |
+
"epoch": 0.21390374331550802,
|
| 1015 |
+
"grad_norm": 0.3893519341945648,
|
| 1016 |
+
"learning_rate": 0.0004984518990995234,
|
| 1017 |
+
"loss": 7.5984,
|
| 1018 |
+
"step": 1430
|
| 1019 |
+
},
|
| 1020 |
+
{
|
| 1021 |
+
"epoch": 0.21539957368834375,
|
| 1022 |
+
"grad_norm": 0.3353728950023651,
|
| 1023 |
+
"learning_rate": 0.0004984184629503318,
|
| 1024 |
+
"loss": 7.5662,
|
| 1025 |
+
"step": 1440
|
| 1026 |
+
},
|
| 1027 |
+
{
|
| 1028 |
+
"epoch": 0.21689540406117946,
|
| 1029 |
+
"grad_norm": 0.36072856187820435,
|
| 1030 |
+
"learning_rate": 0.0004983846707143174,
|
| 1031 |
+
"loss": 7.5784,
|
| 1032 |
+
"step": 1450
|
| 1033 |
+
},
|
| 1034 |
+
{
|
| 1035 |
+
"epoch": 0.21839123443401517,
|
| 1036 |
+
"grad_norm": 0.385777086019516,
|
| 1037 |
+
"learning_rate": 0.0004983505224399187,
|
| 1038 |
+
"loss": 7.593,
|
| 1039 |
+
"step": 1460
|
| 1040 |
+
},
|
| 1041 |
+
{
|
| 1042 |
+
"epoch": 0.2198870648068509,
|
| 1043 |
+
"grad_norm": 0.3383258581161499,
|
| 1044 |
+
"learning_rate": 0.0004983160181760845,
|
| 1045 |
+
"loss": 7.5652,
|
| 1046 |
+
"step": 1470
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 0.22138289517968662,
|
| 1050 |
+
"grad_norm": 0.409434050321579,
|
| 1051 |
+
"learning_rate": 0.0004982811579722735,
|
| 1052 |
+
"loss": 7.5578,
|
| 1053 |
+
"step": 1480
|
| 1054 |
+
},
|
| 1055 |
+
{
|
| 1056 |
+
"epoch": 0.22287872555252233,
|
| 1057 |
+
"grad_norm": 0.40944933891296387,
|
| 1058 |
+
"learning_rate": 0.0004982459418784549,
|
| 1059 |
+
"loss": 7.5988,
|
| 1060 |
+
"step": 1490
|
| 1061 |
+
},
|
| 1062 |
+
{
|
| 1063 |
+
"epoch": 0.22437455592535807,
|
| 1064 |
+
"grad_norm": 0.3934713900089264,
|
| 1065 |
+
"learning_rate": 0.0004982103699451082,
|
| 1066 |
+
"loss": 7.5879,
|
| 1067 |
+
"step": 1500
|
| 1068 |
+
},
|
| 1069 |
+
{
|
| 1070 |
+
"epoch": 0.22587038629819378,
|
| 1071 |
+
"grad_norm": 0.3455521762371063,
|
| 1072 |
+
"learning_rate": 0.0004981744422232224,
|
| 1073 |
+
"loss": 7.6158,
|
| 1074 |
+
"step": 1510
|
| 1075 |
+
},
|
| 1076 |
+
{
|
| 1077 |
+
"epoch": 0.2273662166710295,
|
| 1078 |
+
"grad_norm": 0.35238131880760193,
|
| 1079 |
+
"learning_rate": 0.000498138158764297,
|
| 1080 |
+
"loss": 7.6229,
|
| 1081 |
+
"step": 1520
|
| 1082 |
+
},
|
| 1083 |
+
{
|
| 1084 |
+
"epoch": 0.22886204704386523,
|
| 1085 |
+
"grad_norm": 0.3862851858139038,
|
| 1086 |
+
"learning_rate": 0.0004981015196203414,
|
| 1087 |
+
"loss": 7.5969,
|
| 1088 |
+
"step": 1530
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 0.23035787741670094,
|
| 1092 |
+
"grad_norm": 0.39322030544281006,
|
| 1093 |
+
"learning_rate": 0.0004980645248438745,
|
| 1094 |
+
"loss": 7.5595,
|
| 1095 |
+
"step": 1540
|
| 1096 |
+
},
|
| 1097 |
+
{
|
| 1098 |
+
"epoch": 0.23185370778953668,
|
| 1099 |
+
"grad_norm": 0.4199928045272827,
|
| 1100 |
+
"learning_rate": 0.0004980271744879254,
|
| 1101 |
+
"loss": 7.5719,
|
| 1102 |
+
"step": 1550
|
| 1103 |
+
},
|
| 1104 |
+
{
|
| 1105 |
+
"epoch": 0.2333495381623724,
|
| 1106 |
+
"grad_norm": 0.6090747714042664,
|
| 1107 |
+
"learning_rate": 0.0004979894686060325,
|
| 1108 |
+
"loss": 7.544,
|
| 1109 |
+
"step": 1560
|
| 1110 |
+
},
|
| 1111 |
+
{
|
| 1112 |
+
"epoch": 0.2348453685352081,
|
| 1113 |
+
"grad_norm": 0.3939005136489868,
|
| 1114 |
+
"learning_rate": 0.0004979514072522439,
|
| 1115 |
+
"loss": 7.5816,
|
| 1116 |
+
"step": 1570
|
| 1117 |
+
},
|
| 1118 |
+
{
|
| 1119 |
+
"epoch": 0.23634119890804384,
|
| 1120 |
+
"grad_norm": 0.3856000006198883,
|
| 1121 |
+
"learning_rate": 0.0004979129904811176,
|
| 1122 |
+
"loss": 7.5489,
|
| 1123 |
+
"step": 1580
|
| 1124 |
+
},
|
| 1125 |
+
{
|
| 1126 |
+
"epoch": 0.23783702928087955,
|
| 1127 |
+
"grad_norm": 0.42172351479530334,
|
| 1128 |
+
"learning_rate": 0.0004978742183477206,
|
| 1129 |
+
"loss": 7.5819,
|
| 1130 |
+
"step": 1590
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 0.23933285965371526,
|
| 1134 |
+
"grad_norm": 0.3501926064491272,
|
| 1135 |
+
"learning_rate": 0.0004978350909076295,
|
| 1136 |
+
"loss": 7.5846,
|
| 1137 |
+
"step": 1600
|
| 1138 |
+
},
|
| 1139 |
+
{
|
| 1140 |
+
"epoch": 0.240828690026551,
|
| 1141 |
+
"grad_norm": 0.39298102259635925,
|
| 1142 |
+
"learning_rate": 0.0004977956082169303,
|
| 1143 |
+
"loss": 7.5995,
|
| 1144 |
+
"step": 1610
|
| 1145 |
+
},
|
| 1146 |
+
{
|
| 1147 |
+
"epoch": 0.2423245203993867,
|
| 1148 |
+
"grad_norm": 0.430789977312088,
|
| 1149 |
+
"learning_rate": 0.0004977557703322178,
|
| 1150 |
+
"loss": 7.6124,
|
| 1151 |
+
"step": 1620
|
| 1152 |
+
},
|
| 1153 |
+
{
|
| 1154 |
+
"epoch": 0.24382035077222242,
|
| 1155 |
+
"grad_norm": 0.4584248661994934,
|
| 1156 |
+
"learning_rate": 0.0004977155773105965,
|
| 1157 |
+
"loss": 7.5875,
|
| 1158 |
+
"step": 1630
|
| 1159 |
+
},
|
| 1160 |
+
{
|
| 1161 |
+
"epoch": 0.24531618114505815,
|
| 1162 |
+
"grad_norm": 0.39846205711364746,
|
| 1163 |
+
"learning_rate": 0.0004976750292096796,
|
| 1164 |
+
"loss": 7.6027,
|
| 1165 |
+
"step": 1640
|
| 1166 |
+
},
|
| 1167 |
+
{
|
| 1168 |
+
"epoch": 0.24681201151789386,
|
| 1169 |
+
"grad_norm": 0.37316691875457764,
|
| 1170 |
+
"learning_rate": 0.0004976341260875894,
|
| 1171 |
+
"loss": 7.6182,
|
| 1172 |
+
"step": 1650
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 0.2483078418907296,
|
| 1176 |
+
"grad_norm": 0.3936406373977661,
|
| 1177 |
+
"learning_rate": 0.000497592868002957,
|
| 1178 |
+
"loss": 7.5557,
|
| 1179 |
+
"step": 1660
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 0.2498036722635653,
|
| 1183 |
+
"grad_norm": 0.337677538394928,
|
| 1184 |
+
"learning_rate": 0.0004975512550149224,
|
| 1185 |
+
"loss": 7.5995,
|
| 1186 |
+
"step": 1670
|
| 1187 |
+
},
|
| 1188 |
+
{
|
| 1189 |
+
"epoch": 0.25129950263640105,
|
| 1190 |
+
"grad_norm": 0.31229522824287415,
|
| 1191 |
+
"learning_rate": 0.0004975092871831343,
|
| 1192 |
+
"loss": 7.5816,
|
| 1193 |
+
"step": 1680
|
| 1194 |
+
},
|
| 1195 |
+
{
|
| 1196 |
+
"epoch": 0.25279533300923673,
|
| 1197 |
+
"grad_norm": 0.31100282073020935,
|
| 1198 |
+
"learning_rate": 0.00049746696456775,
|
| 1199 |
+
"loss": 7.5957,
|
| 1200 |
+
"step": 1690
|
| 1201 |
+
},
|
| 1202 |
+
{
|
| 1203 |
+
"epoch": 0.25429116338207247,
|
| 1204 |
+
"grad_norm": 0.3856932520866394,
|
| 1205 |
+
"learning_rate": 0.0004974242872294354,
|
| 1206 |
+
"loss": 7.6212,
|
| 1207 |
+
"step": 1700
|
| 1208 |
+
},
|
| 1209 |
+
{
|
| 1210 |
+
"epoch": 0.2557869937549082,
|
| 1211 |
+
"grad_norm": 0.34433454275131226,
|
| 1212 |
+
"learning_rate": 0.000497381255229365,
|
| 1213 |
+
"loss": 7.5182,
|
| 1214 |
+
"step": 1710
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 0.2572828241277439,
|
| 1218 |
+
"grad_norm": 0.3633713722229004,
|
| 1219 |
+
"learning_rate": 0.0004973378686292211,
|
| 1220 |
+
"loss": 7.619,
|
| 1221 |
+
"step": 1720
|
| 1222 |
+
},
|
| 1223 |
+
{
|
| 1224 |
+
"epoch": 0.25877865450057963,
|
| 1225 |
+
"grad_norm": 0.4015657901763916,
|
| 1226 |
+
"learning_rate": 0.0004972941274911952,
|
| 1227 |
+
"loss": 7.6097,
|
| 1228 |
+
"step": 1730
|
| 1229 |
+
},
|
| 1230 |
+
{
|
| 1231 |
+
"epoch": 0.26027448487341537,
|
| 1232 |
+
"grad_norm": 0.36401477456092834,
|
| 1233 |
+
"learning_rate": 0.0004972500318779863,
|
| 1234 |
+
"loss": 7.586,
|
| 1235 |
+
"step": 1740
|
| 1236 |
+
},
|
| 1237 |
+
{
|
| 1238 |
+
"epoch": 0.26177031524625105,
|
| 1239 |
+
"grad_norm": 0.42530539631843567,
|
| 1240 |
+
"learning_rate": 0.0004972055818528017,
|
| 1241 |
+
"loss": 7.5906,
|
| 1242 |
+
"step": 1750
|
| 1243 |
+
},
|
| 1244 |
+
{
|
| 1245 |
+
"epoch": 0.2632661456190868,
|
| 1246 |
+
"grad_norm": 0.5568501353263855,
|
| 1247 |
+
"learning_rate": 0.0004971607774793569,
|
| 1248 |
+
"loss": 7.5886,
|
| 1249 |
+
"step": 1760
|
| 1250 |
+
},
|
| 1251 |
+
{
|
| 1252 |
+
"epoch": 0.2647619759919225,
|
| 1253 |
+
"grad_norm": 0.33592841029167175,
|
| 1254 |
+
"learning_rate": 0.0004971156188218749,
|
| 1255 |
+
"loss": 7.6141,
|
| 1256 |
+
"step": 1770
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 0.2662578063647582,
|
| 1260 |
+
"grad_norm": 0.48852601647377014,
|
| 1261 |
+
"learning_rate": 0.0004970701059450872,
|
| 1262 |
+
"loss": 7.5809,
|
| 1263 |
+
"step": 1780
|
| 1264 |
+
},
|
| 1265 |
+
{
|
| 1266 |
+
"epoch": 0.26775363673759395,
|
| 1267 |
+
"grad_norm": 0.33326566219329834,
|
| 1268 |
+
"learning_rate": 0.0004970242389142322,
|
| 1269 |
+
"loss": 7.6085,
|
| 1270 |
+
"step": 1790
|
| 1271 |
+
},
|
| 1272 |
+
{
|
| 1273 |
+
"epoch": 0.2692494671104297,
|
| 1274 |
+
"grad_norm": 0.301401287317276,
|
| 1275 |
+
"learning_rate": 0.0004969780177950568,
|
| 1276 |
+
"loss": 7.583,
|
| 1277 |
+
"step": 1800
|
| 1278 |
+
},
|
| 1279 |
+
{
|
| 1280 |
+
"epoch": 0.2707452974832654,
|
| 1281 |
+
"grad_norm": 0.3466792702674866,
|
| 1282 |
+
"learning_rate": 0.0004969314426538147,
|
| 1283 |
+
"loss": 7.5854,
|
| 1284 |
+
"step": 1810
|
| 1285 |
+
},
|
| 1286 |
+
{
|
| 1287 |
+
"epoch": 0.2722411278561011,
|
| 1288 |
+
"grad_norm": 0.3221440017223358,
|
| 1289 |
+
"learning_rate": 0.0004968845135572677,
|
| 1290 |
+
"loss": 7.5701,
|
| 1291 |
+
"step": 1820
|
| 1292 |
+
},
|
| 1293 |
+
{
|
| 1294 |
+
"epoch": 0.27373695822893684,
|
| 1295 |
+
"grad_norm": 0.35062047839164734,
|
| 1296 |
+
"learning_rate": 0.0004968372305726846,
|
| 1297 |
+
"loss": 7.5589,
|
| 1298 |
+
"step": 1830
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 0.2752327886017726,
|
| 1302 |
+
"grad_norm": 0.3333365321159363,
|
| 1303 |
+
"learning_rate": 0.0004967895937678416,
|
| 1304 |
+
"loss": 7.5638,
|
| 1305 |
+
"step": 1840
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 0.27672861897460826,
|
| 1309 |
+
"grad_norm": 0.3974228501319885,
|
| 1310 |
+
"learning_rate": 0.0004967416032110219,
|
| 1311 |
+
"loss": 7.5589,
|
| 1312 |
+
"step": 1850
|
| 1313 |
+
},
|
| 1314 |
+
{
|
| 1315 |
+
"epoch": 0.278224449347444,
|
| 1316 |
+
"grad_norm": 0.3333839178085327,
|
| 1317 |
+
"learning_rate": 0.0004966932589710161,
|
| 1318 |
+
"loss": 7.5926,
|
| 1319 |
+
"step": 1860
|
| 1320 |
+
},
|
| 1321 |
+
{
|
| 1322 |
+
"epoch": 0.27972027972027974,
|
| 1323 |
+
"grad_norm": 0.3255492150783539,
|
| 1324 |
+
"learning_rate": 0.0004966445611171212,
|
| 1325 |
+
"loss": 7.595,
|
| 1326 |
+
"step": 1870
|
| 1327 |
+
},
|
| 1328 |
+
{
|
| 1329 |
+
"epoch": 0.2812161100931154,
|
| 1330 |
+
"grad_norm": 0.3025016486644745,
|
| 1331 |
+
"learning_rate": 0.0004965955097191419,
|
| 1332 |
+
"loss": 7.6292,
|
| 1333 |
+
"step": 1880
|
| 1334 |
+
},
|
| 1335 |
+
{
|
| 1336 |
+
"epoch": 0.28271194046595116,
|
| 1337 |
+
"grad_norm": 0.3485715985298157,
|
| 1338 |
+
"learning_rate": 0.0004965461048473889,
|
| 1339 |
+
"loss": 7.575,
|
| 1340 |
+
"step": 1890
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 0.2842077708387869,
|
| 1344 |
+
"grad_norm": 0.34214696288108826,
|
| 1345 |
+
"learning_rate": 0.00049649634657268,
|
| 1346 |
+
"loss": 7.5899,
|
| 1347 |
+
"step": 1900
|
| 1348 |
+
},
|
| 1349 |
+
{
|
| 1350 |
+
"epoch": 0.2857036012116226,
|
| 1351 |
+
"grad_norm": 0.2842009663581848,
|
| 1352 |
+
"learning_rate": 0.0004964462349663395,
|
| 1353 |
+
"loss": 7.6015,
|
| 1354 |
+
"step": 1910
|
| 1355 |
+
},
|
| 1356 |
+
{
|
| 1357 |
+
"epoch": 0.2871994315844583,
|
| 1358 |
+
"grad_norm": 0.41511401534080505,
|
| 1359 |
+
"learning_rate": 0.0004963957701001982,
|
| 1360 |
+
"loss": 7.5954,
|
| 1361 |
+
"step": 1920
|
| 1362 |
+
},
|
| 1363 |
+
{
|
| 1364 |
+
"epoch": 0.28869526195729406,
|
| 1365 |
+
"grad_norm": 0.3453192412853241,
|
| 1366 |
+
"learning_rate": 0.000496344952046593,
|
| 1367 |
+
"loss": 7.5533,
|
| 1368 |
+
"step": 1930
|
| 1369 |
+
},
|
| 1370 |
+
{
|
| 1371 |
+
"epoch": 0.29019109233012974,
|
| 1372 |
+
"grad_norm": 0.3092595338821411,
|
| 1373 |
+
"learning_rate": 0.0004962937808783675,
|
| 1374 |
+
"loss": 7.5903,
|
| 1375 |
+
"step": 1940
|
| 1376 |
+
},
|
| 1377 |
+
{
|
| 1378 |
+
"epoch": 0.2916869227029655,
|
| 1379 |
+
"grad_norm": 0.3328978419303894,
|
| 1380 |
+
"learning_rate": 0.0004962422566688711,
|
| 1381 |
+
"loss": 7.6138,
|
| 1382 |
+
"step": 1950
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 0.2931827530758012,
|
| 1386 |
+
"grad_norm": 0.34327930212020874,
|
| 1387 |
+
"learning_rate": 0.0004961903794919595,
|
| 1388 |
+
"loss": 7.578,
|
| 1389 |
+
"step": 1960
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 0.2946785834486369,
|
| 1393 |
+
"grad_norm": 0.3706710934638977,
|
| 1394 |
+
"learning_rate": 0.0004961381494219941,
|
| 1395 |
+
"loss": 7.5915,
|
| 1396 |
+
"step": 1970
|
| 1397 |
+
},
|
| 1398 |
+
{
|
| 1399 |
+
"epoch": 0.29617441382147264,
|
| 1400 |
+
"grad_norm": 0.33585718274116516,
|
| 1401 |
+
"learning_rate": 0.0004960855665338424,
|
| 1402 |
+
"loss": 7.6123,
|
| 1403 |
+
"step": 1980
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"epoch": 0.2976702441943084,
|
| 1407 |
+
"grad_norm": 0.34515875577926636,
|
| 1408 |
+
"learning_rate": 0.0004960326309028775,
|
| 1409 |
+
"loss": 7.5872,
|
| 1410 |
+
"step": 1990
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"epoch": 0.2991660745671441,
|
| 1414 |
+
"grad_norm": 0.36676159501075745,
|
| 1415 |
+
"learning_rate": 0.000495979342604978,
|
| 1416 |
+
"loss": 7.5539,
|
| 1417 |
+
"step": 2000
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 0.2991660745671441,
|
| 1421 |
+
"eval_loss": 7.588481426239014,
|
| 1422 |
+
"eval_runtime": 16.541,
|
| 1423 |
+
"eval_samples_per_second": 6.529,
|
| 1424 |
+
"eval_steps_per_second": 0.846,
|
| 1425 |
+
"step": 2000
|
| 1426 |
}
|
| 1427 |
],
|
| 1428 |
"logging_steps": 10,
|
|
|
|
| 1442 |
"attributes": {}
|
| 1443 |
}
|
| 1444 |
},
|
| 1445 |
+
"total_flos": 1.4943906767782748e+18,
|
| 1446 |
"train_batch_size": 4,
|
| 1447 |
"trial_name": null,
|
| 1448 |
"trial_params": null
|