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 73911112
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f045355f938f666786ad36d5a7bc9028d0be96a54b23bd9a02def905f8ac3f5f
|
3 |
size 73911112
|
last-checkpoint/optimizer.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 37965684
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:9189f4679ea4da6f1b3935af0c0173ea9b0016a35be3117e47e4abc4073f0d74
|
3 |
size 37965684
|
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 sha256:7645d4bab96924146718c10368503cddfbaffa9b9752acf45c4c650d90ccb1e7
|
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:f39bff186481b2691af54be9cda77b9297711925f7bc2907a12df9c78be5954c
|
3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
@@ -1,9 +1,9 @@
|
|
1 |
{
|
2 |
"best_metric": 0.574374258518219,
|
3 |
"best_model_checkpoint": "miner_id_24/checkpoint-700",
|
4 |
-
"epoch": 1.
|
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,
|
@@ -5679,6 +5679,714 @@
|
|
5679 |
"eval_samples_per_second": 31.748,
|
5680 |
"eval_steps_per_second": 7.937,
|
5681 |
"step": 800
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
5682 |
}
|
5683 |
],
|
5684 |
"logging_steps": 1,
|
@@ -5693,7 +6401,7 @@
|
|
5693 |
"early_stopping_threshold": 0.0
|
5694 |
},
|
5695 |
"attributes": {
|
5696 |
-
"early_stopping_patience_counter":
|
5697 |
}
|
5698 |
},
|
5699 |
"TrainerControl": {
|
@@ -5702,12 +6410,12 @@
|
|
5702 |
"should_evaluate": false,
|
5703 |
"should_log": false,
|
5704 |
"should_save": true,
|
5705 |
-
"should_training_stop":
|
5706 |
},
|
5707 |
"attributes": {}
|
5708 |
}
|
5709 |
},
|
5710 |
-
"total_flos": 2.
|
5711 |
"train_batch_size": 4,
|
5712 |
"trial_name": null,
|
5713 |
"trial_params": null
|
|
|
1 |
{
|
2 |
"best_metric": 0.574374258518219,
|
3 |
"best_model_checkpoint": "miner_id_24/checkpoint-700",
|
4 |
+
"epoch": 1.230979654641819,
|
5 |
"eval_steps": 100,
|
6 |
+
"global_step": 900,
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
|
|
5679 |
"eval_samples_per_second": 31.748,
|
5680 |
"eval_steps_per_second": 7.937,
|
5681 |
"step": 800
|
5682 |
+
},
|
5683 |
+
{
|
5684 |
+
"epoch": 1.095571892631219,
|
5685 |
+
"grad_norm": 0.28456875681877136,
|
5686 |
+
"learning_rate": 8.609934310386318e-05,
|
5687 |
+
"loss": 0.6427,
|
5688 |
+
"step": 801
|
5689 |
+
},
|
5690 |
+
{
|
5691 |
+
"epoch": 1.0969396478030433,
|
5692 |
+
"grad_norm": 0.29876023530960083,
|
5693 |
+
"learning_rate": 8.588526066457907e-05,
|
5694 |
+
"loss": 0.5213,
|
5695 |
+
"step": 802
|
5696 |
+
},
|
5697 |
+
{
|
5698 |
+
"epoch": 1.0983074029748674,
|
5699 |
+
"grad_norm": 0.2823258936405182,
|
5700 |
+
"learning_rate": 8.567124420968549e-05,
|
5701 |
+
"loss": 0.6555,
|
5702 |
+
"step": 803
|
5703 |
+
},
|
5704 |
+
{
|
5705 |
+
"epoch": 1.0996751581466917,
|
5706 |
+
"grad_norm": 0.28708407282829285,
|
5707 |
+
"learning_rate": 8.545729473967884e-05,
|
5708 |
+
"loss": 0.618,
|
5709 |
+
"step": 804
|
5710 |
+
},
|
5711 |
+
{
|
5712 |
+
"epoch": 1.101042913318516,
|
5713 |
+
"grad_norm": 0.2833726108074188,
|
5714 |
+
"learning_rate": 8.52434132547423e-05,
|
5715 |
+
"loss": 0.6349,
|
5716 |
+
"step": 805
|
5717 |
+
},
|
5718 |
+
{
|
5719 |
+
"epoch": 1.1024106684903403,
|
5720 |
+
"grad_norm": 0.33605238795280457,
|
5721 |
+
"learning_rate": 8.502960075474131e-05,
|
5722 |
+
"loss": 0.4921,
|
5723 |
+
"step": 806
|
5724 |
+
},
|
5725 |
+
{
|
5726 |
+
"epoch": 1.1037784236621644,
|
5727 |
+
"grad_norm": 0.3451543152332306,
|
5728 |
+
"learning_rate": 8.481585823921873e-05,
|
5729 |
+
"loss": 0.5024,
|
5730 |
+
"step": 807
|
5731 |
+
},
|
5732 |
+
{
|
5733 |
+
"epoch": 1.1051461788339887,
|
5734 |
+
"grad_norm": 0.23923547565937042,
|
5735 |
+
"learning_rate": 8.460218670739031e-05,
|
5736 |
+
"loss": 0.3788,
|
5737 |
+
"step": 808
|
5738 |
+
},
|
5739 |
+
{
|
5740 |
+
"epoch": 1.106513934005813,
|
5741 |
+
"grad_norm": 0.2655201852321625,
|
5742 |
+
"learning_rate": 8.438858715813996e-05,
|
5743 |
+
"loss": 0.5876,
|
5744 |
+
"step": 809
|
5745 |
+
},
|
5746 |
+
{
|
5747 |
+
"epoch": 1.1078816891776373,
|
5748 |
+
"grad_norm": 0.2713606357574463,
|
5749 |
+
"learning_rate": 8.417506059001505e-05,
|
5750 |
+
"loss": 0.397,
|
5751 |
+
"step": 810
|
5752 |
+
},
|
5753 |
+
{
|
5754 |
+
"epoch": 1.1092494443494614,
|
5755 |
+
"grad_norm": 0.24983306229114532,
|
5756 |
+
"learning_rate": 8.396160800122176e-05,
|
5757 |
+
"loss": 0.3463,
|
5758 |
+
"step": 811
|
5759 |
+
},
|
5760 |
+
{
|
5761 |
+
"epoch": 1.1106171995212857,
|
5762 |
+
"grad_norm": 0.28522688150405884,
|
5763 |
+
"learning_rate": 8.374823038962056e-05,
|
5764 |
+
"loss": 0.4376,
|
5765 |
+
"step": 812
|
5766 |
+
},
|
5767 |
+
{
|
5768 |
+
"epoch": 1.11198495469311,
|
5769 |
+
"grad_norm": 0.2977089583873749,
|
5770 |
+
"learning_rate": 8.353492875272118e-05,
|
5771 |
+
"loss": 0.4729,
|
5772 |
+
"step": 813
|
5773 |
+
},
|
5774 |
+
{
|
5775 |
+
"epoch": 1.1133527098649343,
|
5776 |
+
"grad_norm": 0.2716682553291321,
|
5777 |
+
"learning_rate": 8.33217040876784e-05,
|
5778 |
+
"loss": 0.5256,
|
5779 |
+
"step": 814
|
5780 |
+
},
|
5781 |
+
{
|
5782 |
+
"epoch": 1.1147204650367584,
|
5783 |
+
"grad_norm": 0.34533944725990295,
|
5784 |
+
"learning_rate": 8.310855739128703e-05,
|
5785 |
+
"loss": 0.5092,
|
5786 |
+
"step": 815
|
5787 |
+
},
|
5788 |
+
{
|
5789 |
+
"epoch": 1.1160882202085827,
|
5790 |
+
"grad_norm": 0.2501341998577118,
|
5791 |
+
"learning_rate": 8.289548965997748e-05,
|
5792 |
+
"loss": 0.4022,
|
5793 |
+
"step": 816
|
5794 |
+
},
|
5795 |
+
{
|
5796 |
+
"epoch": 1.117455975380407,
|
5797 |
+
"grad_norm": 0.2617763578891754,
|
5798 |
+
"learning_rate": 8.26825018898109e-05,
|
5799 |
+
"loss": 0.3671,
|
5800 |
+
"step": 817
|
5801 |
+
},
|
5802 |
+
{
|
5803 |
+
"epoch": 1.118823730552231,
|
5804 |
+
"grad_norm": 0.2673083543777466,
|
5805 |
+
"learning_rate": 8.246959507647478e-05,
|
5806 |
+
"loss": 0.4135,
|
5807 |
+
"step": 818
|
5808 |
+
},
|
5809 |
+
{
|
5810 |
+
"epoch": 1.1201914857240554,
|
5811 |
+
"grad_norm": 0.31147336959838867,
|
5812 |
+
"learning_rate": 8.225677021527798e-05,
|
5813 |
+
"loss": 0.6971,
|
5814 |
+
"step": 819
|
5815 |
+
},
|
5816 |
+
{
|
5817 |
+
"epoch": 1.1215592408958797,
|
5818 |
+
"grad_norm": 0.23961347341537476,
|
5819 |
+
"learning_rate": 8.20440283011464e-05,
|
5820 |
+
"loss": 0.3846,
|
5821 |
+
"step": 820
|
5822 |
+
},
|
5823 |
+
{
|
5824 |
+
"epoch": 1.122926996067704,
|
5825 |
+
"grad_norm": 0.2794473469257355,
|
5826 |
+
"learning_rate": 8.183137032861804e-05,
|
5827 |
+
"loss": 0.3729,
|
5828 |
+
"step": 821
|
5829 |
+
},
|
5830 |
+
{
|
5831 |
+
"epoch": 1.124294751239528,
|
5832 |
+
"grad_norm": 0.32503047585487366,
|
5833 |
+
"learning_rate": 8.16187972918386e-05,
|
5834 |
+
"loss": 0.6592,
|
5835 |
+
"step": 822
|
5836 |
+
},
|
5837 |
+
{
|
5838 |
+
"epoch": 1.1256625064113523,
|
5839 |
+
"grad_norm": 0.3039405643939972,
|
5840 |
+
"learning_rate": 8.140631018455664e-05,
|
5841 |
+
"loss": 0.4657,
|
5842 |
+
"step": 823
|
5843 |
+
},
|
5844 |
+
{
|
5845 |
+
"epoch": 1.1270302615831767,
|
5846 |
+
"grad_norm": 0.36968639492988586,
|
5847 |
+
"learning_rate": 8.11939100001191e-05,
|
5848 |
+
"loss": 0.5455,
|
5849 |
+
"step": 824
|
5850 |
+
},
|
5851 |
+
{
|
5852 |
+
"epoch": 1.1283980167550007,
|
5853 |
+
"grad_norm": 0.27475252747535706,
|
5854 |
+
"learning_rate": 8.098159773146639e-05,
|
5855 |
+
"loss": 0.4267,
|
5856 |
+
"step": 825
|
5857 |
+
},
|
5858 |
+
{
|
5859 |
+
"epoch": 1.129765771926825,
|
5860 |
+
"grad_norm": 0.29142552614212036,
|
5861 |
+
"learning_rate": 8.07693743711282e-05,
|
5862 |
+
"loss": 0.3276,
|
5863 |
+
"step": 826
|
5864 |
+
},
|
5865 |
+
{
|
5866 |
+
"epoch": 1.1311335270986493,
|
5867 |
+
"grad_norm": 0.282996267080307,
|
5868 |
+
"learning_rate": 8.055724091121826e-05,
|
5869 |
+
"loss": 0.4624,
|
5870 |
+
"step": 827
|
5871 |
+
},
|
5872 |
+
{
|
5873 |
+
"epoch": 1.1325012822704736,
|
5874 |
+
"grad_norm": 0.24529734253883362,
|
5875 |
+
"learning_rate": 8.034519834343033e-05,
|
5876 |
+
"loss": 0.5003,
|
5877 |
+
"step": 828
|
5878 |
+
},
|
5879 |
+
{
|
5880 |
+
"epoch": 1.133869037442298,
|
5881 |
+
"grad_norm": 0.2863565683364868,
|
5882 |
+
"learning_rate": 8.013324765903305e-05,
|
5883 |
+
"loss": 0.552,
|
5884 |
+
"step": 829
|
5885 |
+
},
|
5886 |
+
{
|
5887 |
+
"epoch": 1.135236792614122,
|
5888 |
+
"grad_norm": 0.2725461423397064,
|
5889 |
+
"learning_rate": 7.992138984886567e-05,
|
5890 |
+
"loss": 0.5248,
|
5891 |
+
"step": 830
|
5892 |
+
},
|
5893 |
+
{
|
5894 |
+
"epoch": 1.1366045477859463,
|
5895 |
+
"grad_norm": 0.33715713024139404,
|
5896 |
+
"learning_rate": 7.970962590333316e-05,
|
5897 |
+
"loss": 0.6072,
|
5898 |
+
"step": 831
|
5899 |
+
},
|
5900 |
+
{
|
5901 |
+
"epoch": 1.1379723029577706,
|
5902 |
+
"grad_norm": 0.26902320981025696,
|
5903 |
+
"learning_rate": 7.949795681240168e-05,
|
5904 |
+
"loss": 0.4953,
|
5905 |
+
"step": 832
|
5906 |
+
},
|
5907 |
+
{
|
5908 |
+
"epoch": 1.1393400581295947,
|
5909 |
+
"grad_norm": 0.2683982849121094,
|
5910 |
+
"learning_rate": 7.928638356559407e-05,
|
5911 |
+
"loss": 0.6736,
|
5912 |
+
"step": 833
|
5913 |
+
},
|
5914 |
+
{
|
5915 |
+
"epoch": 1.140707813301419,
|
5916 |
+
"grad_norm": 0.32076001167297363,
|
5917 |
+
"learning_rate": 7.9074907151985e-05,
|
5918 |
+
"loss": 0.5829,
|
5919 |
+
"step": 834
|
5920 |
+
},
|
5921 |
+
{
|
5922 |
+
"epoch": 1.1420755684732433,
|
5923 |
+
"grad_norm": 0.29337945580482483,
|
5924 |
+
"learning_rate": 7.886352856019653e-05,
|
5925 |
+
"loss": 0.4851,
|
5926 |
+
"step": 835
|
5927 |
+
},
|
5928 |
+
{
|
5929 |
+
"epoch": 1.1434433236450676,
|
5930 |
+
"grad_norm": 0.24045072495937347,
|
5931 |
+
"learning_rate": 7.865224877839333e-05,
|
5932 |
+
"loss": 0.3578,
|
5933 |
+
"step": 836
|
5934 |
+
},
|
5935 |
+
{
|
5936 |
+
"epoch": 1.1448110788168917,
|
5937 |
+
"grad_norm": 0.257567822933197,
|
5938 |
+
"learning_rate": 7.844106879427829e-05,
|
5939 |
+
"loss": 0.4057,
|
5940 |
+
"step": 837
|
5941 |
+
},
|
5942 |
+
{
|
5943 |
+
"epoch": 1.146178833988716,
|
5944 |
+
"grad_norm": 0.2878836989402771,
|
5945 |
+
"learning_rate": 7.82299895950876e-05,
|
5946 |
+
"loss": 0.4812,
|
5947 |
+
"step": 838
|
5948 |
+
},
|
5949 |
+
{
|
5950 |
+
"epoch": 1.1475465891605403,
|
5951 |
+
"grad_norm": 0.3263901174068451,
|
5952 |
+
"learning_rate": 7.801901216758642e-05,
|
5953 |
+
"loss": 0.6451,
|
5954 |
+
"step": 839
|
5955 |
+
},
|
5956 |
+
{
|
5957 |
+
"epoch": 1.1489143443323644,
|
5958 |
+
"grad_norm": 0.25824493169784546,
|
5959 |
+
"learning_rate": 7.780813749806407e-05,
|
5960 |
+
"loss": 0.4484,
|
5961 |
+
"step": 840
|
5962 |
+
},
|
5963 |
+
{
|
5964 |
+
"epoch": 1.1502820995041887,
|
5965 |
+
"grad_norm": 0.25362029671669006,
|
5966 |
+
"learning_rate": 7.759736657232951e-05,
|
5967 |
+
"loss": 0.4383,
|
5968 |
+
"step": 841
|
5969 |
+
},
|
5970 |
+
{
|
5971 |
+
"epoch": 1.151649854676013,
|
5972 |
+
"grad_norm": 0.3550529479980469,
|
5973 |
+
"learning_rate": 7.73867003757067e-05,
|
5974 |
+
"loss": 0.5098,
|
5975 |
+
"step": 842
|
5976 |
+
},
|
5977 |
+
{
|
5978 |
+
"epoch": 1.1530176098478373,
|
5979 |
+
"grad_norm": 0.3048471510410309,
|
5980 |
+
"learning_rate": 7.717613989303011e-05,
|
5981 |
+
"loss": 0.5214,
|
5982 |
+
"step": 843
|
5983 |
+
},
|
5984 |
+
{
|
5985 |
+
"epoch": 1.1543853650196616,
|
5986 |
+
"grad_norm": 0.29699233174324036,
|
5987 |
+
"learning_rate": 7.696568610863978e-05,
|
5988 |
+
"loss": 0.4679,
|
5989 |
+
"step": 844
|
5990 |
+
},
|
5991 |
+
{
|
5992 |
+
"epoch": 1.1557531201914857,
|
5993 |
+
"grad_norm": 0.29977262020111084,
|
5994 |
+
"learning_rate": 7.675534000637724e-05,
|
5995 |
+
"loss": 0.4834,
|
5996 |
+
"step": 845
|
5997 |
+
},
|
5998 |
+
{
|
5999 |
+
"epoch": 1.15712087536331,
|
6000 |
+
"grad_norm": 0.3024733066558838,
|
6001 |
+
"learning_rate": 7.654510256958039e-05,
|
6002 |
+
"loss": 0.5326,
|
6003 |
+
"step": 846
|
6004 |
+
},
|
6005 |
+
{
|
6006 |
+
"epoch": 1.1584886305351343,
|
6007 |
+
"grad_norm": 0.3622778058052063,
|
6008 |
+
"learning_rate": 7.633497478107925e-05,
|
6009 |
+
"loss": 0.6334,
|
6010 |
+
"step": 847
|
6011 |
+
},
|
6012 |
+
{
|
6013 |
+
"epoch": 1.1598563857069584,
|
6014 |
+
"grad_norm": 0.3505115509033203,
|
6015 |
+
"learning_rate": 7.612495762319121e-05,
|
6016 |
+
"loss": 0.4456,
|
6017 |
+
"step": 848
|
6018 |
+
},
|
6019 |
+
{
|
6020 |
+
"epoch": 1.1612241408787827,
|
6021 |
+
"grad_norm": 0.30360838770866394,
|
6022 |
+
"learning_rate": 7.59150520777165e-05,
|
6023 |
+
"loss": 0.4395,
|
6024 |
+
"step": 849
|
6025 |
+
},
|
6026 |
+
{
|
6027 |
+
"epoch": 1.162591896050607,
|
6028 |
+
"grad_norm": 0.2924692630767822,
|
6029 |
+
"learning_rate": 7.570525912593354e-05,
|
6030 |
+
"loss": 0.4935,
|
6031 |
+
"step": 850
|
6032 |
+
},
|
6033 |
+
{
|
6034 |
+
"epoch": 1.1639596512224313,
|
6035 |
+
"grad_norm": 0.30422306060791016,
|
6036 |
+
"learning_rate": 7.549557974859446e-05,
|
6037 |
+
"loss": 0.5371,
|
6038 |
+
"step": 851
|
6039 |
+
},
|
6040 |
+
{
|
6041 |
+
"epoch": 1.1653274063942554,
|
6042 |
+
"grad_norm": 0.3041650354862213,
|
6043 |
+
"learning_rate": 7.528601492592037e-05,
|
6044 |
+
"loss": 0.4409,
|
6045 |
+
"step": 852
|
6046 |
+
},
|
6047 |
+
{
|
6048 |
+
"epoch": 1.1666951615660797,
|
6049 |
+
"grad_norm": 0.29774245619773865,
|
6050 |
+
"learning_rate": 7.507656563759691e-05,
|
6051 |
+
"loss": 0.5018,
|
6052 |
+
"step": 853
|
6053 |
+
},
|
6054 |
+
{
|
6055 |
+
"epoch": 1.168062916737904,
|
6056 |
+
"grad_norm": 0.3011612892150879,
|
6057 |
+
"learning_rate": 7.486723286276953e-05,
|
6058 |
+
"loss": 0.4171,
|
6059 |
+
"step": 854
|
6060 |
+
},
|
6061 |
+
{
|
6062 |
+
"epoch": 1.169430671909728,
|
6063 |
+
"grad_norm": 0.2770425081253052,
|
6064 |
+
"learning_rate": 7.465801758003918e-05,
|
6065 |
+
"loss": 0.3773,
|
6066 |
+
"step": 855
|
6067 |
+
},
|
6068 |
+
{
|
6069 |
+
"epoch": 1.1707984270815524,
|
6070 |
+
"grad_norm": 0.2873842716217041,
|
6071 |
+
"learning_rate": 7.444892076745728e-05,
|
6072 |
+
"loss": 0.5557,
|
6073 |
+
"step": 856
|
6074 |
+
},
|
6075 |
+
{
|
6076 |
+
"epoch": 1.1721661822533767,
|
6077 |
+
"grad_norm": 0.2865909934043884,
|
6078 |
+
"learning_rate": 7.423994340252169e-05,
|
6079 |
+
"loss": 0.39,
|
6080 |
+
"step": 857
|
6081 |
+
},
|
6082 |
+
{
|
6083 |
+
"epoch": 1.173533937425201,
|
6084 |
+
"grad_norm": 0.2772740423679352,
|
6085 |
+
"learning_rate": 7.403108646217165e-05,
|
6086 |
+
"loss": 0.5739,
|
6087 |
+
"step": 858
|
6088 |
+
},
|
6089 |
+
{
|
6090 |
+
"epoch": 1.174901692597025,
|
6091 |
+
"grad_norm": 0.3060879409313202,
|
6092 |
+
"learning_rate": 7.382235092278359e-05,
|
6093 |
+
"loss": 0.5719,
|
6094 |
+
"step": 859
|
6095 |
+
},
|
6096 |
+
{
|
6097 |
+
"epoch": 1.1762694477688493,
|
6098 |
+
"grad_norm": 0.33835381269454956,
|
6099 |
+
"learning_rate": 7.361373776016631e-05,
|
6100 |
+
"loss": 0.536,
|
6101 |
+
"step": 860
|
6102 |
+
},
|
6103 |
+
{
|
6104 |
+
"epoch": 1.1776372029406736,
|
6105 |
+
"grad_norm": 0.28969088196754456,
|
6106 |
+
"learning_rate": 7.340524794955659e-05,
|
6107 |
+
"loss": 0.4098,
|
6108 |
+
"step": 861
|
6109 |
+
},
|
6110 |
+
{
|
6111 |
+
"epoch": 1.179004958112498,
|
6112 |
+
"grad_norm": 0.29033181071281433,
|
6113 |
+
"learning_rate": 7.319688246561447e-05,
|
6114 |
+
"loss": 0.5777,
|
6115 |
+
"step": 862
|
6116 |
+
},
|
6117 |
+
{
|
6118 |
+
"epoch": 1.180372713284322,
|
6119 |
+
"grad_norm": 0.3036806583404541,
|
6120 |
+
"learning_rate": 7.298864228241887e-05,
|
6121 |
+
"loss": 0.4617,
|
6122 |
+
"step": 863
|
6123 |
+
},
|
6124 |
+
{
|
6125 |
+
"epoch": 1.1817404684561463,
|
6126 |
+
"grad_norm": 0.313621461391449,
|
6127 |
+
"learning_rate": 7.278052837346288e-05,
|
6128 |
+
"loss": 0.6076,
|
6129 |
+
"step": 864
|
6130 |
+
},
|
6131 |
+
{
|
6132 |
+
"epoch": 1.1831082236279706,
|
6133 |
+
"grad_norm": 0.27229294180870056,
|
6134 |
+
"learning_rate": 7.257254171164932e-05,
|
6135 |
+
"loss": 0.4272,
|
6136 |
+
"step": 865
|
6137 |
+
},
|
6138 |
+
{
|
6139 |
+
"epoch": 1.184475978799795,
|
6140 |
+
"grad_norm": 0.308609277009964,
|
6141 |
+
"learning_rate": 7.236468326928612e-05,
|
6142 |
+
"loss": 0.5481,
|
6143 |
+
"step": 866
|
6144 |
+
},
|
6145 |
+
{
|
6146 |
+
"epoch": 1.185843733971619,
|
6147 |
+
"grad_norm": 0.29016566276550293,
|
6148 |
+
"learning_rate": 7.21569540180818e-05,
|
6149 |
+
"loss": 0.4163,
|
6150 |
+
"step": 867
|
6151 |
+
},
|
6152 |
+
{
|
6153 |
+
"epoch": 1.1872114891434433,
|
6154 |
+
"grad_norm": 0.33708828687667847,
|
6155 |
+
"learning_rate": 7.1949354929141e-05,
|
6156 |
+
"loss": 0.5613,
|
6157 |
+
"step": 868
|
6158 |
+
},
|
6159 |
+
{
|
6160 |
+
"epoch": 1.1885792443152676,
|
6161 |
+
"grad_norm": 0.32777532935142517,
|
6162 |
+
"learning_rate": 7.174188697295971e-05,
|
6163 |
+
"loss": 0.4931,
|
6164 |
+
"step": 869
|
6165 |
+
},
|
6166 |
+
{
|
6167 |
+
"epoch": 1.1899469994870917,
|
6168 |
+
"grad_norm": 0.2704114615917206,
|
6169 |
+
"learning_rate": 7.153455111942108e-05,
|
6170 |
+
"loss": 0.4251,
|
6171 |
+
"step": 870
|
6172 |
+
},
|
6173 |
+
{
|
6174 |
+
"epoch": 1.191314754658916,
|
6175 |
+
"grad_norm": 0.3016801178455353,
|
6176 |
+
"learning_rate": 7.132734833779054e-05,
|
6177 |
+
"loss": 0.4761,
|
6178 |
+
"step": 871
|
6179 |
+
},
|
6180 |
+
{
|
6181 |
+
"epoch": 1.1926825098307403,
|
6182 |
+
"grad_norm": 0.31739842891693115,
|
6183 |
+
"learning_rate": 7.112027959671162e-05,
|
6184 |
+
"loss": 0.5516,
|
6185 |
+
"step": 872
|
6186 |
+
},
|
6187 |
+
{
|
6188 |
+
"epoch": 1.1940502650025646,
|
6189 |
+
"grad_norm": 0.3556269705295563,
|
6190 |
+
"learning_rate": 7.091334586420097e-05,
|
6191 |
+
"loss": 0.5474,
|
6192 |
+
"step": 873
|
6193 |
+
},
|
6194 |
+
{
|
6195 |
+
"epoch": 1.1954180201743887,
|
6196 |
+
"grad_norm": 0.31290698051452637,
|
6197 |
+
"learning_rate": 7.070654810764438e-05,
|
6198 |
+
"loss": 0.5217,
|
6199 |
+
"step": 874
|
6200 |
+
},
|
6201 |
+
{
|
6202 |
+
"epoch": 1.196785775346213,
|
6203 |
+
"grad_norm": 0.2770576775074005,
|
6204 |
+
"learning_rate": 7.049988729379168e-05,
|
6205 |
+
"loss": 0.4103,
|
6206 |
+
"step": 875
|
6207 |
+
},
|
6208 |
+
{
|
6209 |
+
"epoch": 1.1981535305180373,
|
6210 |
+
"grad_norm": 0.2821738123893738,
|
6211 |
+
"learning_rate": 7.029336438875279e-05,
|
6212 |
+
"loss": 0.5438,
|
6213 |
+
"step": 876
|
6214 |
+
},
|
6215 |
+
{
|
6216 |
+
"epoch": 1.1995212856898616,
|
6217 |
+
"grad_norm": 0.3358316123485565,
|
6218 |
+
"learning_rate": 7.008698035799268e-05,
|
6219 |
+
"loss": 0.8765,
|
6220 |
+
"step": 877
|
6221 |
+
},
|
6222 |
+
{
|
6223 |
+
"epoch": 1.2008890408616857,
|
6224 |
+
"grad_norm": 0.31030896306037903,
|
6225 |
+
"learning_rate": 6.988073616632732e-05,
|
6226 |
+
"loss": 0.5541,
|
6227 |
+
"step": 878
|
6228 |
+
},
|
6229 |
+
{
|
6230 |
+
"epoch": 1.20225679603351,
|
6231 |
+
"grad_norm": 0.30155354738235474,
|
6232 |
+
"learning_rate": 6.967463277791879e-05,
|
6233 |
+
"loss": 0.4087,
|
6234 |
+
"step": 879
|
6235 |
+
},
|
6236 |
+
{
|
6237 |
+
"epoch": 1.2036245512053343,
|
6238 |
+
"grad_norm": 0.2397313416004181,
|
6239 |
+
"learning_rate": 6.946867115627103e-05,
|
6240 |
+
"loss": 0.4809,
|
6241 |
+
"step": 880
|
6242 |
+
},
|
6243 |
+
{
|
6244 |
+
"epoch": 1.2049923063771586,
|
6245 |
+
"grad_norm": 0.2920851707458496,
|
6246 |
+
"learning_rate": 6.926285226422515e-05,
|
6247 |
+
"loss": 0.5256,
|
6248 |
+
"step": 881
|
6249 |
+
},
|
6250 |
+
{
|
6251 |
+
"epoch": 1.2063600615489827,
|
6252 |
+
"grad_norm": 0.2639416754245758,
|
6253 |
+
"learning_rate": 6.905717706395516e-05,
|
6254 |
+
"loss": 0.4735,
|
6255 |
+
"step": 882
|
6256 |
+
},
|
6257 |
+
{
|
6258 |
+
"epoch": 1.207727816720807,
|
6259 |
+
"grad_norm": 0.32958149909973145,
|
6260 |
+
"learning_rate": 6.885164651696317e-05,
|
6261 |
+
"loss": 0.5341,
|
6262 |
+
"step": 883
|
6263 |
+
},
|
6264 |
+
{
|
6265 |
+
"epoch": 1.2090955718926313,
|
6266 |
+
"grad_norm": 0.2795344591140747,
|
6267 |
+
"learning_rate": 6.86462615840752e-05,
|
6268 |
+
"loss": 0.4619,
|
6269 |
+
"step": 884
|
6270 |
+
},
|
6271 |
+
{
|
6272 |
+
"epoch": 1.2104633270644554,
|
6273 |
+
"grad_norm": 0.2936275005340576,
|
6274 |
+
"learning_rate": 6.844102322543638e-05,
|
6275 |
+
"loss": 0.4208,
|
6276 |
+
"step": 885
|
6277 |
+
},
|
6278 |
+
{
|
6279 |
+
"epoch": 1.2118310822362797,
|
6280 |
+
"grad_norm": 0.4172208905220032,
|
6281 |
+
"learning_rate": 6.823593240050685e-05,
|
6282 |
+
"loss": 0.6143,
|
6283 |
+
"step": 886
|
6284 |
+
},
|
6285 |
+
{
|
6286 |
+
"epoch": 1.213198837408104,
|
6287 |
+
"grad_norm": 0.2924855351448059,
|
6288 |
+
"learning_rate": 6.80309900680568e-05,
|
6289 |
+
"loss": 0.403,
|
6290 |
+
"step": 887
|
6291 |
+
},
|
6292 |
+
{
|
6293 |
+
"epoch": 1.2145665925799283,
|
6294 |
+
"grad_norm": 0.30166590213775635,
|
6295 |
+
"learning_rate": 6.782619718616252e-05,
|
6296 |
+
"loss": 0.7077,
|
6297 |
+
"step": 888
|
6298 |
+
},
|
6299 |
+
{
|
6300 |
+
"epoch": 1.2159343477517524,
|
6301 |
+
"grad_norm": 0.26507216691970825,
|
6302 |
+
"learning_rate": 6.762155471220131e-05,
|
6303 |
+
"loss": 0.4196,
|
6304 |
+
"step": 889
|
6305 |
+
},
|
6306 |
+
{
|
6307 |
+
"epoch": 1.2173021029235767,
|
6308 |
+
"grad_norm": 0.2820349931716919,
|
6309 |
+
"learning_rate": 6.741706360284763e-05,
|
6310 |
+
"loss": 0.5192,
|
6311 |
+
"step": 890
|
6312 |
+
},
|
6313 |
+
{
|
6314 |
+
"epoch": 1.218669858095401,
|
6315 |
+
"grad_norm": 0.2856913208961487,
|
6316 |
+
"learning_rate": 6.721272481406817e-05,
|
6317 |
+
"loss": 0.5337,
|
6318 |
+
"step": 891
|
6319 |
+
},
|
6320 |
+
{
|
6321 |
+
"epoch": 1.2200376132672253,
|
6322 |
+
"grad_norm": 0.3205767273902893,
|
6323 |
+
"learning_rate": 6.70085393011176e-05,
|
6324 |
+
"loss": 0.3494,
|
6325 |
+
"step": 892
|
6326 |
+
},
|
6327 |
+
{
|
6328 |
+
"epoch": 1.2214053684390493,
|
6329 |
+
"grad_norm": 0.3018949031829834,
|
6330 |
+
"learning_rate": 6.6804508018534e-05,
|
6331 |
+
"loss": 0.35,
|
6332 |
+
"step": 893
|
6333 |
+
},
|
6334 |
+
{
|
6335 |
+
"epoch": 1.2227731236108736,
|
6336 |
+
"grad_norm": 0.3356417119503021,
|
6337 |
+
"learning_rate": 6.660063192013451e-05,
|
6338 |
+
"loss": 0.4085,
|
6339 |
+
"step": 894
|
6340 |
+
},
|
6341 |
+
{
|
6342 |
+
"epoch": 1.224140878782698,
|
6343 |
+
"grad_norm": 0.31260353326797485,
|
6344 |
+
"learning_rate": 6.639691195901076e-05,
|
6345 |
+
"loss": 0.3837,
|
6346 |
+
"step": 895
|
6347 |
+
},
|
6348 |
+
{
|
6349 |
+
"epoch": 1.2255086339545223,
|
6350 |
+
"grad_norm": 0.2612607181072235,
|
6351 |
+
"learning_rate": 6.619334908752447e-05,
|
6352 |
+
"loss": 0.4347,
|
6353 |
+
"step": 896
|
6354 |
+
},
|
6355 |
+
{
|
6356 |
+
"epoch": 1.2268763891263463,
|
6357 |
+
"grad_norm": 0.26736879348754883,
|
6358 |
+
"learning_rate": 6.598994425730299e-05,
|
6359 |
+
"loss": 0.537,
|
6360 |
+
"step": 897
|
6361 |
+
},
|
6362 |
+
{
|
6363 |
+
"epoch": 1.2282441442981706,
|
6364 |
+
"grad_norm": 0.2847106158733368,
|
6365 |
+
"learning_rate": 6.578669841923486e-05,
|
6366 |
+
"loss": 0.5361,
|
6367 |
+
"step": 898
|
6368 |
+
},
|
6369 |
+
{
|
6370 |
+
"epoch": 1.229611899469995,
|
6371 |
+
"grad_norm": 0.254794716835022,
|
6372 |
+
"learning_rate": 6.558361252346532e-05,
|
6373 |
+
"loss": 0.454,
|
6374 |
+
"step": 899
|
6375 |
+
},
|
6376 |
+
{
|
6377 |
+
"epoch": 1.230979654641819,
|
6378 |
+
"grad_norm": 0.27473878860473633,
|
6379 |
+
"learning_rate": 6.5380687519392e-05,
|
6380 |
+
"loss": 0.5392,
|
6381 |
+
"step": 900
|
6382 |
+
},
|
6383 |
+
{
|
6384 |
+
"epoch": 1.230979654641819,
|
6385 |
+
"eval_loss": 0.5753397941589355,
|
6386 |
+
"eval_runtime": 38.7964,
|
6387 |
+
"eval_samples_per_second": 31.756,
|
6388 |
+
"eval_steps_per_second": 7.939,
|
6389 |
+
"step": 900
|
6390 |
}
|
6391 |
],
|
6392 |
"logging_steps": 1,
|
|
|
6401 |
"early_stopping_threshold": 0.0
|
6402 |
},
|
6403 |
"attributes": {
|
6404 |
+
"early_stopping_patience_counter": 2
|
6405 |
}
|
6406 |
},
|
6407 |
"TrainerControl": {
|
|
|
6410 |
"should_evaluate": false,
|
6411 |
"should_log": false,
|
6412 |
"should_save": true,
|
6413 |
+
"should_training_stop": true
|
6414 |
},
|
6415 |
"attributes": {}
|
6416 |
}
|
6417 |
},
|
6418 |
+
"total_flos": 2.57057378726314e+17,
|
6419 |
"train_batch_size": 4,
|
6420 |
"trial_name": null,
|
6421 |
"trial_params": null
|