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 289452128
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:36670a5cb07570b526cc68b5b9d0ebaa892ac926af726fbdbe84eaef96121c65
|
| 3 |
size 289452128
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 147359892
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fe5c90e3905bbef6d07cf5a823555596fd9e072abb4be8bdc0ec4434b20e757e
|
| 3 |
size 147359892
|
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 sha256:630e2711af6547215ca84e215db44507617e11099bbc440a92e4d380f49e9a73
|
| 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 sha256:530505d607699f384741067a5f9139d72f043713adb680898a3f1b5714170c97
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 0.
|
| 3 |
-
"best_model_checkpoint": "miner_id_24/checkpoint-
|
| 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": 31.902,
|
| 724 |
"eval_steps_per_second": 7.977,
|
| 725 |
"step": 100
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 726 |
}
|
| 727 |
],
|
| 728 |
"logging_steps": 1,
|
|
@@ -751,7 +1459,7 @@
|
|
| 751 |
"attributes": {}
|
| 752 |
}
|
| 753 |
},
|
| 754 |
-
"total_flos":
|
| 755 |
"train_batch_size": 8,
|
| 756 |
"trial_name": null,
|
| 757 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 0.7008334398269653,
|
| 3 |
+
"best_model_checkpoint": "miner_id_24/checkpoint-200",
|
| 4 |
+
"epoch": 0.039846590626089556,
|
| 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": 31.902,
|
| 724 |
"eval_steps_per_second": 7.977,
|
| 725 |
"step": 100
|
| 726 |
+
},
|
| 727 |
+
{
|
| 728 |
+
"epoch": 0.020122528266175225,
|
| 729 |
+
"grad_norm": 0.21482613682746887,
|
| 730 |
+
"learning_rate": 9.31367192988896e-05,
|
| 731 |
+
"loss": 0.5356,
|
| 732 |
+
"step": 101
|
| 733 |
+
},
|
| 734 |
+
{
|
| 735 |
+
"epoch": 0.02032176121930567,
|
| 736 |
+
"grad_norm": 0.28820475935935974,
|
| 737 |
+
"learning_rate": 9.297032057507264e-05,
|
| 738 |
+
"loss": 0.6606,
|
| 739 |
+
"step": 102
|
| 740 |
+
},
|
| 741 |
+
{
|
| 742 |
+
"epoch": 0.02052099417243612,
|
| 743 |
+
"grad_norm": 0.22919978201389313,
|
| 744 |
+
"learning_rate": 9.280208114573859e-05,
|
| 745 |
+
"loss": 0.6681,
|
| 746 |
+
"step": 103
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 0.02072022712556657,
|
| 750 |
+
"grad_norm": 0.20591995120048523,
|
| 751 |
+
"learning_rate": 9.263200821770461e-05,
|
| 752 |
+
"loss": 0.6523,
|
| 753 |
+
"step": 104
|
| 754 |
+
},
|
| 755 |
+
{
|
| 756 |
+
"epoch": 0.020919460078697015,
|
| 757 |
+
"grad_norm": 0.20084571838378906,
|
| 758 |
+
"learning_rate": 9.246010907632895e-05,
|
| 759 |
+
"loss": 0.629,
|
| 760 |
+
"step": 105
|
| 761 |
+
},
|
| 762 |
+
{
|
| 763 |
+
"epoch": 0.021118693031827465,
|
| 764 |
+
"grad_norm": 0.25101473927497864,
|
| 765 |
+
"learning_rate": 9.228639108519868e-05,
|
| 766 |
+
"loss": 0.7384,
|
| 767 |
+
"step": 106
|
| 768 |
+
},
|
| 769 |
+
{
|
| 770 |
+
"epoch": 0.021317925984957912,
|
| 771 |
+
"grad_norm": 0.23734250664710999,
|
| 772 |
+
"learning_rate": 9.211086168581433e-05,
|
| 773 |
+
"loss": 0.6895,
|
| 774 |
+
"step": 107
|
| 775 |
+
},
|
| 776 |
+
{
|
| 777 |
+
"epoch": 0.02151715893808836,
|
| 778 |
+
"grad_norm": 0.22536040842533112,
|
| 779 |
+
"learning_rate": 9.193352839727121e-05,
|
| 780 |
+
"loss": 0.7137,
|
| 781 |
+
"step": 108
|
| 782 |
+
},
|
| 783 |
+
{
|
| 784 |
+
"epoch": 0.02171639189121881,
|
| 785 |
+
"grad_norm": 0.2355283945798874,
|
| 786 |
+
"learning_rate": 9.175439881593716e-05,
|
| 787 |
+
"loss": 0.7662,
|
| 788 |
+
"step": 109
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 0.021915624844349255,
|
| 792 |
+
"grad_norm": 0.2228918820619583,
|
| 793 |
+
"learning_rate": 9.157348061512727e-05,
|
| 794 |
+
"loss": 0.6534,
|
| 795 |
+
"step": 110
|
| 796 |
+
},
|
| 797 |
+
{
|
| 798 |
+
"epoch": 0.022114857797479702,
|
| 799 |
+
"grad_norm": 0.2184191644191742,
|
| 800 |
+
"learning_rate": 9.139078154477512e-05,
|
| 801 |
+
"loss": 0.7054,
|
| 802 |
+
"step": 111
|
| 803 |
+
},
|
| 804 |
+
{
|
| 805 |
+
"epoch": 0.022314090750610152,
|
| 806 |
+
"grad_norm": 0.21408338844776154,
|
| 807 |
+
"learning_rate": 9.120630943110077e-05,
|
| 808 |
+
"loss": 0.6103,
|
| 809 |
+
"step": 112
|
| 810 |
+
},
|
| 811 |
+
{
|
| 812 |
+
"epoch": 0.0225133237037406,
|
| 813 |
+
"grad_norm": 0.25427061319351196,
|
| 814 |
+
"learning_rate": 9.102007217627568e-05,
|
| 815 |
+
"loss": 0.8366,
|
| 816 |
+
"step": 113
|
| 817 |
+
},
|
| 818 |
+
{
|
| 819 |
+
"epoch": 0.022712556656871045,
|
| 820 |
+
"grad_norm": 0.21583066880702972,
|
| 821 |
+
"learning_rate": 9.083207775808396e-05,
|
| 822 |
+
"loss": 0.7165,
|
| 823 |
+
"step": 114
|
| 824 |
+
},
|
| 825 |
+
{
|
| 826 |
+
"epoch": 0.022911789610001496,
|
| 827 |
+
"grad_norm": 0.2396174967288971,
|
| 828 |
+
"learning_rate": 9.064233422958077e-05,
|
| 829 |
+
"loss": 0.6794,
|
| 830 |
+
"step": 115
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 0.023111022563131942,
|
| 834 |
+
"grad_norm": 0.2334282398223877,
|
| 835 |
+
"learning_rate": 9.045084971874738e-05,
|
| 836 |
+
"loss": 0.7945,
|
| 837 |
+
"step": 116
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 0.02331025551626239,
|
| 841 |
+
"grad_norm": 0.20607277750968933,
|
| 842 |
+
"learning_rate": 9.025763242814291e-05,
|
| 843 |
+
"loss": 0.5739,
|
| 844 |
+
"step": 117
|
| 845 |
+
},
|
| 846 |
+
{
|
| 847 |
+
"epoch": 0.02350948846939284,
|
| 848 |
+
"grad_norm": 0.2125677615404129,
|
| 849 |
+
"learning_rate": 9.006269063455304e-05,
|
| 850 |
+
"loss": 0.6976,
|
| 851 |
+
"step": 118
|
| 852 |
+
},
|
| 853 |
+
{
|
| 854 |
+
"epoch": 0.023708721422523286,
|
| 855 |
+
"grad_norm": 0.24233980476856232,
|
| 856 |
+
"learning_rate": 8.986603268863536e-05,
|
| 857 |
+
"loss": 0.7768,
|
| 858 |
+
"step": 119
|
| 859 |
+
},
|
| 860 |
+
{
|
| 861 |
+
"epoch": 0.023907954375653732,
|
| 862 |
+
"grad_norm": 0.23903363943099976,
|
| 863 |
+
"learning_rate": 8.966766701456177e-05,
|
| 864 |
+
"loss": 0.7925,
|
| 865 |
+
"step": 120
|
| 866 |
+
},
|
| 867 |
+
{
|
| 868 |
+
"epoch": 0.024107187328784183,
|
| 869 |
+
"grad_norm": 0.2270466685295105,
|
| 870 |
+
"learning_rate": 8.94676021096575e-05,
|
| 871 |
+
"loss": 0.679,
|
| 872 |
+
"step": 121
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 0.02430642028191463,
|
| 876 |
+
"grad_norm": 0.22134599089622498,
|
| 877 |
+
"learning_rate": 8.926584654403724e-05,
|
| 878 |
+
"loss": 0.7155,
|
| 879 |
+
"step": 122
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"epoch": 0.024505653235045076,
|
| 883 |
+
"grad_norm": 0.2539938986301422,
|
| 884 |
+
"learning_rate": 8.906240896023794e-05,
|
| 885 |
+
"loss": 0.7156,
|
| 886 |
+
"step": 123
|
| 887 |
+
},
|
| 888 |
+
{
|
| 889 |
+
"epoch": 0.024704886188175523,
|
| 890 |
+
"grad_norm": 0.24332107603549957,
|
| 891 |
+
"learning_rate": 8.885729807284856e-05,
|
| 892 |
+
"loss": 0.6988,
|
| 893 |
+
"step": 124
|
| 894 |
+
},
|
| 895 |
+
{
|
| 896 |
+
"epoch": 0.024904119141305973,
|
| 897 |
+
"grad_norm": 0.23084275424480438,
|
| 898 |
+
"learning_rate": 8.865052266813685e-05,
|
| 899 |
+
"loss": 0.7142,
|
| 900 |
+
"step": 125
|
| 901 |
+
},
|
| 902 |
+
{
|
| 903 |
+
"epoch": 0.02510335209443642,
|
| 904 |
+
"grad_norm": 0.24665014445781708,
|
| 905 |
+
"learning_rate": 8.844209160367299e-05,
|
| 906 |
+
"loss": 0.7471,
|
| 907 |
+
"step": 126
|
| 908 |
+
},
|
| 909 |
+
{
|
| 910 |
+
"epoch": 0.025302585047566866,
|
| 911 |
+
"grad_norm": 0.23389537632465363,
|
| 912 |
+
"learning_rate": 8.823201380795001e-05,
|
| 913 |
+
"loss": 0.676,
|
| 914 |
+
"step": 127
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 0.025501818000697316,
|
| 918 |
+
"grad_norm": 0.2225302904844284,
|
| 919 |
+
"learning_rate": 8.802029828000156e-05,
|
| 920 |
+
"loss": 0.6702,
|
| 921 |
+
"step": 128
|
| 922 |
+
},
|
| 923 |
+
{
|
| 924 |
+
"epoch": 0.025701050953827763,
|
| 925 |
+
"grad_norm": 0.2194124162197113,
|
| 926 |
+
"learning_rate": 8.780695408901613e-05,
|
| 927 |
+
"loss": 0.7173,
|
| 928 |
+
"step": 129
|
| 929 |
+
},
|
| 930 |
+
{
|
| 931 |
+
"epoch": 0.02590028390695821,
|
| 932 |
+
"grad_norm": 0.22123487293720245,
|
| 933 |
+
"learning_rate": 8.759199037394887e-05,
|
| 934 |
+
"loss": 0.679,
|
| 935 |
+
"step": 130
|
| 936 |
+
},
|
| 937 |
+
{
|
| 938 |
+
"epoch": 0.02609951686008866,
|
| 939 |
+
"grad_norm": 0.2465553730726242,
|
| 940 |
+
"learning_rate": 8.737541634312985e-05,
|
| 941 |
+
"loss": 0.7151,
|
| 942 |
+
"step": 131
|
| 943 |
+
},
|
| 944 |
+
{
|
| 945 |
+
"epoch": 0.026298749813219106,
|
| 946 |
+
"grad_norm": 0.2150459736585617,
|
| 947 |
+
"learning_rate": 8.715724127386972e-05,
|
| 948 |
+
"loss": 0.7106,
|
| 949 |
+
"step": 132
|
| 950 |
+
},
|
| 951 |
+
{
|
| 952 |
+
"epoch": 0.026497982766349553,
|
| 953 |
+
"grad_norm": 0.1966014802455902,
|
| 954 |
+
"learning_rate": 8.693747451206232e-05,
|
| 955 |
+
"loss": 0.7174,
|
| 956 |
+
"step": 133
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 0.026697215719480003,
|
| 960 |
+
"grad_norm": 0.25692620873451233,
|
| 961 |
+
"learning_rate": 8.671612547178428e-05,
|
| 962 |
+
"loss": 0.6709,
|
| 963 |
+
"step": 134
|
| 964 |
+
},
|
| 965 |
+
{
|
| 966 |
+
"epoch": 0.02689644867261045,
|
| 967 |
+
"grad_norm": 0.19534507393836975,
|
| 968 |
+
"learning_rate": 8.649320363489179e-05,
|
| 969 |
+
"loss": 0.5871,
|
| 970 |
+
"step": 135
|
| 971 |
+
},
|
| 972 |
+
{
|
| 973 |
+
"epoch": 0.027095681625740897,
|
| 974 |
+
"grad_norm": 0.25263822078704834,
|
| 975 |
+
"learning_rate": 8.626871855061438e-05,
|
| 976 |
+
"loss": 0.7815,
|
| 977 |
+
"step": 136
|
| 978 |
+
},
|
| 979 |
+
{
|
| 980 |
+
"epoch": 0.027294914578871347,
|
| 981 |
+
"grad_norm": 0.24761663377285004,
|
| 982 |
+
"learning_rate": 8.604267983514594e-05,
|
| 983 |
+
"loss": 0.7253,
|
| 984 |
+
"step": 137
|
| 985 |
+
},
|
| 986 |
+
{
|
| 987 |
+
"epoch": 0.027494147532001793,
|
| 988 |
+
"grad_norm": 0.2336164116859436,
|
| 989 |
+
"learning_rate": 8.581509717123273e-05,
|
| 990 |
+
"loss": 0.6871,
|
| 991 |
+
"step": 138
|
| 992 |
+
},
|
| 993 |
+
{
|
| 994 |
+
"epoch": 0.02769338048513224,
|
| 995 |
+
"grad_norm": 0.2253238558769226,
|
| 996 |
+
"learning_rate": 8.558598030775857e-05,
|
| 997 |
+
"loss": 0.6806,
|
| 998 |
+
"step": 139
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 0.02789261343826269,
|
| 1002 |
+
"grad_norm": 0.2473248988389969,
|
| 1003 |
+
"learning_rate": 8.535533905932738e-05,
|
| 1004 |
+
"loss": 0.8147,
|
| 1005 |
+
"step": 140
|
| 1006 |
+
},
|
| 1007 |
+
{
|
| 1008 |
+
"epoch": 0.028091846391393137,
|
| 1009 |
+
"grad_norm": 0.2515900731086731,
|
| 1010 |
+
"learning_rate": 8.51231833058426e-05,
|
| 1011 |
+
"loss": 0.8055,
|
| 1012 |
+
"step": 141
|
| 1013 |
+
},
|
| 1014 |
+
{
|
| 1015 |
+
"epoch": 0.028291079344523583,
|
| 1016 |
+
"grad_norm": 0.2372109442949295,
|
| 1017 |
+
"learning_rate": 8.488952299208401e-05,
|
| 1018 |
+
"loss": 0.6404,
|
| 1019 |
+
"step": 142
|
| 1020 |
+
},
|
| 1021 |
+
{
|
| 1022 |
+
"epoch": 0.028490312297654034,
|
| 1023 |
+
"grad_norm": 0.24354617297649384,
|
| 1024 |
+
"learning_rate": 8.46543681272818e-05,
|
| 1025 |
+
"loss": 0.7355,
|
| 1026 |
+
"step": 143
|
| 1027 |
+
},
|
| 1028 |
+
{
|
| 1029 |
+
"epoch": 0.02868954525078448,
|
| 1030 |
+
"grad_norm": 0.2426522821187973,
|
| 1031 |
+
"learning_rate": 8.44177287846877e-05,
|
| 1032 |
+
"loss": 0.7561,
|
| 1033 |
+
"step": 144
|
| 1034 |
+
},
|
| 1035 |
+
{
|
| 1036 |
+
"epoch": 0.028888778203914927,
|
| 1037 |
+
"grad_norm": 0.24764999747276306,
|
| 1038 |
+
"learning_rate": 8.417961510114356e-05,
|
| 1039 |
+
"loss": 0.7203,
|
| 1040 |
+
"step": 145
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 0.029088011157045374,
|
| 1044 |
+
"grad_norm": 0.2602037191390991,
|
| 1045 |
+
"learning_rate": 8.39400372766471e-05,
|
| 1046 |
+
"loss": 0.7111,
|
| 1047 |
+
"step": 146
|
| 1048 |
+
},
|
| 1049 |
+
{
|
| 1050 |
+
"epoch": 0.029287244110175824,
|
| 1051 |
+
"grad_norm": 0.25291872024536133,
|
| 1052 |
+
"learning_rate": 8.36990055739149e-05,
|
| 1053 |
+
"loss": 0.7011,
|
| 1054 |
+
"step": 147
|
| 1055 |
+
},
|
| 1056 |
+
{
|
| 1057 |
+
"epoch": 0.02948647706330627,
|
| 1058 |
+
"grad_norm": 0.2726818323135376,
|
| 1059 |
+
"learning_rate": 8.345653031794292e-05,
|
| 1060 |
+
"loss": 0.8132,
|
| 1061 |
+
"step": 148
|
| 1062 |
+
},
|
| 1063 |
+
{
|
| 1064 |
+
"epoch": 0.029685710016436717,
|
| 1065 |
+
"grad_norm": 0.2830042839050293,
|
| 1066 |
+
"learning_rate": 8.321262189556409e-05,
|
| 1067 |
+
"loss": 0.7577,
|
| 1068 |
+
"step": 149
|
| 1069 |
+
},
|
| 1070 |
+
{
|
| 1071 |
+
"epoch": 0.029884942969567167,
|
| 1072 |
+
"grad_norm": 0.2831919193267822,
|
| 1073 |
+
"learning_rate": 8.296729075500344e-05,
|
| 1074 |
+
"loss": 0.7566,
|
| 1075 |
+
"step": 150
|
| 1076 |
+
},
|
| 1077 |
+
{
|
| 1078 |
+
"epoch": 0.030084175922697614,
|
| 1079 |
+
"grad_norm": 0.33581486344337463,
|
| 1080 |
+
"learning_rate": 8.272054740543052e-05,
|
| 1081 |
+
"loss": 0.7664,
|
| 1082 |
+
"step": 151
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 0.03028340887582806,
|
| 1086 |
+
"grad_norm": 0.3191780149936676,
|
| 1087 |
+
"learning_rate": 8.247240241650918e-05,
|
| 1088 |
+
"loss": 0.6583,
|
| 1089 |
+
"step": 152
|
| 1090 |
+
},
|
| 1091 |
+
{
|
| 1092 |
+
"epoch": 0.03048264182895851,
|
| 1093 |
+
"grad_norm": 0.22133736312389374,
|
| 1094 |
+
"learning_rate": 8.222286641794488e-05,
|
| 1095 |
+
"loss": 0.6328,
|
| 1096 |
+
"step": 153
|
| 1097 |
+
},
|
| 1098 |
+
{
|
| 1099 |
+
"epoch": 0.030681874782088957,
|
| 1100 |
+
"grad_norm": 0.21831941604614258,
|
| 1101 |
+
"learning_rate": 8.197195009902924e-05,
|
| 1102 |
+
"loss": 0.698,
|
| 1103 |
+
"step": 154
|
| 1104 |
+
},
|
| 1105 |
+
{
|
| 1106 |
+
"epoch": 0.030881107735219404,
|
| 1107 |
+
"grad_norm": 0.1761123090982437,
|
| 1108 |
+
"learning_rate": 8.171966420818228e-05,
|
| 1109 |
+
"loss": 0.5441,
|
| 1110 |
+
"step": 155
|
| 1111 |
+
},
|
| 1112 |
+
{
|
| 1113 |
+
"epoch": 0.031080340688349854,
|
| 1114 |
+
"grad_norm": 0.20375514030456543,
|
| 1115 |
+
"learning_rate": 8.146601955249188e-05,
|
| 1116 |
+
"loss": 0.6309,
|
| 1117 |
+
"step": 156
|
| 1118 |
+
},
|
| 1119 |
+
{
|
| 1120 |
+
"epoch": 0.0312795736414803,
|
| 1121 |
+
"grad_norm": 0.2410995066165924,
|
| 1122 |
+
"learning_rate": 8.121102699725089e-05,
|
| 1123 |
+
"loss": 0.6337,
|
| 1124 |
+
"step": 157
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 0.03147880659461075,
|
| 1128 |
+
"grad_norm": 0.2229624092578888,
|
| 1129 |
+
"learning_rate": 8.095469746549172e-05,
|
| 1130 |
+
"loss": 0.7014,
|
| 1131 |
+
"step": 158
|
| 1132 |
+
},
|
| 1133 |
+
{
|
| 1134 |
+
"epoch": 0.0316780395477412,
|
| 1135 |
+
"grad_norm": 0.23791897296905518,
|
| 1136 |
+
"learning_rate": 8.069704193751832e-05,
|
| 1137 |
+
"loss": 0.6582,
|
| 1138 |
+
"step": 159
|
| 1139 |
+
},
|
| 1140 |
+
{
|
| 1141 |
+
"epoch": 0.03187727250087164,
|
| 1142 |
+
"grad_norm": 0.22524884343147278,
|
| 1143 |
+
"learning_rate": 8.043807145043604e-05,
|
| 1144 |
+
"loss": 0.7178,
|
| 1145 |
+
"step": 160
|
| 1146 |
+
},
|
| 1147 |
+
{
|
| 1148 |
+
"epoch": 0.03207650545400209,
|
| 1149 |
+
"grad_norm": 0.21897757053375244,
|
| 1150 |
+
"learning_rate": 8.017779709767858e-05,
|
| 1151 |
+
"loss": 0.7032,
|
| 1152 |
+
"step": 161
|
| 1153 |
+
},
|
| 1154 |
+
{
|
| 1155 |
+
"epoch": 0.03227573840713254,
|
| 1156 |
+
"grad_norm": 0.19019927084445953,
|
| 1157 |
+
"learning_rate": 7.991623002853296e-05,
|
| 1158 |
+
"loss": 0.5842,
|
| 1159 |
+
"step": 162
|
| 1160 |
+
},
|
| 1161 |
+
{
|
| 1162 |
+
"epoch": 0.032474971360262984,
|
| 1163 |
+
"grad_norm": 0.21591834723949432,
|
| 1164 |
+
"learning_rate": 7.965338144766186e-05,
|
| 1165 |
+
"loss": 0.7243,
|
| 1166 |
+
"step": 163
|
| 1167 |
+
},
|
| 1168 |
+
{
|
| 1169 |
+
"epoch": 0.032674204313393435,
|
| 1170 |
+
"grad_norm": 0.20649899542331696,
|
| 1171 |
+
"learning_rate": 7.938926261462366e-05,
|
| 1172 |
+
"loss": 0.6814,
|
| 1173 |
+
"step": 164
|
| 1174 |
+
},
|
| 1175 |
+
{
|
| 1176 |
+
"epoch": 0.032873437266523885,
|
| 1177 |
+
"grad_norm": 0.24540702998638153,
|
| 1178 |
+
"learning_rate": 7.912388484339012e-05,
|
| 1179 |
+
"loss": 0.7356,
|
| 1180 |
+
"step": 165
|
| 1181 |
+
},
|
| 1182 |
+
{
|
| 1183 |
+
"epoch": 0.03307267021965433,
|
| 1184 |
+
"grad_norm": 0.22577622532844543,
|
| 1185 |
+
"learning_rate": 7.88572595018617e-05,
|
| 1186 |
+
"loss": 0.6468,
|
| 1187 |
+
"step": 166
|
| 1188 |
+
},
|
| 1189 |
+
{
|
| 1190 |
+
"epoch": 0.03327190317278478,
|
| 1191 |
+
"grad_norm": 0.2168670892715454,
|
| 1192 |
+
"learning_rate": 7.858939801138061e-05,
|
| 1193 |
+
"loss": 0.6448,
|
| 1194 |
+
"step": 167
|
| 1195 |
+
},
|
| 1196 |
+
{
|
| 1197 |
+
"epoch": 0.03347113612591523,
|
| 1198 |
+
"grad_norm": 0.22892935574054718,
|
| 1199 |
+
"learning_rate": 7.832031184624164e-05,
|
| 1200 |
+
"loss": 0.6375,
|
| 1201 |
+
"step": 168
|
| 1202 |
+
},
|
| 1203 |
+
{
|
| 1204 |
+
"epoch": 0.03367036907904567,
|
| 1205 |
+
"grad_norm": 0.2086174190044403,
|
| 1206 |
+
"learning_rate": 7.80500125332005e-05,
|
| 1207 |
+
"loss": 0.6993,
|
| 1208 |
+
"step": 169
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 0.03386960203217612,
|
| 1212 |
+
"grad_norm": 0.19050797820091248,
|
| 1213 |
+
"learning_rate": 7.777851165098012e-05,
|
| 1214 |
+
"loss": 0.5909,
|
| 1215 |
+
"step": 170
|
| 1216 |
+
},
|
| 1217 |
+
{
|
| 1218 |
+
"epoch": 0.03406883498530657,
|
| 1219 |
+
"grad_norm": 0.22635716199874878,
|
| 1220 |
+
"learning_rate": 7.750582082977467e-05,
|
| 1221 |
+
"loss": 0.6799,
|
| 1222 |
+
"step": 171
|
| 1223 |
+
},
|
| 1224 |
+
{
|
| 1225 |
+
"epoch": 0.034268067938437015,
|
| 1226 |
+
"grad_norm": 0.2369690090417862,
|
| 1227 |
+
"learning_rate": 7.723195175075136e-05,
|
| 1228 |
+
"loss": 0.6823,
|
| 1229 |
+
"step": 172
|
| 1230 |
+
},
|
| 1231 |
+
{
|
| 1232 |
+
"epoch": 0.034467300891567465,
|
| 1233 |
+
"grad_norm": 0.2258961796760559,
|
| 1234 |
+
"learning_rate": 7.695691614555003e-05,
|
| 1235 |
+
"loss": 0.6768,
|
| 1236 |
+
"step": 173
|
| 1237 |
+
},
|
| 1238 |
+
{
|
| 1239 |
+
"epoch": 0.034666533844697915,
|
| 1240 |
+
"grad_norm": 0.2175053060054779,
|
| 1241 |
+
"learning_rate": 7.668072579578058e-05,
|
| 1242 |
+
"loss": 0.6501,
|
| 1243 |
+
"step": 174
|
| 1244 |
+
},
|
| 1245 |
+
{
|
| 1246 |
+
"epoch": 0.03486576679782836,
|
| 1247 |
+
"grad_norm": 0.23859256505966187,
|
| 1248 |
+
"learning_rate": 7.64033925325184e-05,
|
| 1249 |
+
"loss": 0.7198,
|
| 1250 |
+
"step": 175
|
| 1251 |
+
},
|
| 1252 |
+
{
|
| 1253 |
+
"epoch": 0.03506499975095881,
|
| 1254 |
+
"grad_norm": 0.2090621143579483,
|
| 1255 |
+
"learning_rate": 7.612492823579745e-05,
|
| 1256 |
+
"loss": 0.6075,
|
| 1257 |
+
"step": 176
|
| 1258 |
+
},
|
| 1259 |
+
{
|
| 1260 |
+
"epoch": 0.03526423270408926,
|
| 1261 |
+
"grad_norm": 0.23009976744651794,
|
| 1262 |
+
"learning_rate": 7.584534483410137e-05,
|
| 1263 |
+
"loss": 0.6993,
|
| 1264 |
+
"step": 177
|
| 1265 |
+
},
|
| 1266 |
+
{
|
| 1267 |
+
"epoch": 0.0354634656572197,
|
| 1268 |
+
"grad_norm": 0.19365736842155457,
|
| 1269 |
+
"learning_rate": 7.55646543038526e-05,
|
| 1270 |
+
"loss": 0.61,
|
| 1271 |
+
"step": 178
|
| 1272 |
+
},
|
| 1273 |
+
{
|
| 1274 |
+
"epoch": 0.03566269861035015,
|
| 1275 |
+
"grad_norm": 0.24506577849388123,
|
| 1276 |
+
"learning_rate": 7.528286866889924e-05,
|
| 1277 |
+
"loss": 0.68,
|
| 1278 |
+
"step": 179
|
| 1279 |
+
},
|
| 1280 |
+
{
|
| 1281 |
+
"epoch": 0.0358619315634806,
|
| 1282 |
+
"grad_norm": 0.2776792049407959,
|
| 1283 |
+
"learning_rate": 7.500000000000001e-05,
|
| 1284 |
+
"loss": 0.7224,
|
| 1285 |
+
"step": 180
|
| 1286 |
+
},
|
| 1287 |
+
{
|
| 1288 |
+
"epoch": 0.036061164516611045,
|
| 1289 |
+
"grad_norm": 0.21556320786476135,
|
| 1290 |
+
"learning_rate": 7.471606041430723e-05,
|
| 1291 |
+
"loss": 0.6742,
|
| 1292 |
+
"step": 181
|
| 1293 |
+
},
|
| 1294 |
+
{
|
| 1295 |
+
"epoch": 0.036260397469741495,
|
| 1296 |
+
"grad_norm": 0.2549598217010498,
|
| 1297 |
+
"learning_rate": 7.443106207484776e-05,
|
| 1298 |
+
"loss": 0.7866,
|
| 1299 |
+
"step": 182
|
| 1300 |
+
},
|
| 1301 |
+
{
|
| 1302 |
+
"epoch": 0.036459630422871946,
|
| 1303 |
+
"grad_norm": 0.2427287995815277,
|
| 1304 |
+
"learning_rate": 7.414501719000187e-05,
|
| 1305 |
+
"loss": 0.7006,
|
| 1306 |
+
"step": 183
|
| 1307 |
+
},
|
| 1308 |
+
{
|
| 1309 |
+
"epoch": 0.03665886337600239,
|
| 1310 |
+
"grad_norm": 0.2485671192407608,
|
| 1311 |
+
"learning_rate": 7.385793801298042e-05,
|
| 1312 |
+
"loss": 0.678,
|
| 1313 |
+
"step": 184
|
| 1314 |
+
},
|
| 1315 |
+
{
|
| 1316 |
+
"epoch": 0.03685809632913284,
|
| 1317 |
+
"grad_norm": 0.23628251254558563,
|
| 1318 |
+
"learning_rate": 7.35698368412999e-05,
|
| 1319 |
+
"loss": 0.7176,
|
| 1320 |
+
"step": 185
|
| 1321 |
+
},
|
| 1322 |
+
{
|
| 1323 |
+
"epoch": 0.03705732928226329,
|
| 1324 |
+
"grad_norm": 0.245437353849411,
|
| 1325 |
+
"learning_rate": 7.328072601625557e-05,
|
| 1326 |
+
"loss": 0.6784,
|
| 1327 |
+
"step": 186
|
| 1328 |
+
},
|
| 1329 |
+
{
|
| 1330 |
+
"epoch": 0.03725656223539373,
|
| 1331 |
+
"grad_norm": 0.26534438133239746,
|
| 1332 |
+
"learning_rate": 7.2990617922393e-05,
|
| 1333 |
+
"loss": 0.7389,
|
| 1334 |
+
"step": 187
|
| 1335 |
+
},
|
| 1336 |
+
{
|
| 1337 |
+
"epoch": 0.03745579518852418,
|
| 1338 |
+
"grad_norm": 0.26061758399009705,
|
| 1339 |
+
"learning_rate": 7.269952498697734e-05,
|
| 1340 |
+
"loss": 0.637,
|
| 1341 |
+
"step": 188
|
| 1342 |
+
},
|
| 1343 |
+
{
|
| 1344 |
+
"epoch": 0.03765502814165463,
|
| 1345 |
+
"grad_norm": 0.247264102101326,
|
| 1346 |
+
"learning_rate": 7.240745967946113e-05,
|
| 1347 |
+
"loss": 0.7007,
|
| 1348 |
+
"step": 189
|
| 1349 |
+
},
|
| 1350 |
+
{
|
| 1351 |
+
"epoch": 0.037854261094785076,
|
| 1352 |
+
"grad_norm": 0.2501027584075928,
|
| 1353 |
+
"learning_rate": 7.211443451095007e-05,
|
| 1354 |
+
"loss": 0.7552,
|
| 1355 |
+
"step": 190
|
| 1356 |
+
},
|
| 1357 |
+
{
|
| 1358 |
+
"epoch": 0.038053494047915526,
|
| 1359 |
+
"grad_norm": 0.24374301731586456,
|
| 1360 |
+
"learning_rate": 7.18204620336671e-05,
|
| 1361 |
+
"loss": 0.7168,
|
| 1362 |
+
"step": 191
|
| 1363 |
+
},
|
| 1364 |
+
{
|
| 1365 |
+
"epoch": 0.038252727001045976,
|
| 1366 |
+
"grad_norm": 0.2584417760372162,
|
| 1367 |
+
"learning_rate": 7.152555484041476e-05,
|
| 1368 |
+
"loss": 0.6993,
|
| 1369 |
+
"step": 192
|
| 1370 |
+
},
|
| 1371 |
+
{
|
| 1372 |
+
"epoch": 0.03845195995417642,
|
| 1373 |
+
"grad_norm": 0.2678215205669403,
|
| 1374 |
+
"learning_rate": 7.122972556403567e-05,
|
| 1375 |
+
"loss": 0.7069,
|
| 1376 |
+
"step": 193
|
| 1377 |
+
},
|
| 1378 |
+
{
|
| 1379 |
+
"epoch": 0.03865119290730687,
|
| 1380 |
+
"grad_norm": 0.27493688464164734,
|
| 1381 |
+
"learning_rate": 7.09329868768714e-05,
|
| 1382 |
+
"loss": 0.7271,
|
| 1383 |
+
"step": 194
|
| 1384 |
+
},
|
| 1385 |
+
{
|
| 1386 |
+
"epoch": 0.03885042586043732,
|
| 1387 |
+
"grad_norm": 0.28842246532440186,
|
| 1388 |
+
"learning_rate": 7.063535149021973e-05,
|
| 1389 |
+
"loss": 0.7254,
|
| 1390 |
+
"step": 195
|
| 1391 |
+
},
|
| 1392 |
+
{
|
| 1393 |
+
"epoch": 0.03904965881356776,
|
| 1394 |
+
"grad_norm": 0.26705753803253174,
|
| 1395 |
+
"learning_rate": 7.033683215379002e-05,
|
| 1396 |
+
"loss": 0.7115,
|
| 1397 |
+
"step": 196
|
| 1398 |
+
},
|
| 1399 |
+
{
|
| 1400 |
+
"epoch": 0.03924889176669821,
|
| 1401 |
+
"grad_norm": 0.25876060128211975,
|
| 1402 |
+
"learning_rate": 7.003744165515705e-05,
|
| 1403 |
+
"loss": 0.7228,
|
| 1404 |
+
"step": 197
|
| 1405 |
+
},
|
| 1406 |
+
{
|
| 1407 |
+
"epoch": 0.03944812471982866,
|
| 1408 |
+
"grad_norm": 0.30024516582489014,
|
| 1409 |
+
"learning_rate": 6.973719281921335e-05,
|
| 1410 |
+
"loss": 0.7615,
|
| 1411 |
+
"step": 198
|
| 1412 |
+
},
|
| 1413 |
+
{
|
| 1414 |
+
"epoch": 0.039647357672959106,
|
| 1415 |
+
"grad_norm": 0.27740225195884705,
|
| 1416 |
+
"learning_rate": 6.943609850761979e-05,
|
| 1417 |
+
"loss": 0.7928,
|
| 1418 |
+
"step": 199
|
| 1419 |
+
},
|
| 1420 |
+
{
|
| 1421 |
+
"epoch": 0.039846590626089556,
|
| 1422 |
+
"grad_norm": 0.3024348020553589,
|
| 1423 |
+
"learning_rate": 6.91341716182545e-05,
|
| 1424 |
+
"loss": 0.7479,
|
| 1425 |
+
"step": 200
|
| 1426 |
+
},
|
| 1427 |
+
{
|
| 1428 |
+
"epoch": 0.039846590626089556,
|
| 1429 |
+
"eval_loss": 0.7008334398269653,
|
| 1430 |
+
"eval_runtime": 265.1826,
|
| 1431 |
+
"eval_samples_per_second": 31.88,
|
| 1432 |
+
"eval_steps_per_second": 7.972,
|
| 1433 |
+
"step": 200
|
| 1434 |
}
|
| 1435 |
],
|
| 1436 |
"logging_steps": 1,
|
|
|
|
| 1459 |
"attributes": {}
|
| 1460 |
}
|
| 1461 |
},
|
| 1462 |
+
"total_flos": 6.651170610846106e+16,
|
| 1463 |
"train_batch_size": 8,
|
| 1464 |
"trial_name": null,
|
| 1465 |
"trial_params": null
|