Training in progress, step 588, 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 2185308216
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:1690b3d5a2c277fc6d6512b55275ba3778a2592e60d5e322d34dc5d26fb2f4a4
|
3 |
size 2185308216
|
last-checkpoint/optimizer.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 43123028
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f16ef706e2e18f5563735548a4b19de22b63f2f741ee8b31cba796df0b8b3182
|
3 |
size 43123028
|
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 sha256:010498655c7b8b87762448cf6be0c6688b7c4f40afdef8f1f5cf3e23305fc882
|
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 sha256:07b6de6a88182f3740cd3bffe2777cfccf71d3bd0df4c2888daece4f220876d2
|
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": 196,
|
6 |
-
"global_step":
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
@@ -2767,6 +2767,1386 @@
|
|
2767 |
"eval_samples_per_second": 14.366,
|
2768 |
"eval_steps_per_second": 7.205,
|
2769 |
"step": 392
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2770 |
}
|
2771 |
],
|
2772 |
"logging_steps": 1,
|
@@ -2786,7 +4166,7 @@
|
|
2786 |
"attributes": {}
|
2787 |
}
|
2788 |
},
|
2789 |
-
"total_flos":
|
2790 |
"train_batch_size": 2,
|
2791 |
"trial_name": null,
|
2792 |
"trial_params": null
|
|
|
1 |
{
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
+
"epoch": 0.7533632286995515,
|
5 |
"eval_steps": 196,
|
6 |
+
"global_step": 588,
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
|
|
2767 |
"eval_samples_per_second": 14.366,
|
2768 |
"eval_steps_per_second": 7.205,
|
2769 |
"step": 392
|
2770 |
+
},
|
2771 |
+
{
|
2772 |
+
"epoch": 0.5035233824471492,
|
2773 |
+
"grad_norm": 3.427626371383667,
|
2774 |
+
"learning_rate": 0.00010101865704284311,
|
2775 |
+
"loss": 1.4643,
|
2776 |
+
"step": 393
|
2777 |
+
},
|
2778 |
+
{
|
2779 |
+
"epoch": 0.5048046124279308,
|
2780 |
+
"grad_norm": 6.498578071594238,
|
2781 |
+
"learning_rate": 0.00010061120099094916,
|
2782 |
+
"loss": 2.0397,
|
2783 |
+
"step": 394
|
2784 |
+
},
|
2785 |
+
{
|
2786 |
+
"epoch": 0.5060858424087123,
|
2787 |
+
"grad_norm": 4.0323076248168945,
|
2788 |
+
"learning_rate": 0.0001002037347911959,
|
2789 |
+
"loss": 1.3606,
|
2790 |
+
"step": 395
|
2791 |
+
},
|
2792 |
+
{
|
2793 |
+
"epoch": 0.507367072389494,
|
2794 |
+
"grad_norm": 4.282688140869141,
|
2795 |
+
"learning_rate": 9.979626520880413e-05,
|
2796 |
+
"loss": 1.7831,
|
2797 |
+
"step": 396
|
2798 |
+
},
|
2799 |
+
{
|
2800 |
+
"epoch": 0.5086483023702755,
|
2801 |
+
"grad_norm": 4.8094048500061035,
|
2802 |
+
"learning_rate": 9.938879900905086e-05,
|
2803 |
+
"loss": 2.0107,
|
2804 |
+
"step": 397
|
2805 |
+
},
|
2806 |
+
{
|
2807 |
+
"epoch": 0.509929532351057,
|
2808 |
+
"grad_norm": 4.786434650421143,
|
2809 |
+
"learning_rate": 9.898134295715692e-05,
|
2810 |
+
"loss": 1.8023,
|
2811 |
+
"step": 398
|
2812 |
+
},
|
2813 |
+
{
|
2814 |
+
"epoch": 0.5112107623318386,
|
2815 |
+
"grad_norm": 6.043309688568115,
|
2816 |
+
"learning_rate": 9.857390381817463e-05,
|
2817 |
+
"loss": 1.6191,
|
2818 |
+
"step": 399
|
2819 |
+
},
|
2820 |
+
{
|
2821 |
+
"epoch": 0.5124919923126201,
|
2822 |
+
"grad_norm": 5.742677211761475,
|
2823 |
+
"learning_rate": 9.816648835687556e-05,
|
2824 |
+
"loss": 2.0572,
|
2825 |
+
"step": 400
|
2826 |
+
},
|
2827 |
+
{
|
2828 |
+
"epoch": 0.5137732222934016,
|
2829 |
+
"grad_norm": 1.4952232837677002,
|
2830 |
+
"learning_rate": 9.775910333763809e-05,
|
2831 |
+
"loss": 0.1042,
|
2832 |
+
"step": 401
|
2833 |
+
},
|
2834 |
+
{
|
2835 |
+
"epoch": 0.5150544522741832,
|
2836 |
+
"grad_norm": 3.1341373920440674,
|
2837 |
+
"learning_rate": 9.73517555243352e-05,
|
2838 |
+
"loss": 0.2243,
|
2839 |
+
"step": 402
|
2840 |
+
},
|
2841 |
+
{
|
2842 |
+
"epoch": 0.5163356822549647,
|
2843 |
+
"grad_norm": 0.6678221225738525,
|
2844 |
+
"learning_rate": 9.694445168022213e-05,
|
2845 |
+
"loss": 0.0248,
|
2846 |
+
"step": 403
|
2847 |
+
},
|
2848 |
+
{
|
2849 |
+
"epoch": 0.5176169122357464,
|
2850 |
+
"grad_norm": 1.330030083656311,
|
2851 |
+
"learning_rate": 9.653719856782412e-05,
|
2852 |
+
"loss": 0.1323,
|
2853 |
+
"step": 404
|
2854 |
+
},
|
2855 |
+
{
|
2856 |
+
"epoch": 0.5188981422165279,
|
2857 |
+
"grad_norm": 0.7335686683654785,
|
2858 |
+
"learning_rate": 9.613000294882405e-05,
|
2859 |
+
"loss": 0.0483,
|
2860 |
+
"step": 405
|
2861 |
+
},
|
2862 |
+
{
|
2863 |
+
"epoch": 0.5201793721973094,
|
2864 |
+
"grad_norm": 1.0830739736557007,
|
2865 |
+
"learning_rate": 9.572287158395025e-05,
|
2866 |
+
"loss": 0.0544,
|
2867 |
+
"step": 406
|
2868 |
+
},
|
2869 |
+
{
|
2870 |
+
"epoch": 0.521460602178091,
|
2871 |
+
"grad_norm": 2.981494903564453,
|
2872 |
+
"learning_rate": 9.531581123286425e-05,
|
2873 |
+
"loss": 0.5615,
|
2874 |
+
"step": 407
|
2875 |
+
},
|
2876 |
+
{
|
2877 |
+
"epoch": 0.5227418321588725,
|
2878 |
+
"grad_norm": 3.155945301055908,
|
2879 |
+
"learning_rate": 9.490882865404849e-05,
|
2880 |
+
"loss": 0.1728,
|
2881 |
+
"step": 408
|
2882 |
+
},
|
2883 |
+
{
|
2884 |
+
"epoch": 0.5240230621396541,
|
2885 |
+
"grad_norm": 0.6578575372695923,
|
2886 |
+
"learning_rate": 9.45019306046942e-05,
|
2887 |
+
"loss": 0.055,
|
2888 |
+
"step": 409
|
2889 |
+
},
|
2890 |
+
{
|
2891 |
+
"epoch": 0.5253042921204356,
|
2892 |
+
"grad_norm": 0.7729504108428955,
|
2893 |
+
"learning_rate": 9.409512384058915e-05,
|
2894 |
+
"loss": 0.0547,
|
2895 |
+
"step": 410
|
2896 |
+
},
|
2897 |
+
{
|
2898 |
+
"epoch": 0.5265855221012171,
|
2899 |
+
"grad_norm": 1.1161659955978394,
|
2900 |
+
"learning_rate": 9.368841511600539e-05,
|
2901 |
+
"loss": 0.1084,
|
2902 |
+
"step": 411
|
2903 |
+
},
|
2904 |
+
{
|
2905 |
+
"epoch": 0.5278667520819987,
|
2906 |
+
"grad_norm": 1.1031287908554077,
|
2907 |
+
"learning_rate": 9.328181118358735e-05,
|
2908 |
+
"loss": 0.1128,
|
2909 |
+
"step": 412
|
2910 |
+
},
|
2911 |
+
{
|
2912 |
+
"epoch": 0.5291479820627802,
|
2913 |
+
"grad_norm": 1.5451914072036743,
|
2914 |
+
"learning_rate": 9.287531879423943e-05,
|
2915 |
+
"loss": 0.1296,
|
2916 |
+
"step": 413
|
2917 |
+
},
|
2918 |
+
{
|
2919 |
+
"epoch": 0.5304292120435619,
|
2920 |
+
"grad_norm": 0.7384908199310303,
|
2921 |
+
"learning_rate": 9.246894469701422e-05,
|
2922 |
+
"loss": 0.053,
|
2923 |
+
"step": 414
|
2924 |
+
},
|
2925 |
+
{
|
2926 |
+
"epoch": 0.5317104420243434,
|
2927 |
+
"grad_norm": 0.9731848239898682,
|
2928 |
+
"learning_rate": 9.206269563900013e-05,
|
2929 |
+
"loss": 0.0716,
|
2930 |
+
"step": 415
|
2931 |
+
},
|
2932 |
+
{
|
2933 |
+
"epoch": 0.5329916720051249,
|
2934 |
+
"grad_norm": 3.4970998764038086,
|
2935 |
+
"learning_rate": 9.165657836520969e-05,
|
2936 |
+
"loss": 0.3843,
|
2937 |
+
"step": 416
|
2938 |
+
},
|
2939 |
+
{
|
2940 |
+
"epoch": 0.5342729019859065,
|
2941 |
+
"grad_norm": 1.0508249998092651,
|
2942 |
+
"learning_rate": 9.125059961846723e-05,
|
2943 |
+
"loss": 0.1178,
|
2944 |
+
"step": 417
|
2945 |
+
},
|
2946 |
+
{
|
2947 |
+
"epoch": 0.535554131966688,
|
2948 |
+
"grad_norm": 1.1686064004898071,
|
2949 |
+
"learning_rate": 9.084476613929725e-05,
|
2950 |
+
"loss": 0.2156,
|
2951 |
+
"step": 418
|
2952 |
+
},
|
2953 |
+
{
|
2954 |
+
"epoch": 0.5368353619474696,
|
2955 |
+
"grad_norm": 1.9709417819976807,
|
2956 |
+
"learning_rate": 9.043908466581224e-05,
|
2957 |
+
"loss": 0.1223,
|
2958 |
+
"step": 419
|
2959 |
+
},
|
2960 |
+
{
|
2961 |
+
"epoch": 0.5381165919282511,
|
2962 |
+
"grad_norm": 0.6963585615158081,
|
2963 |
+
"learning_rate": 9.003356193360098e-05,
|
2964 |
+
"loss": 0.0213,
|
2965 |
+
"step": 420
|
2966 |
+
},
|
2967 |
+
{
|
2968 |
+
"epoch": 0.5393978219090326,
|
2969 |
+
"grad_norm": 1.5098761320114136,
|
2970 |
+
"learning_rate": 8.962820467561659e-05,
|
2971 |
+
"loss": 0.1003,
|
2972 |
+
"step": 421
|
2973 |
+
},
|
2974 |
+
{
|
2975 |
+
"epoch": 0.5406790518898142,
|
2976 |
+
"grad_norm": 1.374016523361206,
|
2977 |
+
"learning_rate": 8.922301962206486e-05,
|
2978 |
+
"loss": 0.1586,
|
2979 |
+
"step": 422
|
2980 |
+
},
|
2981 |
+
{
|
2982 |
+
"epoch": 0.5419602818705958,
|
2983 |
+
"grad_norm": 1.8228262662887573,
|
2984 |
+
"learning_rate": 8.881801350029241e-05,
|
2985 |
+
"loss": 0.2103,
|
2986 |
+
"step": 423
|
2987 |
+
},
|
2988 |
+
{
|
2989 |
+
"epoch": 0.5432415118513774,
|
2990 |
+
"grad_norm": 1.180014967918396,
|
2991 |
+
"learning_rate": 8.841319303467502e-05,
|
2992 |
+
"loss": 0.1432,
|
2993 |
+
"step": 424
|
2994 |
+
},
|
2995 |
+
{
|
2996 |
+
"epoch": 0.5445227418321589,
|
2997 |
+
"grad_norm": 2.189598560333252,
|
2998 |
+
"learning_rate": 8.800856494650601e-05,
|
2999 |
+
"loss": 0.1549,
|
3000 |
+
"step": 425
|
3001 |
+
},
|
3002 |
+
{
|
3003 |
+
"epoch": 0.5458039718129404,
|
3004 |
+
"grad_norm": 1.5259344577789307,
|
3005 |
+
"learning_rate": 8.760413595388462e-05,
|
3006 |
+
"loss": 0.2188,
|
3007 |
+
"step": 426
|
3008 |
+
},
|
3009 |
+
{
|
3010 |
+
"epoch": 0.547085201793722,
|
3011 |
+
"grad_norm": 1.1365989446640015,
|
3012 |
+
"learning_rate": 8.719991277160452e-05,
|
3013 |
+
"loss": 0.1117,
|
3014 |
+
"step": 427
|
3015 |
+
},
|
3016 |
+
{
|
3017 |
+
"epoch": 0.5483664317745035,
|
3018 |
+
"grad_norm": 3.994901418685913,
|
3019 |
+
"learning_rate": 8.679590211104221e-05,
|
3020 |
+
"loss": 0.3669,
|
3021 |
+
"step": 428
|
3022 |
+
},
|
3023 |
+
{
|
3024 |
+
"epoch": 0.5496476617552851,
|
3025 |
+
"grad_norm": 0.49799466133117676,
|
3026 |
+
"learning_rate": 8.639211068004572e-05,
|
3027 |
+
"loss": 0.0361,
|
3028 |
+
"step": 429
|
3029 |
+
},
|
3030 |
+
{
|
3031 |
+
"epoch": 0.5509288917360666,
|
3032 |
+
"grad_norm": 4.062549114227295,
|
3033 |
+
"learning_rate": 8.598854518282311e-05,
|
3034 |
+
"loss": 0.3875,
|
3035 |
+
"step": 430
|
3036 |
+
},
|
3037 |
+
{
|
3038 |
+
"epoch": 0.5522101217168481,
|
3039 |
+
"grad_norm": 2.4972500801086426,
|
3040 |
+
"learning_rate": 8.558521231983133e-05,
|
3041 |
+
"loss": 0.361,
|
3042 |
+
"step": 431
|
3043 |
+
},
|
3044 |
+
{
|
3045 |
+
"epoch": 0.5534913516976298,
|
3046 |
+
"grad_norm": 1.7171235084533691,
|
3047 |
+
"learning_rate": 8.518211878766476e-05,
|
3048 |
+
"loss": 0.2779,
|
3049 |
+
"step": 432
|
3050 |
+
},
|
3051 |
+
{
|
3052 |
+
"epoch": 0.5547725816784113,
|
3053 |
+
"grad_norm": 1.568215012550354,
|
3054 |
+
"learning_rate": 8.477927127894424e-05,
|
3055 |
+
"loss": 0.1972,
|
3056 |
+
"step": 433
|
3057 |
+
},
|
3058 |
+
{
|
3059 |
+
"epoch": 0.5560538116591929,
|
3060 |
+
"grad_norm": 1.996573567390442,
|
3061 |
+
"learning_rate": 8.437667648220579e-05,
|
3062 |
+
"loss": 0.295,
|
3063 |
+
"step": 434
|
3064 |
+
},
|
3065 |
+
{
|
3066 |
+
"epoch": 0.5573350416399744,
|
3067 |
+
"grad_norm": 0.35408028960227966,
|
3068 |
+
"learning_rate": 8.39743410817896e-05,
|
3069 |
+
"loss": 0.0285,
|
3070 |
+
"step": 435
|
3071 |
+
},
|
3072 |
+
{
|
3073 |
+
"epoch": 0.5586162716207559,
|
3074 |
+
"grad_norm": 0.7227452397346497,
|
3075 |
+
"learning_rate": 8.35722717577291e-05,
|
3076 |
+
"loss": 0.0462,
|
3077 |
+
"step": 436
|
3078 |
+
},
|
3079 |
+
{
|
3080 |
+
"epoch": 0.5598975016015375,
|
3081 |
+
"grad_norm": 0.648615300655365,
|
3082 |
+
"learning_rate": 8.317047518564e-05,
|
3083 |
+
"loss": 0.0869,
|
3084 |
+
"step": 437
|
3085 |
+
},
|
3086 |
+
{
|
3087 |
+
"epoch": 0.561178731582319,
|
3088 |
+
"grad_norm": 0.2627539038658142,
|
3089 |
+
"learning_rate": 8.276895803660947e-05,
|
3090 |
+
"loss": 0.0195,
|
3091 |
+
"step": 438
|
3092 |
+
},
|
3093 |
+
{
|
3094 |
+
"epoch": 0.5624599615631006,
|
3095 |
+
"grad_norm": 1.0580172538757324,
|
3096 |
+
"learning_rate": 8.236772697708537e-05,
|
3097 |
+
"loss": 0.1161,
|
3098 |
+
"step": 439
|
3099 |
+
},
|
3100 |
+
{
|
3101 |
+
"epoch": 0.5637411915438821,
|
3102 |
+
"grad_norm": 0.5673841238021851,
|
3103 |
+
"learning_rate": 8.196678866876561e-05,
|
3104 |
+
"loss": 0.0311,
|
3105 |
+
"step": 440
|
3106 |
+
},
|
3107 |
+
{
|
3108 |
+
"epoch": 0.5650224215246636,
|
3109 |
+
"grad_norm": 0.844033420085907,
|
3110 |
+
"learning_rate": 8.156614976848747e-05,
|
3111 |
+
"loss": 0.2755,
|
3112 |
+
"step": 441
|
3113 |
+
},
|
3114 |
+
{
|
3115 |
+
"epoch": 0.5663036515054453,
|
3116 |
+
"grad_norm": 0.27234673500061035,
|
3117 |
+
"learning_rate": 8.116581692811711e-05,
|
3118 |
+
"loss": 0.0246,
|
3119 |
+
"step": 442
|
3120 |
+
},
|
3121 |
+
{
|
3122 |
+
"epoch": 0.5675848814862268,
|
3123 |
+
"grad_norm": 3.3813579082489014,
|
3124 |
+
"learning_rate": 8.076579679443919e-05,
|
3125 |
+
"loss": 2.2328,
|
3126 |
+
"step": 443
|
3127 |
+
},
|
3128 |
+
{
|
3129 |
+
"epoch": 0.5688661114670083,
|
3130 |
+
"grad_norm": 3.2380542755126953,
|
3131 |
+
"learning_rate": 8.036609600904636e-05,
|
3132 |
+
"loss": 1.694,
|
3133 |
+
"step": 444
|
3134 |
+
},
|
3135 |
+
{
|
3136 |
+
"epoch": 0.5701473414477899,
|
3137 |
+
"grad_norm": 3.2565391063690186,
|
3138 |
+
"learning_rate": 7.996672120822918e-05,
|
3139 |
+
"loss": 1.3883,
|
3140 |
+
"step": 445
|
3141 |
+
},
|
3142 |
+
{
|
3143 |
+
"epoch": 0.5714285714285714,
|
3144 |
+
"grad_norm": 3.4468626976013184,
|
3145 |
+
"learning_rate": 7.956767902286579e-05,
|
3146 |
+
"loss": 1.653,
|
3147 |
+
"step": 446
|
3148 |
+
},
|
3149 |
+
{
|
3150 |
+
"epoch": 0.572709801409353,
|
3151 |
+
"grad_norm": 4.3740315437316895,
|
3152 |
+
"learning_rate": 7.916897607831188e-05,
|
3153 |
+
"loss": 1.9767,
|
3154 |
+
"step": 447
|
3155 |
+
},
|
3156 |
+
{
|
3157 |
+
"epoch": 0.5739910313901345,
|
3158 |
+
"grad_norm": 4.42133903503418,
|
3159 |
+
"learning_rate": 7.877061899429066e-05,
|
3160 |
+
"loss": 1.7129,
|
3161 |
+
"step": 448
|
3162 |
+
},
|
3163 |
+
{
|
3164 |
+
"epoch": 0.575272261370916,
|
3165 |
+
"grad_norm": 3.089538097381592,
|
3166 |
+
"learning_rate": 7.8372614384783e-05,
|
3167 |
+
"loss": 1.0838,
|
3168 |
+
"step": 449
|
3169 |
+
},
|
3170 |
+
{
|
3171 |
+
"epoch": 0.5765534913516976,
|
3172 |
+
"grad_norm": 4.295858383178711,
|
3173 |
+
"learning_rate": 7.797496885791753e-05,
|
3174 |
+
"loss": 2.2591,
|
3175 |
+
"step": 450
|
3176 |
+
},
|
3177 |
+
{
|
3178 |
+
"epoch": 0.5778347213324792,
|
3179 |
+
"grad_norm": 3.2733049392700195,
|
3180 |
+
"learning_rate": 7.757768901586106e-05,
|
3181 |
+
"loss": 0.4728,
|
3182 |
+
"step": 451
|
3183 |
+
},
|
3184 |
+
{
|
3185 |
+
"epoch": 0.5791159513132608,
|
3186 |
+
"grad_norm": 3.444890022277832,
|
3187 |
+
"learning_rate": 7.718078145470877e-05,
|
3188 |
+
"loss": 0.4599,
|
3189 |
+
"step": 452
|
3190 |
+
},
|
3191 |
+
{
|
3192 |
+
"epoch": 0.5803971812940423,
|
3193 |
+
"grad_norm": 1.9771506786346436,
|
3194 |
+
"learning_rate": 7.678425276437493e-05,
|
3195 |
+
"loss": 0.2354,
|
3196 |
+
"step": 453
|
3197 |
+
},
|
3198 |
+
{
|
3199 |
+
"epoch": 0.5816784112748238,
|
3200 |
+
"grad_norm": 2.4951820373535156,
|
3201 |
+
"learning_rate": 7.638810952848326e-05,
|
3202 |
+
"loss": 0.2286,
|
3203 |
+
"step": 454
|
3204 |
+
},
|
3205 |
+
{
|
3206 |
+
"epoch": 0.5829596412556054,
|
3207 |
+
"grad_norm": 1.1922607421875,
|
3208 |
+
"learning_rate": 7.599235832425781e-05,
|
3209 |
+
"loss": 0.1114,
|
3210 |
+
"step": 455
|
3211 |
+
},
|
3212 |
+
{
|
3213 |
+
"epoch": 0.5842408712363869,
|
3214 |
+
"grad_norm": 1.61778724193573,
|
3215 |
+
"learning_rate": 7.559700572241354e-05,
|
3216 |
+
"loss": 0.2136,
|
3217 |
+
"step": 456
|
3218 |
+
},
|
3219 |
+
{
|
3220 |
+
"epoch": 0.5855221012171685,
|
3221 |
+
"grad_norm": 1.7456979751586914,
|
3222 |
+
"learning_rate": 7.520205828704753e-05,
|
3223 |
+
"loss": 0.2614,
|
3224 |
+
"step": 457
|
3225 |
+
},
|
3226 |
+
{
|
3227 |
+
"epoch": 0.58680333119795,
|
3228 |
+
"grad_norm": 0.6737706065177917,
|
3229 |
+
"learning_rate": 7.480752257552971e-05,
|
3230 |
+
"loss": 0.108,
|
3231 |
+
"step": 458
|
3232 |
+
},
|
3233 |
+
{
|
3234 |
+
"epoch": 0.5880845611787315,
|
3235 |
+
"grad_norm": 1.5337079763412476,
|
3236 |
+
"learning_rate": 7.441340513839413e-05,
|
3237 |
+
"loss": 0.2073,
|
3238 |
+
"step": 459
|
3239 |
+
},
|
3240 |
+
{
|
3241 |
+
"epoch": 0.5893657911595132,
|
3242 |
+
"grad_norm": 0.7655806541442871,
|
3243 |
+
"learning_rate": 7.401971251923014e-05,
|
3244 |
+
"loss": 0.084,
|
3245 |
+
"step": 460
|
3246 |
+
},
|
3247 |
+
{
|
3248 |
+
"epoch": 0.5906470211402947,
|
3249 |
+
"grad_norm": 0.8033196926116943,
|
3250 |
+
"learning_rate": 7.362645125457382e-05,
|
3251 |
+
"loss": 0.1327,
|
3252 |
+
"step": 461
|
3253 |
+
},
|
3254 |
+
{
|
3255 |
+
"epoch": 0.5919282511210763,
|
3256 |
+
"grad_norm": 2.1315033435821533,
|
3257 |
+
"learning_rate": 7.323362787379935e-05,
|
3258 |
+
"loss": 0.2699,
|
3259 |
+
"step": 462
|
3260 |
+
},
|
3261 |
+
{
|
3262 |
+
"epoch": 0.5932094811018578,
|
3263 |
+
"grad_norm": 0.2412717342376709,
|
3264 |
+
"learning_rate": 7.284124889901074e-05,
|
3265 |
+
"loss": 0.0146,
|
3266 |
+
"step": 463
|
3267 |
+
},
|
3268 |
+
{
|
3269 |
+
"epoch": 0.5944907110826393,
|
3270 |
+
"grad_norm": 0.9515855312347412,
|
3271 |
+
"learning_rate": 7.244932084493338e-05,
|
3272 |
+
"loss": 0.0853,
|
3273 |
+
"step": 464
|
3274 |
+
},
|
3275 |
+
{
|
3276 |
+
"epoch": 0.5957719410634209,
|
3277 |
+
"grad_norm": 0.3724261522293091,
|
3278 |
+
"learning_rate": 7.205785021880602e-05,
|
3279 |
+
"loss": 0.0323,
|
3280 |
+
"step": 465
|
3281 |
+
},
|
3282 |
+
{
|
3283 |
+
"epoch": 0.5970531710442024,
|
3284 |
+
"grad_norm": 1.7310974597930908,
|
3285 |
+
"learning_rate": 7.166684352027265e-05,
|
3286 |
+
"loss": 0.2304,
|
3287 |
+
"step": 466
|
3288 |
+
},
|
3289 |
+
{
|
3290 |
+
"epoch": 0.598334401024984,
|
3291 |
+
"grad_norm": 1.0784680843353271,
|
3292 |
+
"learning_rate": 7.127630724127456e-05,
|
3293 |
+
"loss": 0.1609,
|
3294 |
+
"step": 467
|
3295 |
+
},
|
3296 |
+
{
|
3297 |
+
"epoch": 0.5996156310057655,
|
3298 |
+
"grad_norm": 2.769951105117798,
|
3299 |
+
"learning_rate": 7.088624786594265e-05,
|
3300 |
+
"loss": 0.4748,
|
3301 |
+
"step": 468
|
3302 |
+
},
|
3303 |
+
{
|
3304 |
+
"epoch": 0.600896860986547,
|
3305 |
+
"grad_norm": 2.351362943649292,
|
3306 |
+
"learning_rate": 7.049667187048974e-05,
|
3307 |
+
"loss": 0.405,
|
3308 |
+
"step": 469
|
3309 |
+
},
|
3310 |
+
{
|
3311 |
+
"epoch": 0.6021780909673287,
|
3312 |
+
"grad_norm": 1.630505084991455,
|
3313 |
+
"learning_rate": 7.01075857231029e-05,
|
3314 |
+
"loss": 0.1992,
|
3315 |
+
"step": 470
|
3316 |
+
},
|
3317 |
+
{
|
3318 |
+
"epoch": 0.6034593209481102,
|
3319 |
+
"grad_norm": 1.5354177951812744,
|
3320 |
+
"learning_rate": 6.971899588383635e-05,
|
3321 |
+
"loss": 0.1843,
|
3322 |
+
"step": 471
|
3323 |
+
},
|
3324 |
+
{
|
3325 |
+
"epoch": 0.6047405509288918,
|
3326 |
+
"grad_norm": 0.5430020093917847,
|
3327 |
+
"learning_rate": 6.933090880450386e-05,
|
3328 |
+
"loss": 0.0598,
|
3329 |
+
"step": 472
|
3330 |
+
},
|
3331 |
+
{
|
3332 |
+
"epoch": 0.6060217809096733,
|
3333 |
+
"grad_norm": 0.13371486961841583,
|
3334 |
+
"learning_rate": 6.894333092857199e-05,
|
3335 |
+
"loss": 0.0066,
|
3336 |
+
"step": 473
|
3337 |
+
},
|
3338 |
+
{
|
3339 |
+
"epoch": 0.6073030108904548,
|
3340 |
+
"grad_norm": 2.9969446659088135,
|
3341 |
+
"learning_rate": 6.855626869105272e-05,
|
3342 |
+
"loss": 0.0892,
|
3343 |
+
"step": 474
|
3344 |
+
},
|
3345 |
+
{
|
3346 |
+
"epoch": 0.6085842408712364,
|
3347 |
+
"grad_norm": 2.9847252368927,
|
3348 |
+
"learning_rate": 6.816972851839701e-05,
|
3349 |
+
"loss": 0.5168,
|
3350 |
+
"step": 475
|
3351 |
+
},
|
3352 |
+
{
|
3353 |
+
"epoch": 0.6098654708520179,
|
3354 |
+
"grad_norm": 1.1502326726913452,
|
3355 |
+
"learning_rate": 6.778371682838774e-05,
|
3356 |
+
"loss": 0.169,
|
3357 |
+
"step": 476
|
3358 |
+
},
|
3359 |
+
{
|
3360 |
+
"epoch": 0.6111467008327995,
|
3361 |
+
"grad_norm": 1.1469147205352783,
|
3362 |
+
"learning_rate": 6.739824003003339e-05,
|
3363 |
+
"loss": 0.2161,
|
3364 |
+
"step": 477
|
3365 |
+
},
|
3366 |
+
{
|
3367 |
+
"epoch": 0.612427930813581,
|
3368 |
+
"grad_norm": 0.36965712904930115,
|
3369 |
+
"learning_rate": 6.701330452346155e-05,
|
3370 |
+
"loss": 0.0274,
|
3371 |
+
"step": 478
|
3372 |
+
},
|
3373 |
+
{
|
3374 |
+
"epoch": 0.6137091607943626,
|
3375 |
+
"grad_norm": 1.3748255968093872,
|
3376 |
+
"learning_rate": 6.662891669981266e-05,
|
3377 |
+
"loss": 0.1288,
|
3378 |
+
"step": 479
|
3379 |
+
},
|
3380 |
+
{
|
3381 |
+
"epoch": 0.6149903907751442,
|
3382 |
+
"grad_norm": 0.31070318818092346,
|
3383 |
+
"learning_rate": 6.624508294113387e-05,
|
3384 |
+
"loss": 0.0173,
|
3385 |
+
"step": 480
|
3386 |
+
},
|
3387 |
+
{
|
3388 |
+
"epoch": 0.6162716207559257,
|
3389 |
+
"grad_norm": 1.3732202053070068,
|
3390 |
+
"learning_rate": 6.586180962027322e-05,
|
3391 |
+
"loss": 0.1201,
|
3392 |
+
"step": 481
|
3393 |
+
},
|
3394 |
+
{
|
3395 |
+
"epoch": 0.6175528507367073,
|
3396 |
+
"grad_norm": 2.199179172515869,
|
3397 |
+
"learning_rate": 6.547910310077354e-05,
|
3398 |
+
"loss": 0.1139,
|
3399 |
+
"step": 482
|
3400 |
+
},
|
3401 |
+
{
|
3402 |
+
"epoch": 0.6188340807174888,
|
3403 |
+
"grad_norm": 1.4493154287338257,
|
3404 |
+
"learning_rate": 6.509696973676714e-05,
|
3405 |
+
"loss": 0.1728,
|
3406 |
+
"step": 483
|
3407 |
+
},
|
3408 |
+
{
|
3409 |
+
"epoch": 0.6201153106982703,
|
3410 |
+
"grad_norm": 0.4997680187225342,
|
3411 |
+
"learning_rate": 6.471541587287003e-05,
|
3412 |
+
"loss": 0.0271,
|
3413 |
+
"step": 484
|
3414 |
+
},
|
3415 |
+
{
|
3416 |
+
"epoch": 0.6213965406790519,
|
3417 |
+
"grad_norm": 1.1124134063720703,
|
3418 |
+
"learning_rate": 6.433444784407679e-05,
|
3419 |
+
"loss": 0.1223,
|
3420 |
+
"step": 485
|
3421 |
+
},
|
3422 |
+
{
|
3423 |
+
"epoch": 0.6226777706598334,
|
3424 |
+
"grad_norm": 1.0629478693008423,
|
3425 |
+
"learning_rate": 6.395407197565518e-05,
|
3426 |
+
"loss": 0.1356,
|
3427 |
+
"step": 486
|
3428 |
+
},
|
3429 |
+
{
|
3430 |
+
"epoch": 0.6239590006406149,
|
3431 |
+
"grad_norm": 0.7441665530204773,
|
3432 |
+
"learning_rate": 6.357429458304135e-05,
|
3433 |
+
"loss": 0.0452,
|
3434 |
+
"step": 487
|
3435 |
+
},
|
3436 |
+
{
|
3437 |
+
"epoch": 0.6252402306213966,
|
3438 |
+
"grad_norm": 0.6553467512130737,
|
3439 |
+
"learning_rate": 6.319512197173479e-05,
|
3440 |
+
"loss": 0.0352,
|
3441 |
+
"step": 488
|
3442 |
+
},
|
3443 |
+
{
|
3444 |
+
"epoch": 0.6265214606021781,
|
3445 |
+
"grad_norm": 2.700345516204834,
|
3446 |
+
"learning_rate": 6.281656043719378e-05,
|
3447 |
+
"loss": 0.2802,
|
3448 |
+
"step": 489
|
3449 |
+
},
|
3450 |
+
{
|
3451 |
+
"epoch": 0.6278026905829597,
|
3452 |
+
"grad_norm": 0.47781315445899963,
|
3453 |
+
"learning_rate": 6.243861626473073e-05,
|
3454 |
+
"loss": 0.0519,
|
3455 |
+
"step": 490
|
3456 |
+
},
|
3457 |
+
{
|
3458 |
+
"epoch": 0.6290839205637412,
|
3459 |
+
"grad_norm": 3.317213535308838,
|
3460 |
+
"learning_rate": 6.206129572940798e-05,
|
3461 |
+
"loss": 1.6523,
|
3462 |
+
"step": 491
|
3463 |
+
},
|
3464 |
+
{
|
3465 |
+
"epoch": 0.6303651505445227,
|
3466 |
+
"grad_norm": 3.361447334289551,
|
3467 |
+
"learning_rate": 6.168460509593343e-05,
|
3468 |
+
"loss": 1.1515,
|
3469 |
+
"step": 492
|
3470 |
+
},
|
3471 |
+
{
|
3472 |
+
"epoch": 0.6316463805253043,
|
3473 |
+
"grad_norm": 4.26183557510376,
|
3474 |
+
"learning_rate": 6.130855061855676e-05,
|
3475 |
+
"loss": 1.511,
|
3476 |
+
"step": 493
|
3477 |
+
},
|
3478 |
+
{
|
3479 |
+
"epoch": 0.6329276105060858,
|
3480 |
+
"grad_norm": 4.117159366607666,
|
3481 |
+
"learning_rate": 6.093313854096534e-05,
|
3482 |
+
"loss": 1.753,
|
3483 |
+
"step": 494
|
3484 |
+
},
|
3485 |
+
{
|
3486 |
+
"epoch": 0.6342088404868674,
|
3487 |
+
"grad_norm": 4.344640731811523,
|
3488 |
+
"learning_rate": 6.055837509618072e-05,
|
3489 |
+
"loss": 1.7018,
|
3490 |
+
"step": 495
|
3491 |
+
},
|
3492 |
+
{
|
3493 |
+
"epoch": 0.6354900704676489,
|
3494 |
+
"grad_norm": 5.007361888885498,
|
3495 |
+
"learning_rate": 6.018426650645512e-05,
|
3496 |
+
"loss": 2.035,
|
3497 |
+
"step": 496
|
3498 |
+
},
|
3499 |
+
{
|
3500 |
+
"epoch": 0.6367713004484304,
|
3501 |
+
"grad_norm": 4.537297248840332,
|
3502 |
+
"learning_rate": 5.981081898316809e-05,
|
3503 |
+
"loss": 1.7824,
|
3504 |
+
"step": 497
|
3505 |
+
},
|
3506 |
+
{
|
3507 |
+
"epoch": 0.6380525304292121,
|
3508 |
+
"grad_norm": 4.092187881469727,
|
3509 |
+
"learning_rate": 5.943803872672343e-05,
|
3510 |
+
"loss": 1.3558,
|
3511 |
+
"step": 498
|
3512 |
+
},
|
3513 |
+
{
|
3514 |
+
"epoch": 0.6393337604099936,
|
3515 |
+
"grad_norm": 4.109410285949707,
|
3516 |
+
"learning_rate": 5.9065931926446146e-05,
|
3517 |
+
"loss": 1.3504,
|
3518 |
+
"step": 499
|
3519 |
+
},
|
3520 |
+
{
|
3521 |
+
"epoch": 0.6406149903907752,
|
3522 |
+
"grad_norm": 4.25268030166626,
|
3523 |
+
"learning_rate": 5.869450476047984e-05,
|
3524 |
+
"loss": 1.9729,
|
3525 |
+
"step": 500
|
3526 |
+
},
|
3527 |
+
{
|
3528 |
+
"epoch": 0.6418962203715567,
|
3529 |
+
"grad_norm": 0.7322160005569458,
|
3530 |
+
"learning_rate": 5.832376339568394e-05,
|
3531 |
+
"loss": 0.0662,
|
3532 |
+
"step": 501
|
3533 |
+
},
|
3534 |
+
{
|
3535 |
+
"epoch": 0.6431774503523382,
|
3536 |
+
"grad_norm": 0.7809010148048401,
|
3537 |
+
"learning_rate": 5.795371398753153e-05,
|
3538 |
+
"loss": 0.0924,
|
3539 |
+
"step": 502
|
3540 |
+
},
|
3541 |
+
{
|
3542 |
+
"epoch": 0.6444586803331198,
|
3543 |
+
"grad_norm": 0.3986906409263611,
|
3544 |
+
"learning_rate": 5.7584362680006954e-05,
|
3545 |
+
"loss": 0.0227,
|
3546 |
+
"step": 503
|
3547 |
+
},
|
3548 |
+
{
|
3549 |
+
"epoch": 0.6457399103139013,
|
3550 |
+
"grad_norm": 2.2670533657073975,
|
3551 |
+
"learning_rate": 5.72157156055039e-05,
|
3552 |
+
"loss": 0.1165,
|
3553 |
+
"step": 504
|
3554 |
+
},
|
3555 |
+
{
|
3556 |
+
"epoch": 0.6470211402946829,
|
3557 |
+
"grad_norm": 1.28682279586792,
|
3558 |
+
"learning_rate": 5.684777888472359e-05,
|
3559 |
+
"loss": 0.2821,
|
3560 |
+
"step": 505
|
3561 |
+
},
|
3562 |
+
{
|
3563 |
+
"epoch": 0.6483023702754644,
|
3564 |
+
"grad_norm": 1.957733154296875,
|
3565 |
+
"learning_rate": 5.648055862657321e-05,
|
3566 |
+
"loss": 0.1474,
|
3567 |
+
"step": 506
|
3568 |
+
},
|
3569 |
+
{
|
3570 |
+
"epoch": 0.649583600256246,
|
3571 |
+
"grad_norm": 2.411461591720581,
|
3572 |
+
"learning_rate": 5.6114060928064296e-05,
|
3573 |
+
"loss": 0.0918,
|
3574 |
+
"step": 507
|
3575 |
+
},
|
3576 |
+
{
|
3577 |
+
"epoch": 0.6508648302370276,
|
3578 |
+
"grad_norm": 1.7074830532073975,
|
3579 |
+
"learning_rate": 5.574829187421166e-05,
|
3580 |
+
"loss": 0.2007,
|
3581 |
+
"step": 508
|
3582 |
+
},
|
3583 |
+
{
|
3584 |
+
"epoch": 0.6521460602178091,
|
3585 |
+
"grad_norm": 2.526116132736206,
|
3586 |
+
"learning_rate": 5.538325753793228e-05,
|
3587 |
+
"loss": 0.0466,
|
3588 |
+
"step": 509
|
3589 |
+
},
|
3590 |
+
{
|
3591 |
+
"epoch": 0.6534272901985907,
|
3592 |
+
"grad_norm": 1.4254034757614136,
|
3593 |
+
"learning_rate": 5.501896397994465e-05,
|
3594 |
+
"loss": 0.1308,
|
3595 |
+
"step": 510
|
3596 |
+
},
|
3597 |
+
{
|
3598 |
+
"epoch": 0.6547085201793722,
|
3599 |
+
"grad_norm": 2.644568920135498,
|
3600 |
+
"learning_rate": 5.465541724866787e-05,
|
3601 |
+
"loss": 0.2074,
|
3602 |
+
"step": 511
|
3603 |
+
},
|
3604 |
+
{
|
3605 |
+
"epoch": 0.6559897501601537,
|
3606 |
+
"grad_norm": 1.413726568222046,
|
3607 |
+
"learning_rate": 5.429262338012144e-05,
|
3608 |
+
"loss": 0.0418,
|
3609 |
+
"step": 512
|
3610 |
+
},
|
3611 |
+
{
|
3612 |
+
"epoch": 0.6572709801409353,
|
3613 |
+
"grad_norm": 1.1527552604675293,
|
3614 |
+
"learning_rate": 5.393058839782492e-05,
|
3615 |
+
"loss": 0.0518,
|
3616 |
+
"step": 513
|
3617 |
+
},
|
3618 |
+
{
|
3619 |
+
"epoch": 0.6585522101217168,
|
3620 |
+
"grad_norm": 1.0279854536056519,
|
3621 |
+
"learning_rate": 5.356931831269798e-05,
|
3622 |
+
"loss": 0.0889,
|
3623 |
+
"step": 514
|
3624 |
+
},
|
3625 |
+
{
|
3626 |
+
"epoch": 0.6598334401024984,
|
3627 |
+
"grad_norm": 0.05637316778302193,
|
3628 |
+
"learning_rate": 5.320881912296067e-05,
|
3629 |
+
"loss": 0.002,
|
3630 |
+
"step": 515
|
3631 |
+
},
|
3632 |
+
{
|
3633 |
+
"epoch": 0.66111467008328,
|
3634 |
+
"grad_norm": 0.8346595168113708,
|
3635 |
+
"learning_rate": 5.284909681403367e-05,
|
3636 |
+
"loss": 0.0887,
|
3637 |
+
"step": 516
|
3638 |
+
},
|
3639 |
+
{
|
3640 |
+
"epoch": 0.6623959000640615,
|
3641 |
+
"grad_norm": 1.888337254524231,
|
3642 |
+
"learning_rate": 5.249015735843901e-05,
|
3643 |
+
"loss": 0.1611,
|
3644 |
+
"step": 517
|
3645 |
+
},
|
3646 |
+
{
|
3647 |
+
"epoch": 0.6636771300448431,
|
3648 |
+
"grad_norm": 1.0611954927444458,
|
3649 |
+
"learning_rate": 5.213200671570085e-05,
|
3650 |
+
"loss": 0.1979,
|
3651 |
+
"step": 518
|
3652 |
+
},
|
3653 |
+
{
|
3654 |
+
"epoch": 0.6649583600256246,
|
3655 |
+
"grad_norm": 1.1892603635787964,
|
3656 |
+
"learning_rate": 5.177465083224671e-05,
|
3657 |
+
"loss": 0.1216,
|
3658 |
+
"step": 519
|
3659 |
+
},
|
3660 |
+
{
|
3661 |
+
"epoch": 0.6662395900064062,
|
3662 |
+
"grad_norm": 0.4522724151611328,
|
3663 |
+
"learning_rate": 5.141809564130847e-05,
|
3664 |
+
"loss": 0.0147,
|
3665 |
+
"step": 520
|
3666 |
+
},
|
3667 |
+
{
|
3668 |
+
"epoch": 0.6675208199871877,
|
3669 |
+
"grad_norm": 0.7256630063056946,
|
3670 |
+
"learning_rate": 5.106234706282404e-05,
|
3671 |
+
"loss": 0.0219,
|
3672 |
+
"step": 521
|
3673 |
+
},
|
3674 |
+
{
|
3675 |
+
"epoch": 0.6688020499679692,
|
3676 |
+
"grad_norm": 0.587709367275238,
|
3677 |
+
"learning_rate": 5.070741100333901e-05,
|
3678 |
+
"loss": 0.0493,
|
3679 |
+
"step": 522
|
3680 |
+
},
|
3681 |
+
{
|
3682 |
+
"epoch": 0.6700832799487508,
|
3683 |
+
"grad_norm": 2.2745747566223145,
|
3684 |
+
"learning_rate": 5.035329335590868e-05,
|
3685 |
+
"loss": 0.4237,
|
3686 |
+
"step": 523
|
3687 |
+
},
|
3688 |
+
{
|
3689 |
+
"epoch": 0.6713645099295323,
|
3690 |
+
"grad_norm": 0.1348501443862915,
|
3691 |
+
"learning_rate": 5.000000000000002e-05,
|
3692 |
+
"loss": 0.0056,
|
3693 |
+
"step": 524
|
3694 |
+
},
|
3695 |
+
{
|
3696 |
+
"epoch": 0.672645739910314,
|
3697 |
+
"grad_norm": 1.2014498710632324,
|
3698 |
+
"learning_rate": 4.964753680139424e-05,
|
3699 |
+
"loss": 0.0352,
|
3700 |
+
"step": 525
|
3701 |
+
},
|
3702 |
+
{
|
3703 |
+
"epoch": 0.6739269698910955,
|
3704 |
+
"grad_norm": 0.5173647999763489,
|
3705 |
+
"learning_rate": 4.9295909612089265e-05,
|
3706 |
+
"loss": 0.0428,
|
3707 |
+
"step": 526
|
3708 |
+
},
|
3709 |
+
{
|
3710 |
+
"epoch": 0.675208199871877,
|
3711 |
+
"grad_norm": 1.0500999689102173,
|
3712 |
+
"learning_rate": 4.8945124270202743e-05,
|
3713 |
+
"loss": 0.1163,
|
3714 |
+
"step": 527
|
3715 |
+
},
|
3716 |
+
{
|
3717 |
+
"epoch": 0.6764894298526586,
|
3718 |
+
"grad_norm": 1.5263786315917969,
|
3719 |
+
"learning_rate": 4.859518659987491e-05,
|
3720 |
+
"loss": 0.149,
|
3721 |
+
"step": 528
|
3722 |
+
},
|
3723 |
+
{
|
3724 |
+
"epoch": 0.6777706598334401,
|
3725 |
+
"grad_norm": 0.07213965058326721,
|
3726 |
+
"learning_rate": 4.8246102411171976e-05,
|
3727 |
+
"loss": 0.0034,
|
3728 |
+
"step": 529
|
3729 |
+
},
|
3730 |
+
{
|
3731 |
+
"epoch": 0.6790518898142216,
|
3732 |
+
"grad_norm": 1.7902854681015015,
|
3733 |
+
"learning_rate": 4.7897877499989784e-05,
|
3734 |
+
"loss": 0.2018,
|
3735 |
+
"step": 530
|
3736 |
+
},
|
3737 |
+
{
|
3738 |
+
"epoch": 0.6803331197950032,
|
3739 |
+
"grad_norm": 0.6578352451324463,
|
3740 |
+
"learning_rate": 4.7550517647957373e-05,
|
3741 |
+
"loss": 0.0634,
|
3742 |
+
"step": 531
|
3743 |
+
},
|
3744 |
+
{
|
3745 |
+
"epoch": 0.6816143497757847,
|
3746 |
+
"grad_norm": 0.6184600591659546,
|
3747 |
+
"learning_rate": 4.7204028622341047e-05,
|
3748 |
+
"loss": 0.0483,
|
3749 |
+
"step": 532
|
3750 |
+
},
|
3751 |
+
{
|
3752 |
+
"epoch": 0.6828955797565663,
|
3753 |
+
"grad_norm": 0.6419950127601624,
|
3754 |
+
"learning_rate": 4.6858416175948796e-05,
|
3755 |
+
"loss": 0.0317,
|
3756 |
+
"step": 533
|
3757 |
+
},
|
3758 |
+
{
|
3759 |
+
"epoch": 0.6841768097373478,
|
3760 |
+
"grad_norm": 0.3431781530380249,
|
3761 |
+
"learning_rate": 4.651368604703449e-05,
|
3762 |
+
"loss": 0.0176,
|
3763 |
+
"step": 534
|
3764 |
+
},
|
3765 |
+
{
|
3766 |
+
"epoch": 0.6854580397181294,
|
3767 |
+
"grad_norm": 0.1813570261001587,
|
3768 |
+
"learning_rate": 4.616984395920282e-05,
|
3769 |
+
"loss": 0.0105,
|
3770 |
+
"step": 535
|
3771 |
+
},
|
3772 |
+
{
|
3773 |
+
"epoch": 0.686739269698911,
|
3774 |
+
"grad_norm": 0.4483226239681244,
|
3775 |
+
"learning_rate": 4.582689562131416e-05,
|
3776 |
+
"loss": 0.0483,
|
3777 |
+
"step": 536
|
3778 |
+
},
|
3779 |
+
{
|
3780 |
+
"epoch": 0.6880204996796925,
|
3781 |
+
"grad_norm": 0.5256419777870178,
|
3782 |
+
"learning_rate": 4.548484672738992e-05,
|
3783 |
+
"loss": 0.1125,
|
3784 |
+
"step": 537
|
3785 |
+
},
|
3786 |
+
{
|
3787 |
+
"epoch": 0.6893017296604741,
|
3788 |
+
"grad_norm": 1.3066880702972412,
|
3789 |
+
"learning_rate": 4.514370295651781e-05,
|
3790 |
+
"loss": 0.0532,
|
3791 |
+
"step": 538
|
3792 |
+
},
|
3793 |
+
{
|
3794 |
+
"epoch": 0.6905829596412556,
|
3795 |
+
"grad_norm": 1.7708029747009277,
|
3796 |
+
"learning_rate": 4.4803469972757676e-05,
|
3797 |
+
"loss": 0.2584,
|
3798 |
+
"step": 539
|
3799 |
+
},
|
3800 |
+
{
|
3801 |
+
"epoch": 0.6918641896220371,
|
3802 |
+
"grad_norm": 0.3985656201839447,
|
3803 |
+
"learning_rate": 4.446415342504742e-05,
|
3804 |
+
"loss": 0.0379,
|
3805 |
+
"step": 540
|
3806 |
+
},
|
3807 |
+
{
|
3808 |
+
"epoch": 0.6931454196028187,
|
3809 |
+
"grad_norm": 2.026365041732788,
|
3810 |
+
"learning_rate": 4.412575894710931e-05,
|
3811 |
+
"loss": 0.7599,
|
3812 |
+
"step": 541
|
3813 |
+
},
|
3814 |
+
{
|
3815 |
+
"epoch": 0.6944266495836002,
|
3816 |
+
"grad_norm": 3.7336649894714355,
|
3817 |
+
"learning_rate": 4.378829215735624e-05,
|
3818 |
+
"loss": 1.6296,
|
3819 |
+
"step": 542
|
3820 |
+
},
|
3821 |
+
{
|
3822 |
+
"epoch": 0.6957078795643818,
|
3823 |
+
"grad_norm": 4.778042316436768,
|
3824 |
+
"learning_rate": 4.3451758658798626e-05,
|
3825 |
+
"loss": 1.7327,
|
3826 |
+
"step": 543
|
3827 |
+
},
|
3828 |
+
{
|
3829 |
+
"epoch": 0.6969891095451634,
|
3830 |
+
"grad_norm": 3.444413900375366,
|
3831 |
+
"learning_rate": 4.311616403895126e-05,
|
3832 |
+
"loss": 1.0805,
|
3833 |
+
"step": 544
|
3834 |
+
},
|
3835 |
+
{
|
3836 |
+
"epoch": 0.6982703395259449,
|
3837 |
+
"grad_norm": 3.895218849182129,
|
3838 |
+
"learning_rate": 4.27815138697407e-05,
|
3839 |
+
"loss": 1.2967,
|
3840 |
+
"step": 545
|
3841 |
+
},
|
3842 |
+
{
|
3843 |
+
"epoch": 0.6995515695067265,
|
3844 |
+
"grad_norm": 3.487854242324829,
|
3845 |
+
"learning_rate": 4.244781370741254e-05,
|
3846 |
+
"loss": 1.2053,
|
3847 |
+
"step": 546
|
3848 |
+
},
|
3849 |
+
{
|
3850 |
+
"epoch": 0.700832799487508,
|
3851 |
+
"grad_norm": 3.7566707134246826,
|
3852 |
+
"learning_rate": 4.2115069092439316e-05,
|
3853 |
+
"loss": 0.9895,
|
3854 |
+
"step": 547
|
3855 |
+
},
|
3856 |
+
{
|
3857 |
+
"epoch": 0.7021140294682896,
|
3858 |
+
"grad_norm": 4.939494609832764,
|
3859 |
+
"learning_rate": 4.17832855494285e-05,
|
3860 |
+
"loss": 1.6784,
|
3861 |
+
"step": 548
|
3862 |
+
},
|
3863 |
+
{
|
3864 |
+
"epoch": 0.7033952594490711,
|
3865 |
+
"grad_norm": 3.646867036819458,
|
3866 |
+
"learning_rate": 4.145246858703064e-05,
|
3867 |
+
"loss": 1.6463,
|
3868 |
+
"step": 549
|
3869 |
+
},
|
3870 |
+
{
|
3871 |
+
"epoch": 0.7046764894298526,
|
3872 |
+
"grad_norm": 4.433988571166992,
|
3873 |
+
"learning_rate": 4.112262369784816e-05,
|
3874 |
+
"loss": 1.6189,
|
3875 |
+
"step": 550
|
3876 |
+
},
|
3877 |
+
{
|
3878 |
+
"epoch": 0.7059577194106342,
|
3879 |
+
"grad_norm": 1.8268520832061768,
|
3880 |
+
"learning_rate": 4.079375635834392e-05,
|
3881 |
+
"loss": 0.211,
|
3882 |
+
"step": 551
|
3883 |
+
},
|
3884 |
+
{
|
3885 |
+
"epoch": 0.7072389493914157,
|
3886 |
+
"grad_norm": 1.1956140995025635,
|
3887 |
+
"learning_rate": 4.046587202875036e-05,
|
3888 |
+
"loss": 0.1915,
|
3889 |
+
"step": 552
|
3890 |
+
},
|
3891 |
+
{
|
3892 |
+
"epoch": 0.7085201793721974,
|
3893 |
+
"grad_norm": 3.8450396060943604,
|
3894 |
+
"learning_rate": 4.013897615297889e-05,
|
3895 |
+
"loss": 0.3661,
|
3896 |
+
"step": 553
|
3897 |
+
},
|
3898 |
+
{
|
3899 |
+
"epoch": 0.7098014093529789,
|
3900 |
+
"grad_norm": 1.0664808750152588,
|
3901 |
+
"learning_rate": 3.9813074158529496e-05,
|
3902 |
+
"loss": 0.0645,
|
3903 |
+
"step": 554
|
3904 |
+
},
|
3905 |
+
{
|
3906 |
+
"epoch": 0.7110826393337604,
|
3907 |
+
"grad_norm": 1.5560762882232666,
|
3908 |
+
"learning_rate": 3.9488171456400623e-05,
|
3909 |
+
"loss": 0.1012,
|
3910 |
+
"step": 555
|
3911 |
+
},
|
3912 |
+
{
|
3913 |
+
"epoch": 0.712363869314542,
|
3914 |
+
"grad_norm": 1.0144875049591064,
|
3915 |
+
"learning_rate": 3.9164273440999274e-05,
|
3916 |
+
"loss": 0.0609,
|
3917 |
+
"step": 556
|
3918 |
+
},
|
3919 |
+
{
|
3920 |
+
"epoch": 0.7136450992953235,
|
3921 |
+
"grad_norm": 4.624766826629639,
|
3922 |
+
"learning_rate": 3.8841385490051484e-05,
|
3923 |
+
"loss": 0.4452,
|
3924 |
+
"step": 557
|
3925 |
+
},
|
3926 |
+
{
|
3927 |
+
"epoch": 0.7149263292761051,
|
3928 |
+
"grad_norm": 1.2203162908554077,
|
3929 |
+
"learning_rate": 3.851951296451305e-05,
|
3930 |
+
"loss": 0.1105,
|
3931 |
+
"step": 558
|
3932 |
+
},
|
3933 |
+
{
|
3934 |
+
"epoch": 0.7162075592568866,
|
3935 |
+
"grad_norm": 1.3631420135498047,
|
3936 |
+
"learning_rate": 3.819866120848058e-05,
|
3937 |
+
"loss": 0.1389,
|
3938 |
+
"step": 559
|
3939 |
+
},
|
3940 |
+
{
|
3941 |
+
"epoch": 0.7174887892376681,
|
3942 |
+
"grad_norm": 1.411089301109314,
|
3943 |
+
"learning_rate": 3.787883554910262e-05,
|
3944 |
+
"loss": 0.1298,
|
3945 |
+
"step": 560
|
3946 |
+
},
|
3947 |
+
{
|
3948 |
+
"epoch": 0.7187700192184497,
|
3949 |
+
"grad_norm": 0.6210871338844299,
|
3950 |
+
"learning_rate": 3.756004129649129e-05,
|
3951 |
+
"loss": 0.0592,
|
3952 |
+
"step": 561
|
3953 |
+
},
|
3954 |
+
{
|
3955 |
+
"epoch": 0.7200512491992312,
|
3956 |
+
"grad_norm": 1.5564156770706177,
|
3957 |
+
"learning_rate": 3.72422837436341e-05,
|
3958 |
+
"loss": 0.2118,
|
3959 |
+
"step": 562
|
3960 |
+
},
|
3961 |
+
{
|
3962 |
+
"epoch": 0.7213324791800129,
|
3963 |
+
"grad_norm": 4.404601573944092,
|
3964 |
+
"learning_rate": 3.692556816630618e-05,
|
3965 |
+
"loss": 0.7355,
|
3966 |
+
"step": 563
|
3967 |
+
},
|
3968 |
+
{
|
3969 |
+
"epoch": 0.7226137091607944,
|
3970 |
+
"grad_norm": 1.367163896560669,
|
3971 |
+
"learning_rate": 3.6609899822982484e-05,
|
3972 |
+
"loss": 0.2019,
|
3973 |
+
"step": 564
|
3974 |
+
},
|
3975 |
+
{
|
3976 |
+
"epoch": 0.7238949391415759,
|
3977 |
+
"grad_norm": 1.4442821741104126,
|
3978 |
+
"learning_rate": 3.629528395475063e-05,
|
3979 |
+
"loss": 0.1558,
|
3980 |
+
"step": 565
|
3981 |
+
},
|
3982 |
+
{
|
3983 |
+
"epoch": 0.7251761691223575,
|
3984 |
+
"grad_norm": 0.461955726146698,
|
3985 |
+
"learning_rate": 3.598172578522382e-05,
|
3986 |
+
"loss": 0.0276,
|
3987 |
+
"step": 566
|
3988 |
+
},
|
3989 |
+
{
|
3990 |
+
"epoch": 0.726457399103139,
|
3991 |
+
"grad_norm": 0.4186933636665344,
|
3992 |
+
"learning_rate": 3.566923052045411e-05,
|
3993 |
+
"loss": 0.0265,
|
3994 |
+
"step": 567
|
3995 |
+
},
|
3996 |
+
{
|
3997 |
+
"epoch": 0.7277386290839206,
|
3998 |
+
"grad_norm": 4.97804594039917,
|
3999 |
+
"learning_rate": 3.5357803348846085e-05,
|
4000 |
+
"loss": 0.1257,
|
4001 |
+
"step": 568
|
4002 |
+
},
|
4003 |
+
{
|
4004 |
+
"epoch": 0.7290198590647021,
|
4005 |
+
"grad_norm": 1.054079294204712,
|
4006 |
+
"learning_rate": 3.504744944107055e-05,
|
4007 |
+
"loss": 0.0947,
|
4008 |
+
"step": 569
|
4009 |
+
},
|
4010 |
+
{
|
4011 |
+
"epoch": 0.7303010890454836,
|
4012 |
+
"grad_norm": 1.1791518926620483,
|
4013 |
+
"learning_rate": 3.4738173949978735e-05,
|
4014 |
+
"loss": 0.1131,
|
4015 |
+
"step": 570
|
4016 |
+
},
|
4017 |
+
{
|
4018 |
+
"epoch": 0.7315823190262652,
|
4019 |
+
"grad_norm": 1.2824045419692993,
|
4020 |
+
"learning_rate": 3.442998201051677e-05,
|
4021 |
+
"loss": 0.1187,
|
4022 |
+
"step": 571
|
4023 |
+
},
|
4024 |
+
{
|
4025 |
+
"epoch": 0.7328635490070468,
|
4026 |
+
"grad_norm": 0.708232045173645,
|
4027 |
+
"learning_rate": 3.412287873964052e-05,
|
4028 |
+
"loss": 0.0667,
|
4029 |
+
"step": 572
|
4030 |
+
},
|
4031 |
+
{
|
4032 |
+
"epoch": 0.7341447789878283,
|
4033 |
+
"grad_norm": 0.8040647506713867,
|
4034 |
+
"learning_rate": 3.381686923623039e-05,
|
4035 |
+
"loss": 0.1298,
|
4036 |
+
"step": 573
|
4037 |
+
},
|
4038 |
+
{
|
4039 |
+
"epoch": 0.7354260089686099,
|
4040 |
+
"grad_norm": 0.4562820792198181,
|
4041 |
+
"learning_rate": 3.3511958581006874e-05,
|
4042 |
+
"loss": 0.0395,
|
4043 |
+
"step": 574
|
4044 |
+
},
|
4045 |
+
{
|
4046 |
+
"epoch": 0.7367072389493914,
|
4047 |
+
"grad_norm": 0.4617690443992615,
|
4048 |
+
"learning_rate": 3.3208151836446077e-05,
|
4049 |
+
"loss": 0.0645,
|
4050 |
+
"step": 575
|
4051 |
+
},
|
4052 |
+
{
|
4053 |
+
"epoch": 0.737988468930173,
|
4054 |
+
"grad_norm": 0.399964302778244,
|
4055 |
+
"learning_rate": 3.2905454046695815e-05,
|
4056 |
+
"loss": 0.0365,
|
4057 |
+
"step": 576
|
4058 |
+
},
|
4059 |
+
{
|
4060 |
+
"epoch": 0.7392696989109545,
|
4061 |
+
"grad_norm": 1.6209018230438232,
|
4062 |
+
"learning_rate": 3.260387023749165e-05,
|
4063 |
+
"loss": 0.2314,
|
4064 |
+
"step": 577
|
4065 |
+
},
|
4066 |
+
{
|
4067 |
+
"epoch": 0.740550928891736,
|
4068 |
+
"grad_norm": 0.6959272027015686,
|
4069 |
+
"learning_rate": 3.2303405416073677e-05,
|
4070 |
+
"loss": 0.0853,
|
4071 |
+
"step": 578
|
4072 |
+
},
|
4073 |
+
{
|
4074 |
+
"epoch": 0.7418321588725176,
|
4075 |
+
"grad_norm": 4.2295918464660645,
|
4076 |
+
"learning_rate": 3.200406457110315e-05,
|
4077 |
+
"loss": 0.3678,
|
4078 |
+
"step": 579
|
4079 |
+
},
|
4080 |
+
{
|
4081 |
+
"epoch": 0.7431133888532991,
|
4082 |
+
"grad_norm": 0.4365205466747284,
|
4083 |
+
"learning_rate": 3.170585267257985e-05,
|
4084 |
+
"loss": 0.03,
|
4085 |
+
"step": 580
|
4086 |
+
},
|
4087 |
+
{
|
4088 |
+
"epoch": 0.7443946188340808,
|
4089 |
+
"grad_norm": 0.24114122986793518,
|
4090 |
+
"learning_rate": 3.140877467175952e-05,
|
4091 |
+
"loss": 0.0169,
|
4092 |
+
"step": 581
|
4093 |
+
},
|
4094 |
+
{
|
4095 |
+
"epoch": 0.7456758488148623,
|
4096 |
+
"grad_norm": 0.39172670245170593,
|
4097 |
+
"learning_rate": 3.111283550107158e-05,
|
4098 |
+
"loss": 0.044,
|
4099 |
+
"step": 582
|
4100 |
+
},
|
4101 |
+
{
|
4102 |
+
"epoch": 0.7469570787956438,
|
4103 |
+
"grad_norm": 2.391191005706787,
|
4104 |
+
"learning_rate": 3.0818040074037305e-05,
|
4105 |
+
"loss": 0.1675,
|
4106 |
+
"step": 583
|
4107 |
+
},
|
4108 |
+
{
|
4109 |
+
"epoch": 0.7482383087764254,
|
4110 |
+
"grad_norm": 2.9679298400878906,
|
4111 |
+
"learning_rate": 3.052439328518817e-05,
|
4112 |
+
"loss": 0.2899,
|
4113 |
+
"step": 584
|
4114 |
+
},
|
4115 |
+
{
|
4116 |
+
"epoch": 0.7495195387572069,
|
4117 |
+
"grad_norm": 0.3954543471336365,
|
4118 |
+
"learning_rate": 3.023190000998479e-05,
|
4119 |
+
"loss": 0.0257,
|
4120 |
+
"step": 585
|
4121 |
+
},
|
4122 |
+
{
|
4123 |
+
"epoch": 0.7508007687379885,
|
4124 |
+
"grad_norm": 0.30377396941185,
|
4125 |
+
"learning_rate": 2.994056510473571e-05,
|
4126 |
+
"loss": 0.0208,
|
4127 |
+
"step": 586
|
4128 |
+
},
|
4129 |
+
{
|
4130 |
+
"epoch": 0.75208199871877,
|
4131 |
+
"grad_norm": 0.24724654853343964,
|
4132 |
+
"learning_rate": 2.9650393406516886e-05,
|
4133 |
+
"loss": 0.0146,
|
4134 |
+
"step": 587
|
4135 |
+
},
|
4136 |
+
{
|
4137 |
+
"epoch": 0.7533632286995515,
|
4138 |
+
"grad_norm": 0.5047999620437622,
|
4139 |
+
"learning_rate": 2.9361389733091417e-05,
|
4140 |
+
"loss": 0.0644,
|
4141 |
+
"step": 588
|
4142 |
+
},
|
4143 |
+
{
|
4144 |
+
"epoch": 0.7533632286995515,
|
4145 |
+
"eval_loss": 0.32855647802352905,
|
4146 |
+
"eval_runtime": 22.913,
|
4147 |
+
"eval_samples_per_second": 14.359,
|
4148 |
+
"eval_steps_per_second": 7.201,
|
4149 |
+
"step": 588
|
4150 |
}
|
4151 |
],
|
4152 |
"logging_steps": 1,
|
|
|
4166 |
"attributes": {}
|
4167 |
}
|
4168 |
},
|
4169 |
+
"total_flos": 2.175088739619963e+17,
|
4170 |
"train_batch_size": 2,
|
4171 |
"trial_name": null,
|
4172 |
"trial_params": null
|