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 159967880
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:022acefac6286471ee8d2c71c3690fd8a565a014ef058c4e3af4f78951dbd0c6
|
| 3 |
size 159967880
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 81730644
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6815cc9faa2ed01fbe0f7494c50e2dad1e8d76d810e7a58f8875d331c5a18757
|
| 3 |
size 81730644
|
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:a51e93265e2c6a6f43a552e36ccd1c9901c5296dbb2ff8ac30a43fe5205e60a5
|
| 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 sha256:4f7ccd40217515b1adb168de499560d4c6c803e2bb562e36a2dfe417244db986
|
| 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,
|
|
@@ -5679,6 +5679,714 @@
|
|
| 5679 |
"eval_samples_per_second": 7.139,
|
| 5680 |
"eval_steps_per_second": 1.785,
|
| 5681 |
"step": 800
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 5682 |
}
|
| 5683 |
],
|
| 5684 |
"logging_steps": 1,
|
|
@@ -5707,7 +6415,7 @@
|
|
| 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": 0.723136305809021,
|
| 3 |
+
"best_model_checkpoint": "miner_id_24/checkpoint-900",
|
| 4 |
+
"epoch": 0.06268500783562599,
|
| 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": 7.139,
|
| 5680 |
"eval_steps_per_second": 1.785,
|
| 5681 |
"step": 800
|
| 5682 |
+
},
|
| 5683 |
+
{
|
| 5684 |
+
"epoch": 0.05578965697370712,
|
| 5685 |
+
"grad_norm": 0.8243626356124878,
|
| 5686 |
+
"learning_rate": 1.4343506570148846e-05,
|
| 5687 |
+
"loss": 0.9067,
|
| 5688 |
+
"step": 801
|
| 5689 |
+
},
|
| 5690 |
+
{
|
| 5691 |
+
"epoch": 0.05585930698241337,
|
| 5692 |
+
"grad_norm": 0.740206241607666,
|
| 5693 |
+
"learning_rate": 1.4174389342653971e-05,
|
| 5694 |
+
"loss": 1.0956,
|
| 5695 |
+
"step": 802
|
| 5696 |
+
},
|
| 5697 |
+
{
|
| 5698 |
+
"epoch": 0.055928956991119626,
|
| 5699 |
+
"grad_norm": 0.6383155584335327,
|
| 5700 |
+
"learning_rate": 1.4006198945990168e-05,
|
| 5701 |
+
"loss": 0.9274,
|
| 5702 |
+
"step": 803
|
| 5703 |
+
},
|
| 5704 |
+
{
|
| 5705 |
+
"epoch": 0.055998606999825874,
|
| 5706 |
+
"grad_norm": 0.7425148487091064,
|
| 5707 |
+
"learning_rate": 1.3838937196445579e-05,
|
| 5708 |
+
"loss": 1.083,
|
| 5709 |
+
"step": 804
|
| 5710 |
+
},
|
| 5711 |
+
{
|
| 5712 |
+
"epoch": 0.05606825700853212,
|
| 5713 |
+
"grad_norm": 0.6034273505210876,
|
| 5714 |
+
"learning_rate": 1.367260590028e-05,
|
| 5715 |
+
"loss": 0.7125,
|
| 5716 |
+
"step": 805
|
| 5717 |
+
},
|
| 5718 |
+
{
|
| 5719 |
+
"epoch": 0.05613790701723838,
|
| 5720 |
+
"grad_norm": 0.7047588229179382,
|
| 5721 |
+
"learning_rate": 1.3507206853705178e-05,
|
| 5722 |
+
"loss": 0.7749,
|
| 5723 |
+
"step": 806
|
| 5724 |
+
},
|
| 5725 |
+
{
|
| 5726 |
+
"epoch": 0.05620755702594463,
|
| 5727 |
+
"grad_norm": 0.7387014627456665,
|
| 5728 |
+
"learning_rate": 1.334274184286558e-05,
|
| 5729 |
+
"loss": 0.7397,
|
| 5730 |
+
"step": 807
|
| 5731 |
+
},
|
| 5732 |
+
{
|
| 5733 |
+
"epoch": 0.05627720703465088,
|
| 5734 |
+
"grad_norm": 0.6060226559638977,
|
| 5735 |
+
"learning_rate": 1.3179212643818929e-05,
|
| 5736 |
+
"loss": 0.5144,
|
| 5737 |
+
"step": 808
|
| 5738 |
+
},
|
| 5739 |
+
{
|
| 5740 |
+
"epoch": 0.05634685704335713,
|
| 5741 |
+
"grad_norm": 0.7422417402267456,
|
| 5742 |
+
"learning_rate": 1.3016621022517206e-05,
|
| 5743 |
+
"loss": 0.7739,
|
| 5744 |
+
"step": 809
|
| 5745 |
+
},
|
| 5746 |
+
{
|
| 5747 |
+
"epoch": 0.05641650705206338,
|
| 5748 |
+
"grad_norm": 0.6336711645126343,
|
| 5749 |
+
"learning_rate": 1.2854968734787398e-05,
|
| 5750 |
+
"loss": 0.471,
|
| 5751 |
+
"step": 810
|
| 5752 |
+
},
|
| 5753 |
+
{
|
| 5754 |
+
"epoch": 0.056486157060769634,
|
| 5755 |
+
"grad_norm": 0.667668879032135,
|
| 5756 |
+
"learning_rate": 1.2694257526312725e-05,
|
| 5757 |
+
"loss": 0.4143,
|
| 5758 |
+
"step": 811
|
| 5759 |
+
},
|
| 5760 |
+
{
|
| 5761 |
+
"epoch": 0.05655580706947588,
|
| 5762 |
+
"grad_norm": 0.6936927437782288,
|
| 5763 |
+
"learning_rate": 1.2534489132613603e-05,
|
| 5764 |
+
"loss": 0.8842,
|
| 5765 |
+
"step": 812
|
| 5766 |
+
},
|
| 5767 |
+
{
|
| 5768 |
+
"epoch": 0.05662545707818213,
|
| 5769 |
+
"grad_norm": 0.6019664406776428,
|
| 5770 |
+
"learning_rate": 1.2375665279029048e-05,
|
| 5771 |
+
"loss": 0.7445,
|
| 5772 |
+
"step": 813
|
| 5773 |
+
},
|
| 5774 |
+
{
|
| 5775 |
+
"epoch": 0.05669510708688839,
|
| 5776 |
+
"grad_norm": 0.7595625519752502,
|
| 5777 |
+
"learning_rate": 1.221778768069799e-05,
|
| 5778 |
+
"loss": 0.8676,
|
| 5779 |
+
"step": 814
|
| 5780 |
+
},
|
| 5781 |
+
{
|
| 5782 |
+
"epoch": 0.056764757095594635,
|
| 5783 |
+
"grad_norm": 0.593315839767456,
|
| 5784 |
+
"learning_rate": 1.206085804254069e-05,
|
| 5785 |
+
"loss": 0.7546,
|
| 5786 |
+
"step": 815
|
| 5787 |
+
},
|
| 5788 |
+
{
|
| 5789 |
+
"epoch": 0.05683440710430089,
|
| 5790 |
+
"grad_norm": 0.7907949090003967,
|
| 5791 |
+
"learning_rate": 1.1904878059240442e-05,
|
| 5792 |
+
"loss": 1.0131,
|
| 5793 |
+
"step": 816
|
| 5794 |
+
},
|
| 5795 |
+
{
|
| 5796 |
+
"epoch": 0.05690405711300714,
|
| 5797 |
+
"grad_norm": 0.6472040414810181,
|
| 5798 |
+
"learning_rate": 1.174984941522519e-05,
|
| 5799 |
+
"loss": 0.6795,
|
| 5800 |
+
"step": 817
|
| 5801 |
+
},
|
| 5802 |
+
{
|
| 5803 |
+
"epoch": 0.05697370712171339,
|
| 5804 |
+
"grad_norm": 0.6748494505882263,
|
| 5805 |
+
"learning_rate": 1.1595773784649389e-05,
|
| 5806 |
+
"loss": 0.9777,
|
| 5807 |
+
"step": 818
|
| 5808 |
+
},
|
| 5809 |
+
{
|
| 5810 |
+
"epoch": 0.05704335713041964,
|
| 5811 |
+
"grad_norm": 0.7594382166862488,
|
| 5812 |
+
"learning_rate": 1.1442652831375855e-05,
|
| 5813 |
+
"loss": 0.8305,
|
| 5814 |
+
"step": 819
|
| 5815 |
+
},
|
| 5816 |
+
{
|
| 5817 |
+
"epoch": 0.05711300713912589,
|
| 5818 |
+
"grad_norm": 0.5605437159538269,
|
| 5819 |
+
"learning_rate": 1.1290488208957895e-05,
|
| 5820 |
+
"loss": 0.9774,
|
| 5821 |
+
"step": 820
|
| 5822 |
+
},
|
| 5823 |
+
{
|
| 5824 |
+
"epoch": 0.05718265714783214,
|
| 5825 |
+
"grad_norm": 0.7108663320541382,
|
| 5826 |
+
"learning_rate": 1.1139281560621362e-05,
|
| 5827 |
+
"loss": 1.1447,
|
| 5828 |
+
"step": 821
|
| 5829 |
+
},
|
| 5830 |
+
{
|
| 5831 |
+
"epoch": 0.057252307156538396,
|
| 5832 |
+
"grad_norm": 0.7549561858177185,
|
| 5833 |
+
"learning_rate": 1.0989034519246954e-05,
|
| 5834 |
+
"loss": 1.0838,
|
| 5835 |
+
"step": 822
|
| 5836 |
+
},
|
| 5837 |
+
{
|
| 5838 |
+
"epoch": 0.057321957165244644,
|
| 5839 |
+
"grad_norm": 0.5975289940834045,
|
| 5840 |
+
"learning_rate": 1.0839748707352603e-05,
|
| 5841 |
+
"loss": 1.0126,
|
| 5842 |
+
"step": 823
|
| 5843 |
+
},
|
| 5844 |
+
{
|
| 5845 |
+
"epoch": 0.0573916071739509,
|
| 5846 |
+
"grad_norm": 0.6680442094802856,
|
| 5847 |
+
"learning_rate": 1.06914257370759e-05,
|
| 5848 |
+
"loss": 0.5809,
|
| 5849 |
+
"step": 824
|
| 5850 |
+
},
|
| 5851 |
+
{
|
| 5852 |
+
"epoch": 0.05746125718265715,
|
| 5853 |
+
"grad_norm": 0.7288407683372498,
|
| 5854 |
+
"learning_rate": 1.0544067210156671e-05,
|
| 5855 |
+
"loss": 0.9369,
|
| 5856 |
+
"step": 825
|
| 5857 |
+
},
|
| 5858 |
+
{
|
| 5859 |
+
"epoch": 0.0575309071913634,
|
| 5860 |
+
"grad_norm": 0.7064124345779419,
|
| 5861 |
+
"learning_rate": 1.0397674717919802e-05,
|
| 5862 |
+
"loss": 0.8142,
|
| 5863 |
+
"step": 826
|
| 5864 |
+
},
|
| 5865 |
+
{
|
| 5866 |
+
"epoch": 0.05760055720006965,
|
| 5867 |
+
"grad_norm": 0.6422365307807922,
|
| 5868 |
+
"learning_rate": 1.0252249841257877e-05,
|
| 5869 |
+
"loss": 0.5993,
|
| 5870 |
+
"step": 827
|
| 5871 |
+
},
|
| 5872 |
+
{
|
| 5873 |
+
"epoch": 0.0576702072087759,
|
| 5874 |
+
"grad_norm": 0.6080381870269775,
|
| 5875 |
+
"learning_rate": 1.0107794150614281e-05,
|
| 5876 |
+
"loss": 0.6939,
|
| 5877 |
+
"step": 828
|
| 5878 |
+
},
|
| 5879 |
+
{
|
| 5880 |
+
"epoch": 0.05773985721748215,
|
| 5881 |
+
"grad_norm": 0.6256659030914307,
|
| 5882 |
+
"learning_rate": 9.964309205966083e-06,
|
| 5883 |
+
"loss": 0.4506,
|
| 5884 |
+
"step": 829
|
| 5885 |
+
},
|
| 5886 |
+
{
|
| 5887 |
+
"epoch": 0.057809507226188404,
|
| 5888 |
+
"grad_norm": 0.6198416352272034,
|
| 5889 |
+
"learning_rate": 9.821796556807339e-06,
|
| 5890 |
+
"loss": 0.6324,
|
| 5891 |
+
"step": 830
|
| 5892 |
+
},
|
| 5893 |
+
{
|
| 5894 |
+
"epoch": 0.05787915723489465,
|
| 5895 |
+
"grad_norm": 0.6347202658653259,
|
| 5896 |
+
"learning_rate": 9.680257742132215e-06,
|
| 5897 |
+
"loss": 0.6047,
|
| 5898 |
+
"step": 831
|
| 5899 |
+
},
|
| 5900 |
+
{
|
| 5901 |
+
"epoch": 0.05794880724360091,
|
| 5902 |
+
"grad_norm": 0.60918128490448,
|
| 5903 |
+
"learning_rate": 9.539694290418488e-06,
|
| 5904 |
+
"loss": 0.9085,
|
| 5905 |
+
"step": 832
|
| 5906 |
+
},
|
| 5907 |
+
{
|
| 5908 |
+
"epoch": 0.05801845725230716,
|
| 5909 |
+
"grad_norm": 0.6706361174583435,
|
| 5910 |
+
"learning_rate": 9.400107719610995e-06,
|
| 5911 |
+
"loss": 0.9078,
|
| 5912 |
+
"step": 833
|
| 5913 |
+
},
|
| 5914 |
+
{
|
| 5915 |
+
"epoch": 0.058088107261013405,
|
| 5916 |
+
"grad_norm": 0.7337279915809631,
|
| 5917 |
+
"learning_rate": 9.261499537105177e-06,
|
| 5918 |
+
"loss": 1.0197,
|
| 5919 |
+
"step": 834
|
| 5920 |
+
},
|
| 5921 |
+
{
|
| 5922 |
+
"epoch": 0.05815775726971966,
|
| 5923 |
+
"grad_norm": 0.5747254490852356,
|
| 5924 |
+
"learning_rate": 9.12387123973093e-06,
|
| 5925 |
+
"loss": 0.8288,
|
| 5926 |
+
"step": 835
|
| 5927 |
+
},
|
| 5928 |
+
{
|
| 5929 |
+
"epoch": 0.05822740727842591,
|
| 5930 |
+
"grad_norm": 0.6484262347221375,
|
| 5931 |
+
"learning_rate": 8.98722431373631e-06,
|
| 5932 |
+
"loss": 1.1276,
|
| 5933 |
+
"step": 836
|
| 5934 |
+
},
|
| 5935 |
+
{
|
| 5936 |
+
"epoch": 0.05829705728713216,
|
| 5937 |
+
"grad_norm": 0.6793870329856873,
|
| 5938 |
+
"learning_rate": 8.851560234771594e-06,
|
| 5939 |
+
"loss": 0.5941,
|
| 5940 |
+
"step": 837
|
| 5941 |
+
},
|
| 5942 |
+
{
|
| 5943 |
+
"epoch": 0.05836670729583841,
|
| 5944 |
+
"grad_norm": 0.6910689473152161,
|
| 5945 |
+
"learning_rate": 8.716880467873234e-06,
|
| 5946 |
+
"loss": 0.9097,
|
| 5947 |
+
"step": 838
|
| 5948 |
+
},
|
| 5949 |
+
{
|
| 5950 |
+
"epoch": 0.05843635730454466,
|
| 5951 |
+
"grad_norm": 0.7062430381774902,
|
| 5952 |
+
"learning_rate": 8.583186467448167e-06,
|
| 5953 |
+
"loss": 0.9619,
|
| 5954 |
+
"step": 839
|
| 5955 |
+
},
|
| 5956 |
+
{
|
| 5957 |
+
"epoch": 0.05850600731325092,
|
| 5958 |
+
"grad_norm": 0.8270265460014343,
|
| 5959 |
+
"learning_rate": 8.45047967725796e-06,
|
| 5960 |
+
"loss": 1.0196,
|
| 5961 |
+
"step": 840
|
| 5962 |
+
},
|
| 5963 |
+
{
|
| 5964 |
+
"epoch": 0.058575657321957166,
|
| 5965 |
+
"grad_norm": 0.6949748992919922,
|
| 5966 |
+
"learning_rate": 8.318761530403374e-06,
|
| 5967 |
+
"loss": 0.5329,
|
| 5968 |
+
"step": 841
|
| 5969 |
+
},
|
| 5970 |
+
{
|
| 5971 |
+
"epoch": 0.058645307330663414,
|
| 5972 |
+
"grad_norm": 0.7285637855529785,
|
| 5973 |
+
"learning_rate": 8.188033449308719e-06,
|
| 5974 |
+
"loss": 0.6849,
|
| 5975 |
+
"step": 842
|
| 5976 |
+
},
|
| 5977 |
+
{
|
| 5978 |
+
"epoch": 0.05871495733936967,
|
| 5979 |
+
"grad_norm": 0.5861655473709106,
|
| 5980 |
+
"learning_rate": 8.058296845706715e-06,
|
| 5981 |
+
"loss": 0.8638,
|
| 5982 |
+
"step": 843
|
| 5983 |
+
},
|
| 5984 |
+
{
|
| 5985 |
+
"epoch": 0.05878460734807592,
|
| 5986 |
+
"grad_norm": 0.7448881268501282,
|
| 5987 |
+
"learning_rate": 7.929553120622968e-06,
|
| 5988 |
+
"loss": 0.8458,
|
| 5989 |
+
"step": 844
|
| 5990 |
+
},
|
| 5991 |
+
{
|
| 5992 |
+
"epoch": 0.058854257356782166,
|
| 5993 |
+
"grad_norm": 0.5610641241073608,
|
| 5994 |
+
"learning_rate": 7.801803664361095e-06,
|
| 5995 |
+
"loss": 0.4706,
|
| 5996 |
+
"step": 845
|
| 5997 |
+
},
|
| 5998 |
+
{
|
| 5999 |
+
"epoch": 0.05892390736548842,
|
| 6000 |
+
"grad_norm": 0.5610293745994568,
|
| 6001 |
+
"learning_rate": 7.675049856487549e-06,
|
| 6002 |
+
"loss": 0.5503,
|
| 6003 |
+
"step": 846
|
| 6004 |
+
},
|
| 6005 |
+
{
|
| 6006 |
+
"epoch": 0.05899355737419467,
|
| 6007 |
+
"grad_norm": 0.6175963282585144,
|
| 6008 |
+
"learning_rate": 7.5492930658168096e-06,
|
| 6009 |
+
"loss": 0.6195,
|
| 6010 |
+
"step": 847
|
| 6011 |
+
},
|
| 6012 |
+
{
|
| 6013 |
+
"epoch": 0.059063207382900926,
|
| 6014 |
+
"grad_norm": 0.6749705672264099,
|
| 6015 |
+
"learning_rate": 7.42453465039652e-06,
|
| 6016 |
+
"loss": 0.7353,
|
| 6017 |
+
"step": 848
|
| 6018 |
+
},
|
| 6019 |
+
{
|
| 6020 |
+
"epoch": 0.059132857391607174,
|
| 6021 |
+
"grad_norm": 0.6812541484832764,
|
| 6022 |
+
"learning_rate": 7.300775957492923e-06,
|
| 6023 |
+
"loss": 0.6882,
|
| 6024 |
+
"step": 849
|
| 6025 |
+
},
|
| 6026 |
+
{
|
| 6027 |
+
"epoch": 0.05920250740031342,
|
| 6028 |
+
"grad_norm": 0.6131837368011475,
|
| 6029 |
+
"learning_rate": 7.178018323576208e-06,
|
| 6030 |
+
"loss": 0.9945,
|
| 6031 |
+
"step": 850
|
| 6032 |
+
},
|
| 6033 |
+
{
|
| 6034 |
+
"epoch": 0.05927215740901968,
|
| 6035 |
+
"grad_norm": 0.6159570217132568,
|
| 6036 |
+
"learning_rate": 7.056263074306191e-06,
|
| 6037 |
+
"loss": 0.7943,
|
| 6038 |
+
"step": 851
|
| 6039 |
+
},
|
| 6040 |
+
{
|
| 6041 |
+
"epoch": 0.05934180741772593,
|
| 6042 |
+
"grad_norm": 0.7175585627555847,
|
| 6043 |
+
"learning_rate": 6.935511524517835e-06,
|
| 6044 |
+
"loss": 0.8498,
|
| 6045 |
+
"step": 852
|
| 6046 |
+
},
|
| 6047 |
+
{
|
| 6048 |
+
"epoch": 0.059411457426432175,
|
| 6049 |
+
"grad_norm": 0.7083918452262878,
|
| 6050 |
+
"learning_rate": 6.815764978207284e-06,
|
| 6051 |
+
"loss": 0.9473,
|
| 6052 |
+
"step": 853
|
| 6053 |
+
},
|
| 6054 |
+
{
|
| 6055 |
+
"epoch": 0.05948110743513843,
|
| 6056 |
+
"grad_norm": 0.7349149584770203,
|
| 6057 |
+
"learning_rate": 6.6970247285175315e-06,
|
| 6058 |
+
"loss": 0.9025,
|
| 6059 |
+
"step": 854
|
| 6060 |
+
},
|
| 6061 |
+
{
|
| 6062 |
+
"epoch": 0.05955075744384468,
|
| 6063 |
+
"grad_norm": 0.6739192008972168,
|
| 6064 |
+
"learning_rate": 6.579292057724639e-06,
|
| 6065 |
+
"loss": 0.8435,
|
| 6066 |
+
"step": 855
|
| 6067 |
+
},
|
| 6068 |
+
{
|
| 6069 |
+
"epoch": 0.059620407452550934,
|
| 6070 |
+
"grad_norm": 0.6588095426559448,
|
| 6071 |
+
"learning_rate": 6.4625682372237874e-06,
|
| 6072 |
+
"loss": 0.6966,
|
| 6073 |
+
"step": 856
|
| 6074 |
+
},
|
| 6075 |
+
{
|
| 6076 |
+
"epoch": 0.05969005746125718,
|
| 6077 |
+
"grad_norm": 0.5185966491699219,
|
| 6078 |
+
"learning_rate": 6.346854527515622e-06,
|
| 6079 |
+
"loss": 0.6977,
|
| 6080 |
+
"step": 857
|
| 6081 |
+
},
|
| 6082 |
+
{
|
| 6083 |
+
"epoch": 0.05975970746996343,
|
| 6084 |
+
"grad_norm": 0.5705149173736572,
|
| 6085 |
+
"learning_rate": 6.23215217819253e-06,
|
| 6086 |
+
"loss": 0.6574,
|
| 6087 |
+
"step": 858
|
| 6088 |
+
},
|
| 6089 |
+
{
|
| 6090 |
+
"epoch": 0.05982935747866969,
|
| 6091 |
+
"grad_norm": 0.5465989112854004,
|
| 6092 |
+
"learning_rate": 6.11846242792532e-06,
|
| 6093 |
+
"loss": 0.5492,
|
| 6094 |
+
"step": 859
|
| 6095 |
+
},
|
| 6096 |
+
{
|
| 6097 |
+
"epoch": 0.059899007487375935,
|
| 6098 |
+
"grad_norm": 0.7820805311203003,
|
| 6099 |
+
"learning_rate": 6.005786504449651e-06,
|
| 6100 |
+
"loss": 0.8664,
|
| 6101 |
+
"step": 860
|
| 6102 |
+
},
|
| 6103 |
+
{
|
| 6104 |
+
"epoch": 0.059968657496082184,
|
| 6105 |
+
"grad_norm": 0.7436554431915283,
|
| 6106 |
+
"learning_rate": 5.894125624552915e-06,
|
| 6107 |
+
"loss": 0.9035,
|
| 6108 |
+
"step": 861
|
| 6109 |
+
},
|
| 6110 |
+
{
|
| 6111 |
+
"epoch": 0.06003830750478844,
|
| 6112 |
+
"grad_norm": 0.7402638792991638,
|
| 6113 |
+
"learning_rate": 5.7834809940610195e-06,
|
| 6114 |
+
"loss": 0.7703,
|
| 6115 |
+
"step": 862
|
| 6116 |
+
},
|
| 6117 |
+
{
|
| 6118 |
+
"epoch": 0.06010795751349469,
|
| 6119 |
+
"grad_norm": 0.6208961009979248,
|
| 6120 |
+
"learning_rate": 5.673853807825424e-06,
|
| 6121 |
+
"loss": 0.7226,
|
| 6122 |
+
"step": 863
|
| 6123 |
+
},
|
| 6124 |
+
{
|
| 6125 |
+
"epoch": 0.06017760752220094,
|
| 6126 |
+
"grad_norm": 0.5884114503860474,
|
| 6127 |
+
"learning_rate": 5.565245249710194e-06,
|
| 6128 |
+
"loss": 1.0493,
|
| 6129 |
+
"step": 864
|
| 6130 |
+
},
|
| 6131 |
+
{
|
| 6132 |
+
"epoch": 0.06024725753090719,
|
| 6133 |
+
"grad_norm": 0.7064511179924011,
|
| 6134 |
+
"learning_rate": 5.457656492579211e-06,
|
| 6135 |
+
"loss": 1.0538,
|
| 6136 |
+
"step": 865
|
| 6137 |
+
},
|
| 6138 |
+
{
|
| 6139 |
+
"epoch": 0.06031690753961344,
|
| 6140 |
+
"grad_norm": 0.714733362197876,
|
| 6141 |
+
"learning_rate": 5.351088698283558e-06,
|
| 6142 |
+
"loss": 0.7942,
|
| 6143 |
+
"step": 866
|
| 6144 |
+
},
|
| 6145 |
+
{
|
| 6146 |
+
"epoch": 0.060386557548319696,
|
| 6147 |
+
"grad_norm": 0.6394374966621399,
|
| 6148 |
+
"learning_rate": 5.2455430176489014e-06,
|
| 6149 |
+
"loss": 0.7437,
|
| 6150 |
+
"step": 867
|
| 6151 |
+
},
|
| 6152 |
+
{
|
| 6153 |
+
"epoch": 0.060456207557025944,
|
| 6154 |
+
"grad_norm": 0.6636267900466919,
|
| 6155 |
+
"learning_rate": 5.1410205904631415e-06,
|
| 6156 |
+
"loss": 0.8204,
|
| 6157 |
+
"step": 868
|
| 6158 |
+
},
|
| 6159 |
+
{
|
| 6160 |
+
"epoch": 0.06052585756573219,
|
| 6161 |
+
"grad_norm": 0.6036087274551392,
|
| 6162 |
+
"learning_rate": 5.037522545464024e-06,
|
| 6163 |
+
"loss": 0.7066,
|
| 6164 |
+
"step": 869
|
| 6165 |
+
},
|
| 6166 |
+
{
|
| 6167 |
+
"epoch": 0.06059550757443845,
|
| 6168 |
+
"grad_norm": 0.6227147579193115,
|
| 6169 |
+
"learning_rate": 4.9350500003270465e-06,
|
| 6170 |
+
"loss": 0.7101,
|
| 6171 |
+
"step": 870
|
| 6172 |
+
},
|
| 6173 |
+
{
|
| 6174 |
+
"epoch": 0.060665157583144697,
|
| 6175 |
+
"grad_norm": 0.5791090130805969,
|
| 6176 |
+
"learning_rate": 4.833604061653252e-06,
|
| 6177 |
+
"loss": 0.6439,
|
| 6178 |
+
"step": 871
|
| 6179 |
+
},
|
| 6180 |
+
{
|
| 6181 |
+
"epoch": 0.06073480759185095,
|
| 6182 |
+
"grad_norm": 0.5661488771438599,
|
| 6183 |
+
"learning_rate": 4.73318582495742e-06,
|
| 6184 |
+
"loss": 0.5134,
|
| 6185 |
+
"step": 872
|
| 6186 |
+
},
|
| 6187 |
+
{
|
| 6188 |
+
"epoch": 0.0608044576005572,
|
| 6189 |
+
"grad_norm": 0.7721818089485168,
|
| 6190 |
+
"learning_rate": 4.633796374656174e-06,
|
| 6191 |
+
"loss": 0.8566,
|
| 6192 |
+
"step": 873
|
| 6193 |
+
},
|
| 6194 |
+
{
|
| 6195 |
+
"epoch": 0.06087410760926345,
|
| 6196 |
+
"grad_norm": 0.7348571419715881,
|
| 6197 |
+
"learning_rate": 4.535436784056269e-06,
|
| 6198 |
+
"loss": 0.6653,
|
| 6199 |
+
"step": 874
|
| 6200 |
+
},
|
| 6201 |
+
{
|
| 6202 |
+
"epoch": 0.060943757617969704,
|
| 6203 |
+
"grad_norm": 0.6881682872772217,
|
| 6204 |
+
"learning_rate": 4.438108115342965e-06,
|
| 6205 |
+
"loss": 0.7876,
|
| 6206 |
+
"step": 875
|
| 6207 |
+
},
|
| 6208 |
+
{
|
| 6209 |
+
"epoch": 0.06101340762667595,
|
| 6210 |
+
"grad_norm": 0.6156147718429565,
|
| 6211 |
+
"learning_rate": 4.3418114195686536e-06,
|
| 6212 |
+
"loss": 0.8429,
|
| 6213 |
+
"step": 876
|
| 6214 |
+
},
|
| 6215 |
+
{
|
| 6216 |
+
"epoch": 0.0610830576353822,
|
| 6217 |
+
"grad_norm": 0.6420087218284607,
|
| 6218 |
+
"learning_rate": 4.246547736641382e-06,
|
| 6219 |
+
"loss": 0.7274,
|
| 6220 |
+
"step": 877
|
| 6221 |
+
},
|
| 6222 |
+
{
|
| 6223 |
+
"epoch": 0.06115270764408846,
|
| 6224 |
+
"grad_norm": 0.5134680271148682,
|
| 6225 |
+
"learning_rate": 4.152318095313778e-06,
|
| 6226 |
+
"loss": 0.5185,
|
| 6227 |
+
"step": 878
|
| 6228 |
+
},
|
| 6229 |
+
{
|
| 6230 |
+
"epoch": 0.061222357652794705,
|
| 6231 |
+
"grad_norm": 0.6913058757781982,
|
| 6232 |
+
"learning_rate": 4.05912351317177e-06,
|
| 6233 |
+
"loss": 0.9036,
|
| 6234 |
+
"step": 879
|
| 6235 |
+
},
|
| 6236 |
+
{
|
| 6237 |
+
"epoch": 0.06129200766150096,
|
| 6238 |
+
"grad_norm": 0.5641781091690063,
|
| 6239 |
+
"learning_rate": 3.966964996623735e-06,
|
| 6240 |
+
"loss": 0.8567,
|
| 6241 |
+
"step": 880
|
| 6242 |
+
},
|
| 6243 |
+
{
|
| 6244 |
+
"epoch": 0.06136165767020721,
|
| 6245 |
+
"grad_norm": 0.5682424306869507,
|
| 6246 |
+
"learning_rate": 3.875843540889546e-06,
|
| 6247 |
+
"loss": 0.7562,
|
| 6248 |
+
"step": 881
|
| 6249 |
+
},
|
| 6250 |
+
{
|
| 6251 |
+
"epoch": 0.06143130767891346,
|
| 6252 |
+
"grad_norm": 0.5852996110916138,
|
| 6253 |
+
"learning_rate": 3.785760129989868e-06,
|
| 6254 |
+
"loss": 0.4581,
|
| 6255 |
+
"step": 882
|
| 6256 |
+
},
|
| 6257 |
+
{
|
| 6258 |
+
"epoch": 0.06150095768761971,
|
| 6259 |
+
"grad_norm": 0.6625421047210693,
|
| 6260 |
+
"learning_rate": 3.6967157367355567e-06,
|
| 6261 |
+
"loss": 1.0613,
|
| 6262 |
+
"step": 883
|
| 6263 |
+
},
|
| 6264 |
+
{
|
| 6265 |
+
"epoch": 0.06157060769632596,
|
| 6266 |
+
"grad_norm": 0.7365720868110657,
|
| 6267 |
+
"learning_rate": 3.6087113227170287e-06,
|
| 6268 |
+
"loss": 0.8548,
|
| 6269 |
+
"step": 884
|
| 6270 |
+
},
|
| 6271 |
+
{
|
| 6272 |
+
"epoch": 0.06164025770503221,
|
| 6273 |
+
"grad_norm": 0.596820592880249,
|
| 6274 |
+
"learning_rate": 3.5217478382940426e-06,
|
| 6275 |
+
"loss": 0.7301,
|
| 6276 |
+
"step": 885
|
| 6277 |
+
},
|
| 6278 |
+
{
|
| 6279 |
+
"epoch": 0.061709907713738466,
|
| 6280 |
+
"grad_norm": 0.7230522632598877,
|
| 6281 |
+
"learning_rate": 3.4358262225853254e-06,
|
| 6282 |
+
"loss": 1.0264,
|
| 6283 |
+
"step": 886
|
| 6284 |
+
},
|
| 6285 |
+
{
|
| 6286 |
+
"epoch": 0.061779557722444714,
|
| 6287 |
+
"grad_norm": 0.550679087638855,
|
| 6288 |
+
"learning_rate": 3.3509474034584596e-06,
|
| 6289 |
+
"loss": 0.6914,
|
| 6290 |
+
"step": 887
|
| 6291 |
+
},
|
| 6292 |
+
{
|
| 6293 |
+
"epoch": 0.06184920773115097,
|
| 6294 |
+
"grad_norm": 0.6080251932144165,
|
| 6295 |
+
"learning_rate": 3.267112297519881e-06,
|
| 6296 |
+
"loss": 0.8706,
|
| 6297 |
+
"step": 888
|
| 6298 |
+
},
|
| 6299 |
+
{
|
| 6300 |
+
"epoch": 0.06191885773985722,
|
| 6301 |
+
"grad_norm": 0.6070705056190491,
|
| 6302 |
+
"learning_rate": 3.184321810104962e-06,
|
| 6303 |
+
"loss": 1.0111,
|
| 6304 |
+
"step": 889
|
| 6305 |
+
},
|
| 6306 |
+
{
|
| 6307 |
+
"epoch": 0.061988507748563466,
|
| 6308 |
+
"grad_norm": 0.6949368715286255,
|
| 6309 |
+
"learning_rate": 3.102576835268212e-06,
|
| 6310 |
+
"loss": 0.9892,
|
| 6311 |
+
"step": 890
|
| 6312 |
+
},
|
| 6313 |
+
{
|
| 6314 |
+
"epoch": 0.06205815775726972,
|
| 6315 |
+
"grad_norm": 0.7588335275650024,
|
| 6316 |
+
"learning_rate": 3.0218782557737136e-06,
|
| 6317 |
+
"loss": 0.8309,
|
| 6318 |
+
"step": 891
|
| 6319 |
+
},
|
| 6320 |
+
{
|
| 6321 |
+
"epoch": 0.06212780776597597,
|
| 6322 |
+
"grad_norm": 0.5684018135070801,
|
| 6323 |
+
"learning_rate": 2.9422269430854245e-06,
|
| 6324 |
+
"loss": 0.8553,
|
| 6325 |
+
"step": 892
|
| 6326 |
+
},
|
| 6327 |
+
{
|
| 6328 |
+
"epoch": 0.06219745777468222,
|
| 6329 |
+
"grad_norm": 0.554639995098114,
|
| 6330 |
+
"learning_rate": 2.863623757357992e-06,
|
| 6331 |
+
"loss": 0.7984,
|
| 6332 |
+
"step": 893
|
| 6333 |
+
},
|
| 6334 |
+
{
|
| 6335 |
+
"epoch": 0.062267107783388474,
|
| 6336 |
+
"grad_norm": 0.653669536113739,
|
| 6337 |
+
"learning_rate": 2.7860695474272392e-06,
|
| 6338 |
+
"loss": 0.8296,
|
| 6339 |
+
"step": 894
|
| 6340 |
+
},
|
| 6341 |
+
{
|
| 6342 |
+
"epoch": 0.06233675779209472,
|
| 6343 |
+
"grad_norm": 0.610150158405304,
|
| 6344 |
+
"learning_rate": 2.709565150801152e-06,
|
| 6345 |
+
"loss": 0.5203,
|
| 6346 |
+
"step": 895
|
| 6347 |
+
},
|
| 6348 |
+
{
|
| 6349 |
+
"epoch": 0.06240640780080098,
|
| 6350 |
+
"grad_norm": 0.6130475401878357,
|
| 6351 |
+
"learning_rate": 2.634111393650751e-06,
|
| 6352 |
+
"loss": 0.8298,
|
| 6353 |
+
"step": 896
|
| 6354 |
+
},
|
| 6355 |
+
{
|
| 6356 |
+
"epoch": 0.06247605780950723,
|
| 6357 |
+
"grad_norm": 0.5449431538581848,
|
| 6358 |
+
"learning_rate": 2.559709090801221e-06,
|
| 6359 |
+
"loss": 0.7497,
|
| 6360 |
+
"step": 897
|
| 6361 |
+
},
|
| 6362 |
+
{
|
| 6363 |
+
"epoch": 0.06254570781821348,
|
| 6364 |
+
"grad_norm": 0.6247503757476807,
|
| 6365 |
+
"learning_rate": 2.4863590457230743e-06,
|
| 6366 |
+
"loss": 1.1263,
|
| 6367 |
+
"step": 898
|
| 6368 |
+
},
|
| 6369 |
+
{
|
| 6370 |
+
"epoch": 0.06261535782691972,
|
| 6371 |
+
"grad_norm": 0.7267642617225647,
|
| 6372 |
+
"learning_rate": 2.4140620505235135e-06,
|
| 6373 |
+
"loss": 0.7873,
|
| 6374 |
+
"step": 899
|
| 6375 |
+
},
|
| 6376 |
+
{
|
| 6377 |
+
"epoch": 0.06268500783562599,
|
| 6378 |
+
"grad_norm": 0.7534024119377136,
|
| 6379 |
+
"learning_rate": 2.342818885937825e-06,
|
| 6380 |
+
"loss": 1.0745,
|
| 6381 |
+
"step": 900
|
| 6382 |
+
},
|
| 6383 |
+
{
|
| 6384 |
+
"epoch": 0.06268500783562599,
|
| 6385 |
+
"eval_loss": 0.723136305809021,
|
| 6386 |
+
"eval_runtime": 700.403,
|
| 6387 |
+
"eval_samples_per_second": 7.139,
|
| 6388 |
+
"eval_steps_per_second": 1.785,
|
| 6389 |
+
"step": 900
|
| 6390 |
}
|
| 6391 |
],
|
| 6392 |
"logging_steps": 1,
|
|
|
|
| 6415 |
"attributes": {}
|
| 6416 |
}
|
| 6417 |
},
|
| 6418 |
+
"total_flos": 2.448712204661293e+18,
|
| 6419 |
"train_batch_size": 4,
|
| 6420 |
"trial_name": null,
|
| 6421 |
"trial_params": null
|