Training in progress, step 900, 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 35237104
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:c1cab7e9a8a987404549f49afdd2936be09b05aa346de6bd994dcd584e6c7c99
|
3 |
size 35237104
|
last-checkpoint/optimizer.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 18810356
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:a3af5b950a75d8130ae2831e845bdb2d9c0f2568b854448eb1a41e6ca5caf699
|
3 |
size 18810356
|
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 sha256:31fd0ef2088134d693702f76cf93ec3d3456380164b3e8cc27330c341fd530f6
|
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 sha256:0428512ada8c2471b2f37ecbdd4efa5f13e3ba0e777fddbfec0396eebc36c01a
|
3 |
size 1064
|
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": 250,
|
6 |
-
"global_step":
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
@@ -5639,6 +5639,706 @@
|
|
5639 |
"learning_rate": 1.0542974530180327e-05,
|
5640 |
"loss": 3.083,
|
5641 |
"step": 800
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
5642 |
}
|
5643 |
],
|
5644 |
"logging_steps": 1,
|
@@ -5658,7 +6358,7 @@
|
|
5658 |
"attributes": {}
|
5659 |
}
|
5660 |
},
|
5661 |
-
"total_flos":
|
5662 |
"train_batch_size": 2,
|
5663 |
"trial_name": null,
|
5664 |
"trial_params": null
|
|
|
1 |
{
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
+
"epoch": 0.32617559119325906,
|
5 |
"eval_steps": 250,
|
6 |
+
"global_step": 900,
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
|
|
5639 |
"learning_rate": 1.0542974530180327e-05,
|
5640 |
"loss": 3.083,
|
5641 |
"step": 800
|
5642 |
+
},
|
5643 |
+
{
|
5644 |
+
"epoch": 0.29029627616200054,
|
5645 |
+
"grad_norm": 6.44810152053833,
|
5646 |
+
"learning_rate": 1.0441632244932237e-05,
|
5647 |
+
"loss": 2.2946,
|
5648 |
+
"step": 801
|
5649 |
+
},
|
5650 |
+
{
|
5651 |
+
"epoch": 0.2906586934855486,
|
5652 |
+
"grad_norm": 10.6620454788208,
|
5653 |
+
"learning_rate": 1.0340722563656107e-05,
|
5654 |
+
"loss": 3.3972,
|
5655 |
+
"step": 802
|
5656 |
+
},
|
5657 |
+
{
|
5658 |
+
"epoch": 0.29102111080909665,
|
5659 |
+
"grad_norm": 5.806639671325684,
|
5660 |
+
"learning_rate": 1.0240246589884044e-05,
|
5661 |
+
"loss": 2.2731,
|
5662 |
+
"step": 803
|
5663 |
+
},
|
5664 |
+
{
|
5665 |
+
"epoch": 0.29138352813264473,
|
5666 |
+
"grad_norm": 5.967702865600586,
|
5667 |
+
"learning_rate": 1.0140205422405214e-05,
|
5668 |
+
"loss": 3.1209,
|
5669 |
+
"step": 804
|
5670 |
+
},
|
5671 |
+
{
|
5672 |
+
"epoch": 0.2917459454561928,
|
5673 |
+
"grad_norm": 7.064982891082764,
|
5674 |
+
"learning_rate": 1.0040600155253765e-05,
|
5675 |
+
"loss": 3.3447,
|
5676 |
+
"step": 805
|
5677 |
+
},
|
5678 |
+
{
|
5679 |
+
"epoch": 0.2921083627797409,
|
5680 |
+
"grad_norm": 5.999367713928223,
|
5681 |
+
"learning_rate": 9.941431877696955e-06,
|
5682 |
+
"loss": 2.46,
|
5683 |
+
"step": 806
|
5684 |
+
},
|
5685 |
+
{
|
5686 |
+
"epoch": 0.2924707801032889,
|
5687 |
+
"grad_norm": 6.495697498321533,
|
5688 |
+
"learning_rate": 9.842701674223187e-06,
|
5689 |
+
"loss": 1.9858,
|
5690 |
+
"step": 807
|
5691 |
+
},
|
5692 |
+
{
|
5693 |
+
"epoch": 0.292833197426837,
|
5694 |
+
"grad_norm": 7.841910362243652,
|
5695 |
+
"learning_rate": 9.744410624530148e-06,
|
5696 |
+
"loss": 3.1342,
|
5697 |
+
"step": 808
|
5698 |
+
},
|
5699 |
+
{
|
5700 |
+
"epoch": 0.2931956147503851,
|
5701 |
+
"grad_norm": 7.2909111976623535,
|
5702 |
+
"learning_rate": 9.646559803512994e-06,
|
5703 |
+
"loss": 4.0895,
|
5704 |
+
"step": 809
|
5705 |
+
},
|
5706 |
+
{
|
5707 |
+
"epoch": 0.29355803207393316,
|
5708 |
+
"grad_norm": 5.502638816833496,
|
5709 |
+
"learning_rate": 9.549150281252633e-06,
|
5710 |
+
"loss": 2.1715,
|
5711 |
+
"step": 810
|
5712 |
+
},
|
5713 |
+
{
|
5714 |
+
"epoch": 0.2939204493974812,
|
5715 |
+
"grad_norm": 5.319310188293457,
|
5716 |
+
"learning_rate": 9.452183123004e-06,
|
5717 |
+
"loss": 2.8537,
|
5718 |
+
"step": 811
|
5719 |
+
},
|
5720 |
+
{
|
5721 |
+
"epoch": 0.29428286672102927,
|
5722 |
+
"grad_norm": 6.386750221252441,
|
5723 |
+
"learning_rate": 9.355659389184396e-06,
|
5724 |
+
"loss": 3.2566,
|
5725 |
+
"step": 812
|
5726 |
+
},
|
5727 |
+
{
|
5728 |
+
"epoch": 0.29464528404457735,
|
5729 |
+
"grad_norm": 7.152513027191162,
|
5730 |
+
"learning_rate": 9.259580135361929e-06,
|
5731 |
+
"loss": 2.6424,
|
5732 |
+
"step": 813
|
5733 |
+
},
|
5734 |
+
{
|
5735 |
+
"epoch": 0.2950077013681254,
|
5736 |
+
"grad_norm": 7.873605728149414,
|
5737 |
+
"learning_rate": 9.163946412243896e-06,
|
5738 |
+
"loss": 3.1727,
|
5739 |
+
"step": 814
|
5740 |
+
},
|
5741 |
+
{
|
5742 |
+
"epoch": 0.29537011869167346,
|
5743 |
+
"grad_norm": 5.937596321105957,
|
5744 |
+
"learning_rate": 9.068759265665384e-06,
|
5745 |
+
"loss": 2.6007,
|
5746 |
+
"step": 815
|
5747 |
+
},
|
5748 |
+
{
|
5749 |
+
"epoch": 0.29573253601522154,
|
5750 |
+
"grad_norm": 6.4454121589660645,
|
5751 |
+
"learning_rate": 8.974019736577777e-06,
|
5752 |
+
"loss": 2.9121,
|
5753 |
+
"step": 816
|
5754 |
+
},
|
5755 |
+
{
|
5756 |
+
"epoch": 0.2960949533387696,
|
5757 |
+
"grad_norm": 8.89317512512207,
|
5758 |
+
"learning_rate": 8.879728861037384e-06,
|
5759 |
+
"loss": 3.5852,
|
5760 |
+
"step": 817
|
5761 |
+
},
|
5762 |
+
{
|
5763 |
+
"epoch": 0.29645737066231764,
|
5764 |
+
"grad_norm": 6.11660623550415,
|
5765 |
+
"learning_rate": 8.785887670194138e-06,
|
5766 |
+
"loss": 2.1771,
|
5767 |
+
"step": 818
|
5768 |
+
},
|
5769 |
+
{
|
5770 |
+
"epoch": 0.2968197879858657,
|
5771 |
+
"grad_norm": 6.442231178283691,
|
5772 |
+
"learning_rate": 8.692497190280224e-06,
|
5773 |
+
"loss": 2.6609,
|
5774 |
+
"step": 819
|
5775 |
+
},
|
5776 |
+
{
|
5777 |
+
"epoch": 0.2971822053094138,
|
5778 |
+
"grad_norm": 6.466634750366211,
|
5779 |
+
"learning_rate": 8.599558442598998e-06,
|
5780 |
+
"loss": 2.2619,
|
5781 |
+
"step": 820
|
5782 |
+
},
|
5783 |
+
{
|
5784 |
+
"epoch": 0.29754462263296183,
|
5785 |
+
"grad_norm": 7.821259498596191,
|
5786 |
+
"learning_rate": 8.507072443513702e-06,
|
5787 |
+
"loss": 2.7053,
|
5788 |
+
"step": 821
|
5789 |
+
},
|
5790 |
+
{
|
5791 |
+
"epoch": 0.2979070399565099,
|
5792 |
+
"grad_norm": 6.689558506011963,
|
5793 |
+
"learning_rate": 8.415040204436426e-06,
|
5794 |
+
"loss": 2.8319,
|
5795 |
+
"step": 822
|
5796 |
+
},
|
5797 |
+
{
|
5798 |
+
"epoch": 0.298269457280058,
|
5799 |
+
"grad_norm": 8.634140968322754,
|
5800 |
+
"learning_rate": 8.323462731816961e-06,
|
5801 |
+
"loss": 2.8749,
|
5802 |
+
"step": 823
|
5803 |
+
},
|
5804 |
+
{
|
5805 |
+
"epoch": 0.2986318746036061,
|
5806 |
+
"grad_norm": 7.520872592926025,
|
5807 |
+
"learning_rate": 8.232341027131885e-06,
|
5808 |
+
"loss": 2.4212,
|
5809 |
+
"step": 824
|
5810 |
+
},
|
5811 |
+
{
|
5812 |
+
"epoch": 0.2989942919271541,
|
5813 |
+
"grad_norm": 6.704894542694092,
|
5814 |
+
"learning_rate": 8.141676086873572e-06,
|
5815 |
+
"loss": 3.3488,
|
5816 |
+
"step": 825
|
5817 |
+
},
|
5818 |
+
{
|
5819 |
+
"epoch": 0.2993567092507022,
|
5820 |
+
"grad_norm": 7.143224239349365,
|
5821 |
+
"learning_rate": 8.051468902539272e-06,
|
5822 |
+
"loss": 3.364,
|
5823 |
+
"step": 826
|
5824 |
+
},
|
5825 |
+
{
|
5826 |
+
"epoch": 0.29971912657425026,
|
5827 |
+
"grad_norm": 6.381522178649902,
|
5828 |
+
"learning_rate": 7.96172046062032e-06,
|
5829 |
+
"loss": 2.5679,
|
5830 |
+
"step": 827
|
5831 |
+
},
|
5832 |
+
{
|
5833 |
+
"epoch": 0.30008154389779834,
|
5834 |
+
"grad_norm": 5.404728889465332,
|
5835 |
+
"learning_rate": 7.872431742591268e-06,
|
5836 |
+
"loss": 3.0229,
|
5837 |
+
"step": 828
|
5838 |
+
},
|
5839 |
+
{
|
5840 |
+
"epoch": 0.30044396122134637,
|
5841 |
+
"grad_norm": 5.889482498168945,
|
5842 |
+
"learning_rate": 7.783603724899257e-06,
|
5843 |
+
"loss": 2.9306,
|
5844 |
+
"step": 829
|
5845 |
+
},
|
5846 |
+
{
|
5847 |
+
"epoch": 0.30080637854489445,
|
5848 |
+
"grad_norm": 6.722096920013428,
|
5849 |
+
"learning_rate": 7.695237378953223e-06,
|
5850 |
+
"loss": 2.9732,
|
5851 |
+
"step": 830
|
5852 |
+
},
|
5853 |
+
{
|
5854 |
+
"epoch": 0.30116879586844253,
|
5855 |
+
"grad_norm": 4.966902732849121,
|
5856 |
+
"learning_rate": 7.607333671113409e-06,
|
5857 |
+
"loss": 2.5392,
|
5858 |
+
"step": 831
|
5859 |
+
},
|
5860 |
+
{
|
5861 |
+
"epoch": 0.30153121319199055,
|
5862 |
+
"grad_norm": 6.83373498916626,
|
5863 |
+
"learning_rate": 7.519893562680663e-06,
|
5864 |
+
"loss": 3.3694,
|
5865 |
+
"step": 832
|
5866 |
+
},
|
5867 |
+
{
|
5868 |
+
"epoch": 0.30189363051553864,
|
5869 |
+
"grad_norm": 6.037679195404053,
|
5870 |
+
"learning_rate": 7.432918009885997e-06,
|
5871 |
+
"loss": 2.0567,
|
5872 |
+
"step": 833
|
5873 |
+
},
|
5874 |
+
{
|
5875 |
+
"epoch": 0.3022560478390867,
|
5876 |
+
"grad_norm": 6.7791032791137695,
|
5877 |
+
"learning_rate": 7.3464079638801365e-06,
|
5878 |
+
"loss": 2.491,
|
5879 |
+
"step": 834
|
5880 |
+
},
|
5881 |
+
{
|
5882 |
+
"epoch": 0.3026184651626348,
|
5883 |
+
"grad_norm": 6.07615852355957,
|
5884 |
+
"learning_rate": 7.260364370723044e-06,
|
5885 |
+
"loss": 2.4496,
|
5886 |
+
"step": 835
|
5887 |
+
},
|
5888 |
+
{
|
5889 |
+
"epoch": 0.3029808824861828,
|
5890 |
+
"grad_norm": 7.163369655609131,
|
5891 |
+
"learning_rate": 7.174788171373731e-06,
|
5892 |
+
"loss": 2.7414,
|
5893 |
+
"step": 836
|
5894 |
+
},
|
5895 |
+
{
|
5896 |
+
"epoch": 0.3033432998097309,
|
5897 |
+
"grad_norm": 5.891061305999756,
|
5898 |
+
"learning_rate": 7.089680301679752e-06,
|
5899 |
+
"loss": 2.7737,
|
5900 |
+
"step": 837
|
5901 |
+
},
|
5902 |
+
{
|
5903 |
+
"epoch": 0.303705717133279,
|
5904 |
+
"grad_norm": 5.414796352386475,
|
5905 |
+
"learning_rate": 7.005041692367154e-06,
|
5906 |
+
"loss": 2.2463,
|
5907 |
+
"step": 838
|
5908 |
+
},
|
5909 |
+
{
|
5910 |
+
"epoch": 0.304068134456827,
|
5911 |
+
"grad_norm": 5.405869483947754,
|
5912 |
+
"learning_rate": 6.92087326903022e-06,
|
5913 |
+
"loss": 2.4379,
|
5914 |
+
"step": 839
|
5915 |
+
},
|
5916 |
+
{
|
5917 |
+
"epoch": 0.3044305517803751,
|
5918 |
+
"grad_norm": 6.097501277923584,
|
5919 |
+
"learning_rate": 6.837175952121306e-06,
|
5920 |
+
"loss": 2.9844,
|
5921 |
+
"step": 840
|
5922 |
+
},
|
5923 |
+
{
|
5924 |
+
"epoch": 0.3047929691039232,
|
5925 |
+
"grad_norm": 9.414335250854492,
|
5926 |
+
"learning_rate": 6.753950656940905e-06,
|
5927 |
+
"loss": 2.4858,
|
5928 |
+
"step": 841
|
5929 |
+
},
|
5930 |
+
{
|
5931 |
+
"epoch": 0.30515538642747125,
|
5932 |
+
"grad_norm": 8.032082557678223,
|
5933 |
+
"learning_rate": 6.671198293627479e-06,
|
5934 |
+
"loss": 2.3787,
|
5935 |
+
"step": 842
|
5936 |
+
},
|
5937 |
+
{
|
5938 |
+
"epoch": 0.3055178037510193,
|
5939 |
+
"grad_norm": 7.170987606048584,
|
5940 |
+
"learning_rate": 6.588919767147639e-06,
|
5941 |
+
"loss": 2.3088,
|
5942 |
+
"step": 843
|
5943 |
+
},
|
5944 |
+
{
|
5945 |
+
"epoch": 0.30588022107456736,
|
5946 |
+
"grad_norm": 7.258477210998535,
|
5947 |
+
"learning_rate": 6.5071159772861436e-06,
|
5948 |
+
"loss": 3.3815,
|
5949 |
+
"step": 844
|
5950 |
+
},
|
5951 |
+
{
|
5952 |
+
"epoch": 0.30624263839811544,
|
5953 |
+
"grad_norm": 6.890701770782471,
|
5954 |
+
"learning_rate": 6.425787818636131e-06,
|
5955 |
+
"loss": 3.248,
|
5956 |
+
"step": 845
|
5957 |
+
},
|
5958 |
+
{
|
5959 |
+
"epoch": 0.3066050557216635,
|
5960 |
+
"grad_norm": 9.972538948059082,
|
5961 |
+
"learning_rate": 6.344936180589351e-06,
|
5962 |
+
"loss": 3.0751,
|
5963 |
+
"step": 846
|
5964 |
+
},
|
5965 |
+
{
|
5966 |
+
"epoch": 0.30696747304521155,
|
5967 |
+
"grad_norm": 7.4007792472839355,
|
5968 |
+
"learning_rate": 6.264561947326331e-06,
|
5969 |
+
"loss": 3.0693,
|
5970 |
+
"step": 847
|
5971 |
+
},
|
5972 |
+
{
|
5973 |
+
"epoch": 0.30732989036875963,
|
5974 |
+
"grad_norm": 8.686583518981934,
|
5975 |
+
"learning_rate": 6.184665997806832e-06,
|
5976 |
+
"loss": 2.6327,
|
5977 |
+
"step": 848
|
5978 |
+
},
|
5979 |
+
{
|
5980 |
+
"epoch": 0.3076923076923077,
|
5981 |
+
"grad_norm": 9.083965301513672,
|
5982 |
+
"learning_rate": 6.1052492057601275e-06,
|
5983 |
+
"loss": 3.0008,
|
5984 |
+
"step": 849
|
5985 |
+
},
|
5986 |
+
{
|
5987 |
+
"epoch": 0.30805472501585573,
|
5988 |
+
"grad_norm": 6.180975437164307,
|
5989 |
+
"learning_rate": 6.026312439675552e-06,
|
5990 |
+
"loss": 2.9941,
|
5991 |
+
"step": 850
|
5992 |
+
},
|
5993 |
+
{
|
5994 |
+
"epoch": 0.3084171423394038,
|
5995 |
+
"grad_norm": 7.1970930099487305,
|
5996 |
+
"learning_rate": 5.947856562792925e-06,
|
5997 |
+
"loss": 3.0446,
|
5998 |
+
"step": 851
|
5999 |
+
},
|
6000 |
+
{
|
6001 |
+
"epoch": 0.3087795596629519,
|
6002 |
+
"grad_norm": 8.323010444641113,
|
6003 |
+
"learning_rate": 5.869882433093155e-06,
|
6004 |
+
"loss": 2.6583,
|
6005 |
+
"step": 852
|
6006 |
+
},
|
6007 |
+
{
|
6008 |
+
"epoch": 0.3091419769865,
|
6009 |
+
"grad_norm": 6.963425159454346,
|
6010 |
+
"learning_rate": 5.79239090328883e-06,
|
6011 |
+
"loss": 3.3398,
|
6012 |
+
"step": 853
|
6013 |
+
},
|
6014 |
+
{
|
6015 |
+
"epoch": 0.309504394310048,
|
6016 |
+
"grad_norm": 7.508121013641357,
|
6017 |
+
"learning_rate": 5.715382820814885e-06,
|
6018 |
+
"loss": 1.7714,
|
6019 |
+
"step": 854
|
6020 |
+
},
|
6021 |
+
{
|
6022 |
+
"epoch": 0.3098668116335961,
|
6023 |
+
"grad_norm": 7.211197376251221,
|
6024 |
+
"learning_rate": 5.6388590278194096e-06,
|
6025 |
+
"loss": 3.164,
|
6026 |
+
"step": 855
|
6027 |
+
},
|
6028 |
+
{
|
6029 |
+
"epoch": 0.31022922895714417,
|
6030 |
+
"grad_norm": 5.219383716583252,
|
6031 |
+
"learning_rate": 5.562820361154314e-06,
|
6032 |
+
"loss": 3.0525,
|
6033 |
+
"step": 856
|
6034 |
+
},
|
6035 |
+
{
|
6036 |
+
"epoch": 0.3105916462806922,
|
6037 |
+
"grad_norm": 6.372992992401123,
|
6038 |
+
"learning_rate": 5.48726765236629e-06,
|
6039 |
+
"loss": 2.2544,
|
6040 |
+
"step": 857
|
6041 |
+
},
|
6042 |
+
{
|
6043 |
+
"epoch": 0.31095406360424027,
|
6044 |
+
"grad_norm": 13.562681198120117,
|
6045 |
+
"learning_rate": 5.412201727687644e-06,
|
6046 |
+
"loss": 2.8972,
|
6047 |
+
"step": 858
|
6048 |
+
},
|
6049 |
+
{
|
6050 |
+
"epoch": 0.31131648092778835,
|
6051 |
+
"grad_norm": 7.206049919128418,
|
6052 |
+
"learning_rate": 5.337623408027293e-06,
|
6053 |
+
"loss": 3.3129,
|
6054 |
+
"step": 859
|
6055 |
+
},
|
6056 |
+
{
|
6057 |
+
"epoch": 0.31167889825133643,
|
6058 |
+
"grad_norm": 5.508146286010742,
|
6059 |
+
"learning_rate": 5.263533508961827e-06,
|
6060 |
+
"loss": 2.8893,
|
6061 |
+
"step": 860
|
6062 |
+
},
|
6063 |
+
{
|
6064 |
+
"epoch": 0.31204131557488446,
|
6065 |
+
"grad_norm": 5.100135803222656,
|
6066 |
+
"learning_rate": 5.1899328407264855e-06,
|
6067 |
+
"loss": 2.755,
|
6068 |
+
"step": 861
|
6069 |
+
},
|
6070 |
+
{
|
6071 |
+
"epoch": 0.31240373289843254,
|
6072 |
+
"grad_norm": 6.5922675132751465,
|
6073 |
+
"learning_rate": 5.116822208206396e-06,
|
6074 |
+
"loss": 2.2685,
|
6075 |
+
"step": 862
|
6076 |
+
},
|
6077 |
+
{
|
6078 |
+
"epoch": 0.3127661502219806,
|
6079 |
+
"grad_norm": 6.995744705200195,
|
6080 |
+
"learning_rate": 5.044202410927706e-06,
|
6081 |
+
"loss": 2.9639,
|
6082 |
+
"step": 863
|
6083 |
+
},
|
6084 |
+
{
|
6085 |
+
"epoch": 0.3131285675455287,
|
6086 |
+
"grad_norm": 9.096760749816895,
|
6087 |
+
"learning_rate": 4.972074243048897e-06,
|
6088 |
+
"loss": 3.1619,
|
6089 |
+
"step": 864
|
6090 |
+
},
|
6091 |
+
{
|
6092 |
+
"epoch": 0.3134909848690767,
|
6093 |
+
"grad_norm": 6.071923732757568,
|
6094 |
+
"learning_rate": 4.900438493352055e-06,
|
6095 |
+
"loss": 3.0337,
|
6096 |
+
"step": 865
|
6097 |
+
},
|
6098 |
+
{
|
6099 |
+
"epoch": 0.3138534021926248,
|
6100 |
+
"grad_norm": 6.915728569030762,
|
6101 |
+
"learning_rate": 4.829295945234258e-06,
|
6102 |
+
"loss": 2.7878,
|
6103 |
+
"step": 866
|
6104 |
+
},
|
6105 |
+
{
|
6106 |
+
"epoch": 0.3142158195161729,
|
6107 |
+
"grad_norm": 5.116779327392578,
|
6108 |
+
"learning_rate": 4.758647376699032e-06,
|
6109 |
+
"loss": 2.028,
|
6110 |
+
"step": 867
|
6111 |
+
},
|
6112 |
+
{
|
6113 |
+
"epoch": 0.3145782368397209,
|
6114 |
+
"grad_norm": 6.29951286315918,
|
6115 |
+
"learning_rate": 4.688493560347773e-06,
|
6116 |
+
"loss": 2.4187,
|
6117 |
+
"step": 868
|
6118 |
+
},
|
6119 |
+
{
|
6120 |
+
"epoch": 0.314940654163269,
|
6121 |
+
"grad_norm": 4.603816986083984,
|
6122 |
+
"learning_rate": 4.618835263371396e-06,
|
6123 |
+
"loss": 2.1837,
|
6124 |
+
"step": 869
|
6125 |
+
},
|
6126 |
+
{
|
6127 |
+
"epoch": 0.3153030714868171,
|
6128 |
+
"grad_norm": 7.818729877471924,
|
6129 |
+
"learning_rate": 4.549673247541875e-06,
|
6130 |
+
"loss": 3.9697,
|
6131 |
+
"step": 870
|
6132 |
+
},
|
6133 |
+
{
|
6134 |
+
"epoch": 0.31566548881036516,
|
6135 |
+
"grad_norm": 6.050262928009033,
|
6136 |
+
"learning_rate": 4.48100826920394e-06,
|
6137 |
+
"loss": 2.8657,
|
6138 |
+
"step": 871
|
6139 |
+
},
|
6140 |
+
{
|
6141 |
+
"epoch": 0.3160279061339132,
|
6142 |
+
"grad_norm": 5.011636257171631,
|
6143 |
+
"learning_rate": 4.412841079266777e-06,
|
6144 |
+
"loss": 2.155,
|
6145 |
+
"step": 872
|
6146 |
+
},
|
6147 |
+
{
|
6148 |
+
"epoch": 0.31639032345746126,
|
6149 |
+
"grad_norm": 7.263031005859375,
|
6150 |
+
"learning_rate": 4.3451724231958644e-06,
|
6151 |
+
"loss": 2.8195,
|
6152 |
+
"step": 873
|
6153 |
+
},
|
6154 |
+
{
|
6155 |
+
"epoch": 0.31675274078100935,
|
6156 |
+
"grad_norm": 6.203886985778809,
|
6157 |
+
"learning_rate": 4.27800304100478e-06,
|
6158 |
+
"loss": 2.9632,
|
6159 |
+
"step": 874
|
6160 |
+
},
|
6161 |
+
{
|
6162 |
+
"epoch": 0.31711515810455737,
|
6163 |
+
"grad_norm": 5.561065196990967,
|
6164 |
+
"learning_rate": 4.2113336672471245e-06,
|
6165 |
+
"loss": 2.5715,
|
6166 |
+
"step": 875
|
6167 |
+
},
|
6168 |
+
{
|
6169 |
+
"epoch": 0.31747757542810545,
|
6170 |
+
"grad_norm": 7.3997578620910645,
|
6171 |
+
"learning_rate": 4.145165031008508e-06,
|
6172 |
+
"loss": 2.8707,
|
6173 |
+
"step": 876
|
6174 |
+
},
|
6175 |
+
{
|
6176 |
+
"epoch": 0.31783999275165353,
|
6177 |
+
"grad_norm": 6.424815654754639,
|
6178 |
+
"learning_rate": 4.079497855898501e-06,
|
6179 |
+
"loss": 3.787,
|
6180 |
+
"step": 877
|
6181 |
+
},
|
6182 |
+
{
|
6183 |
+
"epoch": 0.3182024100752016,
|
6184 |
+
"grad_norm": 8.192680358886719,
|
6185 |
+
"learning_rate": 4.01433286004283e-06,
|
6186 |
+
"loss": 3.4674,
|
6187 |
+
"step": 878
|
6188 |
+
},
|
6189 |
+
{
|
6190 |
+
"epoch": 0.31856482739874964,
|
6191 |
+
"grad_norm": 6.513679504394531,
|
6192 |
+
"learning_rate": 3.949670756075447e-06,
|
6193 |
+
"loss": 2.7441,
|
6194 |
+
"step": 879
|
6195 |
+
},
|
6196 |
+
{
|
6197 |
+
"epoch": 0.3189272447222977,
|
6198 |
+
"grad_norm": 7.615200042724609,
|
6199 |
+
"learning_rate": 3.885512251130763e-06,
|
6200 |
+
"loss": 3.0239,
|
6201 |
+
"step": 880
|
6202 |
+
},
|
6203 |
+
{
|
6204 |
+
"epoch": 0.3192896620458458,
|
6205 |
+
"grad_norm": 7.766543865203857,
|
6206 |
+
"learning_rate": 3.821858046835913e-06,
|
6207 |
+
"loss": 3.1864,
|
6208 |
+
"step": 881
|
6209 |
+
},
|
6210 |
+
{
|
6211 |
+
"epoch": 0.3196520793693939,
|
6212 |
+
"grad_norm": 6.684014320373535,
|
6213 |
+
"learning_rate": 3.75870883930306e-06,
|
6214 |
+
"loss": 3.0098,
|
6215 |
+
"step": 882
|
6216 |
+
},
|
6217 |
+
{
|
6218 |
+
"epoch": 0.3200144966929419,
|
6219 |
+
"grad_norm": 7.500833034515381,
|
6220 |
+
"learning_rate": 3.696065319121833e-06,
|
6221 |
+
"loss": 3.0346,
|
6222 |
+
"step": 883
|
6223 |
+
},
|
6224 |
+
{
|
6225 |
+
"epoch": 0.32037691401649,
|
6226 |
+
"grad_norm": 6.329296588897705,
|
6227 |
+
"learning_rate": 3.6339281713517303e-06,
|
6228 |
+
"loss": 2.8556,
|
6229 |
+
"step": 884
|
6230 |
+
},
|
6231 |
+
{
|
6232 |
+
"epoch": 0.32073933134003807,
|
6233 |
+
"grad_norm": 5.90148401260376,
|
6234 |
+
"learning_rate": 3.5722980755146517e-06,
|
6235 |
+
"loss": 2.4994,
|
6236 |
+
"step": 885
|
6237 |
+
},
|
6238 |
+
{
|
6239 |
+
"epoch": 0.3211017486635861,
|
6240 |
+
"grad_norm": 7.304757595062256,
|
6241 |
+
"learning_rate": 3.511175705587433e-06,
|
6242 |
+
"loss": 3.3535,
|
6243 |
+
"step": 886
|
6244 |
+
},
|
6245 |
+
{
|
6246 |
+
"epoch": 0.3214641659871342,
|
6247 |
+
"grad_norm": 8.020987510681152,
|
6248 |
+
"learning_rate": 3.4505617299945336e-06,
|
6249 |
+
"loss": 3.9281,
|
6250 |
+
"step": 887
|
6251 |
+
},
|
6252 |
+
{
|
6253 |
+
"epoch": 0.32182658331068226,
|
6254 |
+
"grad_norm": 7.933343887329102,
|
6255 |
+
"learning_rate": 3.390456811600673e-06,
|
6256 |
+
"loss": 3.0785,
|
6257 |
+
"step": 888
|
6258 |
+
},
|
6259 |
+
{
|
6260 |
+
"epoch": 0.32218900063423034,
|
6261 |
+
"grad_norm": 5.5655670166015625,
|
6262 |
+
"learning_rate": 3.3308616077036115e-06,
|
6263 |
+
"loss": 3.0121,
|
6264 |
+
"step": 889
|
6265 |
+
},
|
6266 |
+
{
|
6267 |
+
"epoch": 0.32255141795777836,
|
6268 |
+
"grad_norm": 7.345324516296387,
|
6269 |
+
"learning_rate": 3.271776770026963e-06,
|
6270 |
+
"loss": 2.1397,
|
6271 |
+
"step": 890
|
6272 |
+
},
|
6273 |
+
{
|
6274 |
+
"epoch": 0.32291383528132644,
|
6275 |
+
"grad_norm": 6.862234592437744,
|
6276 |
+
"learning_rate": 3.213202944713023e-06,
|
6277 |
+
"loss": 2.6641,
|
6278 |
+
"step": 891
|
6279 |
+
},
|
6280 |
+
{
|
6281 |
+
"epoch": 0.3232762526048745,
|
6282 |
+
"grad_norm": 5.922866344451904,
|
6283 |
+
"learning_rate": 3.155140772315773e-06,
|
6284 |
+
"loss": 2.8116,
|
6285 |
+
"step": 892
|
6286 |
+
},
|
6287 |
+
{
|
6288 |
+
"epoch": 0.32363866992842255,
|
6289 |
+
"grad_norm": 5.328739643096924,
|
6290 |
+
"learning_rate": 3.0975908877938277e-06,
|
6291 |
+
"loss": 2.9905,
|
6292 |
+
"step": 893
|
6293 |
+
},
|
6294 |
+
{
|
6295 |
+
"epoch": 0.32400108725197063,
|
6296 |
+
"grad_norm": 5.276422500610352,
|
6297 |
+
"learning_rate": 3.040553920503503e-06,
|
6298 |
+
"loss": 2.2472,
|
6299 |
+
"step": 894
|
6300 |
+
},
|
6301 |
+
{
|
6302 |
+
"epoch": 0.3243635045755187,
|
6303 |
+
"grad_norm": 5.449382305145264,
|
6304 |
+
"learning_rate": 2.9840304941919415e-06,
|
6305 |
+
"loss": 2.3198,
|
6306 |
+
"step": 895
|
6307 |
+
},
|
6308 |
+
{
|
6309 |
+
"epoch": 0.3247259218990668,
|
6310 |
+
"grad_norm": 5.51635217666626,
|
6311 |
+
"learning_rate": 2.928021226990263e-06,
|
6312 |
+
"loss": 2.7019,
|
6313 |
+
"step": 896
|
6314 |
+
},
|
6315 |
+
{
|
6316 |
+
"epoch": 0.3250883392226148,
|
6317 |
+
"grad_norm": 5.101719856262207,
|
6318 |
+
"learning_rate": 2.8725267314068495e-06,
|
6319 |
+
"loss": 2.5853,
|
6320 |
+
"step": 897
|
6321 |
+
},
|
6322 |
+
{
|
6323 |
+
"epoch": 0.3254507565461629,
|
6324 |
+
"grad_norm": 8.838541030883789,
|
6325 |
+
"learning_rate": 2.817547614320615e-06,
|
6326 |
+
"loss": 2.5069,
|
6327 |
+
"step": 898
|
6328 |
+
},
|
6329 |
+
{
|
6330 |
+
"epoch": 0.325813173869711,
|
6331 |
+
"grad_norm": 6.860764503479004,
|
6332 |
+
"learning_rate": 2.7630844769743757e-06,
|
6333 |
+
"loss": 2.8399,
|
6334 |
+
"step": 899
|
6335 |
+
},
|
6336 |
+
{
|
6337 |
+
"epoch": 0.32617559119325906,
|
6338 |
+
"grad_norm": 5.290176868438721,
|
6339 |
+
"learning_rate": 2.7091379149682685e-06,
|
6340 |
+
"loss": 2.5194,
|
6341 |
+
"step": 900
|
6342 |
}
|
6343 |
],
|
6344 |
"logging_steps": 1,
|
|
|
6358 |
"attributes": {}
|
6359 |
}
|
6360 |
},
|
6361 |
+
"total_flos": 6.4885887074304e+16,
|
6362 |
"train_batch_size": 2,
|
6363 |
"trial_name": null,
|
6364 |
"trial_params": null
|