Training in progress, step 500, 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 140815952
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ab7de5a4f71f4c55592d7dd0bfec683332e4beef3a399148fd0647f05b016669
|
| 3 |
size 140815952
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 71878996
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fe90ba59438995207d6b2fd90994253ff0224026840ec367200c2735aa5ca6ac
|
| 3 |
size 71878996
|
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:f152b160914c2ede02f6e50de068f39d624bdf7552f9a2b416ec64ec58043054
|
| 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:fe1d153de177b356f9e3a70d6e4ec979560b0c300994e71ca4cb89afc74c5b3a
|
| 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,
|
|
@@ -2847,6 +2847,714 @@
|
|
| 2847 |
"eval_samples_per_second": 16.209,
|
| 2848 |
"eval_steps_per_second": 4.054,
|
| 2849 |
"step": 400
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2850 |
}
|
| 2851 |
],
|
| 2852 |
"logging_steps": 1,
|
|
@@ -2870,12 +3578,12 @@
|
|
| 2870 |
"should_evaluate": false,
|
| 2871 |
"should_log": false,
|
| 2872 |
"should_save": true,
|
| 2873 |
-
"should_training_stop":
|
| 2874 |
},
|
| 2875 |
"attributes": {}
|
| 2876 |
}
|
| 2877 |
},
|
| 2878 |
-
"total_flos": 3.
|
| 2879 |
"train_batch_size": 8,
|
| 2880 |
"trial_name": null,
|
| 2881 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 0.5021965503692627,
|
| 3 |
+
"best_model_checkpoint": "miner_id_24/checkpoint-500",
|
| 4 |
+
"epoch": 0.091324200913242,
|
| 5 |
"eval_steps": 100,
|
| 6 |
+
"global_step": 500,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 2847 |
"eval_samples_per_second": 16.209,
|
| 2848 |
"eval_steps_per_second": 4.054,
|
| 2849 |
"step": 400
|
| 2850 |
+
},
|
| 2851 |
+
{
|
| 2852 |
+
"epoch": 0.07324200913242009,
|
| 2853 |
+
"grad_norm": 1.2463815212249756,
|
| 2854 |
+
"learning_rate": 1.013396731136465e-05,
|
| 2855 |
+
"loss": 0.6997,
|
| 2856 |
+
"step": 401
|
| 2857 |
+
},
|
| 2858 |
+
{
|
| 2859 |
+
"epoch": 0.07342465753424658,
|
| 2860 |
+
"grad_norm": 1.217960000038147,
|
| 2861 |
+
"learning_rate": 9.937309365446973e-06,
|
| 2862 |
+
"loss": 0.4999,
|
| 2863 |
+
"step": 402
|
| 2864 |
+
},
|
| 2865 |
+
{
|
| 2866 |
+
"epoch": 0.07360730593607306,
|
| 2867 |
+
"grad_norm": 1.2658041715621948,
|
| 2868 |
+
"learning_rate": 9.742367571857091e-06,
|
| 2869 |
+
"loss": 0.4848,
|
| 2870 |
+
"step": 403
|
| 2871 |
+
},
|
| 2872 |
+
{
|
| 2873 |
+
"epoch": 0.07378995433789955,
|
| 2874 |
+
"grad_norm": 1.161754846572876,
|
| 2875 |
+
"learning_rate": 9.549150281252633e-06,
|
| 2876 |
+
"loss": 0.3601,
|
| 2877 |
+
"step": 404
|
| 2878 |
+
},
|
| 2879 |
+
{
|
| 2880 |
+
"epoch": 0.07397260273972603,
|
| 2881 |
+
"grad_norm": 1.1665185689926147,
|
| 2882 |
+
"learning_rate": 9.357665770419244e-06,
|
| 2883 |
+
"loss": 0.4717,
|
| 2884 |
+
"step": 405
|
| 2885 |
+
},
|
| 2886 |
+
{
|
| 2887 |
+
"epoch": 0.07415525114155251,
|
| 2888 |
+
"grad_norm": 1.2439416646957397,
|
| 2889 |
+
"learning_rate": 9.167922241916055e-06,
|
| 2890 |
+
"loss": 0.4756,
|
| 2891 |
+
"step": 406
|
| 2892 |
+
},
|
| 2893 |
+
{
|
| 2894 |
+
"epoch": 0.074337899543379,
|
| 2895 |
+
"grad_norm": 1.0245717763900757,
|
| 2896 |
+
"learning_rate": 8.97992782372432e-06,
|
| 2897 |
+
"loss": 0.4582,
|
| 2898 |
+
"step": 407
|
| 2899 |
+
},
|
| 2900 |
+
{
|
| 2901 |
+
"epoch": 0.07452054794520548,
|
| 2902 |
+
"grad_norm": 1.089290738105774,
|
| 2903 |
+
"learning_rate": 8.793690568899216e-06,
|
| 2904 |
+
"loss": 0.5299,
|
| 2905 |
+
"step": 408
|
| 2906 |
+
},
|
| 2907 |
+
{
|
| 2908 |
+
"epoch": 0.07470319634703196,
|
| 2909 |
+
"grad_norm": 1.3025835752487183,
|
| 2910 |
+
"learning_rate": 8.609218455224893e-06,
|
| 2911 |
+
"loss": 0.5364,
|
| 2912 |
+
"step": 409
|
| 2913 |
+
},
|
| 2914 |
+
{
|
| 2915 |
+
"epoch": 0.07488584474885844,
|
| 2916 |
+
"grad_norm": 1.1520413160324097,
|
| 2917 |
+
"learning_rate": 8.426519384872733e-06,
|
| 2918 |
+
"loss": 0.4844,
|
| 2919 |
+
"step": 410
|
| 2920 |
+
},
|
| 2921 |
+
{
|
| 2922 |
+
"epoch": 0.07506849315068494,
|
| 2923 |
+
"grad_norm": 1.42751145362854,
|
| 2924 |
+
"learning_rate": 8.245601184062852e-06,
|
| 2925 |
+
"loss": 0.6758,
|
| 2926 |
+
"step": 411
|
| 2927 |
+
},
|
| 2928 |
+
{
|
| 2929 |
+
"epoch": 0.07525114155251142,
|
| 2930 |
+
"grad_norm": 1.2706706523895264,
|
| 2931 |
+
"learning_rate": 8.066471602728803e-06,
|
| 2932 |
+
"loss": 0.5892,
|
| 2933 |
+
"step": 412
|
| 2934 |
+
},
|
| 2935 |
+
{
|
| 2936 |
+
"epoch": 0.0754337899543379,
|
| 2937 |
+
"grad_norm": 1.106239676475525,
|
| 2938 |
+
"learning_rate": 7.889138314185678e-06,
|
| 2939 |
+
"loss": 0.6417,
|
| 2940 |
+
"step": 413
|
| 2941 |
+
},
|
| 2942 |
+
{
|
| 2943 |
+
"epoch": 0.07561643835616438,
|
| 2944 |
+
"grad_norm": 1.1466580629348755,
|
| 2945 |
+
"learning_rate": 7.71360891480134e-06,
|
| 2946 |
+
"loss": 0.5323,
|
| 2947 |
+
"step": 414
|
| 2948 |
+
},
|
| 2949 |
+
{
|
| 2950 |
+
"epoch": 0.07579908675799087,
|
| 2951 |
+
"grad_norm": 1.415773868560791,
|
| 2952 |
+
"learning_rate": 7.539890923671062e-06,
|
| 2953 |
+
"loss": 0.6715,
|
| 2954 |
+
"step": 415
|
| 2955 |
+
},
|
| 2956 |
+
{
|
| 2957 |
+
"epoch": 0.07598173515981735,
|
| 2958 |
+
"grad_norm": 1.2140964269638062,
|
| 2959 |
+
"learning_rate": 7.367991782295391e-06,
|
| 2960 |
+
"loss": 0.5337,
|
| 2961 |
+
"step": 416
|
| 2962 |
+
},
|
| 2963 |
+
{
|
| 2964 |
+
"epoch": 0.07616438356164383,
|
| 2965 |
+
"grad_norm": 1.058760643005371,
|
| 2966 |
+
"learning_rate": 7.197918854261432e-06,
|
| 2967 |
+
"loss": 0.4444,
|
| 2968 |
+
"step": 417
|
| 2969 |
+
},
|
| 2970 |
+
{
|
| 2971 |
+
"epoch": 0.07634703196347031,
|
| 2972 |
+
"grad_norm": 1.1686381101608276,
|
| 2973 |
+
"learning_rate": 7.029679424927365e-06,
|
| 2974 |
+
"loss": 0.5853,
|
| 2975 |
+
"step": 418
|
| 2976 |
+
},
|
| 2977 |
+
{
|
| 2978 |
+
"epoch": 0.07652968036529681,
|
| 2979 |
+
"grad_norm": 1.3012456893920898,
|
| 2980 |
+
"learning_rate": 6.863280701110408e-06,
|
| 2981 |
+
"loss": 0.6194,
|
| 2982 |
+
"step": 419
|
| 2983 |
+
},
|
| 2984 |
+
{
|
| 2985 |
+
"epoch": 0.07671232876712329,
|
| 2986 |
+
"grad_norm": 1.3436243534088135,
|
| 2987 |
+
"learning_rate": 6.698729810778065e-06,
|
| 2988 |
+
"loss": 0.5871,
|
| 2989 |
+
"step": 420
|
| 2990 |
+
},
|
| 2991 |
+
{
|
| 2992 |
+
"epoch": 0.07689497716894977,
|
| 2993 |
+
"grad_norm": 1.2779135704040527,
|
| 2994 |
+
"learning_rate": 6.536033802742813e-06,
|
| 2995 |
+
"loss": 0.6079,
|
| 2996 |
+
"step": 421
|
| 2997 |
+
},
|
| 2998 |
+
{
|
| 2999 |
+
"epoch": 0.07707762557077626,
|
| 3000 |
+
"grad_norm": 1.21416175365448,
|
| 3001 |
+
"learning_rate": 6.375199646360142e-06,
|
| 3002 |
+
"loss": 0.5959,
|
| 3003 |
+
"step": 422
|
| 3004 |
+
},
|
| 3005 |
+
{
|
| 3006 |
+
"epoch": 0.07726027397260274,
|
| 3007 |
+
"grad_norm": 1.3231496810913086,
|
| 3008 |
+
"learning_rate": 6.216234231230012e-06,
|
| 3009 |
+
"loss": 0.6456,
|
| 3010 |
+
"step": 423
|
| 3011 |
+
},
|
| 3012 |
+
{
|
| 3013 |
+
"epoch": 0.07744292237442922,
|
| 3014 |
+
"grad_norm": 1.204132318496704,
|
| 3015 |
+
"learning_rate": 6.059144366901736e-06,
|
| 3016 |
+
"loss": 0.5165,
|
| 3017 |
+
"step": 424
|
| 3018 |
+
},
|
| 3019 |
+
{
|
| 3020 |
+
"epoch": 0.0776255707762557,
|
| 3021 |
+
"grad_norm": 1.0210753679275513,
|
| 3022 |
+
"learning_rate": 5.903936782582253e-06,
|
| 3023 |
+
"loss": 0.5296,
|
| 3024 |
+
"step": 425
|
| 3025 |
+
},
|
| 3026 |
+
{
|
| 3027 |
+
"epoch": 0.07780821917808219,
|
| 3028 |
+
"grad_norm": 1.1843194961547852,
|
| 3029 |
+
"learning_rate": 5.750618126847912e-06,
|
| 3030 |
+
"loss": 0.5952,
|
| 3031 |
+
"step": 426
|
| 3032 |
+
},
|
| 3033 |
+
{
|
| 3034 |
+
"epoch": 0.07799086757990868,
|
| 3035 |
+
"grad_norm": 1.1271318197250366,
|
| 3036 |
+
"learning_rate": 5.599194967359639e-06,
|
| 3037 |
+
"loss": 0.4949,
|
| 3038 |
+
"step": 427
|
| 3039 |
+
},
|
| 3040 |
+
{
|
| 3041 |
+
"epoch": 0.07817351598173516,
|
| 3042 |
+
"grad_norm": 1.4363117218017578,
|
| 3043 |
+
"learning_rate": 5.449673790581611e-06,
|
| 3044 |
+
"loss": 0.604,
|
| 3045 |
+
"step": 428
|
| 3046 |
+
},
|
| 3047 |
+
{
|
| 3048 |
+
"epoch": 0.07835616438356165,
|
| 3049 |
+
"grad_norm": 1.2224055528640747,
|
| 3050 |
+
"learning_rate": 5.302061001503394e-06,
|
| 3051 |
+
"loss": 0.584,
|
| 3052 |
+
"step": 429
|
| 3053 |
+
},
|
| 3054 |
+
{
|
| 3055 |
+
"epoch": 0.07853881278538813,
|
| 3056 |
+
"grad_norm": 1.8298218250274658,
|
| 3057 |
+
"learning_rate": 5.156362923365588e-06,
|
| 3058 |
+
"loss": 0.6796,
|
| 3059 |
+
"step": 430
|
| 3060 |
+
},
|
| 3061 |
+
{
|
| 3062 |
+
"epoch": 0.07872146118721461,
|
| 3063 |
+
"grad_norm": 1.4991016387939453,
|
| 3064 |
+
"learning_rate": 5.012585797388936e-06,
|
| 3065 |
+
"loss": 0.6289,
|
| 3066 |
+
"step": 431
|
| 3067 |
+
},
|
| 3068 |
+
{
|
| 3069 |
+
"epoch": 0.0789041095890411,
|
| 3070 |
+
"grad_norm": 1.097782015800476,
|
| 3071 |
+
"learning_rate": 4.87073578250698e-06,
|
| 3072 |
+
"loss": 0.4951,
|
| 3073 |
+
"step": 432
|
| 3074 |
+
},
|
| 3075 |
+
{
|
| 3076 |
+
"epoch": 0.07908675799086758,
|
| 3077 |
+
"grad_norm": 1.333260416984558,
|
| 3078 |
+
"learning_rate": 4.730818955102234e-06,
|
| 3079 |
+
"loss": 0.6281,
|
| 3080 |
+
"step": 433
|
| 3081 |
+
},
|
| 3082 |
+
{
|
| 3083 |
+
"epoch": 0.07926940639269406,
|
| 3084 |
+
"grad_norm": 1.578722357749939,
|
| 3085 |
+
"learning_rate": 4.592841308745932e-06,
|
| 3086 |
+
"loss": 0.5824,
|
| 3087 |
+
"step": 434
|
| 3088 |
+
},
|
| 3089 |
+
{
|
| 3090 |
+
"epoch": 0.07945205479452055,
|
| 3091 |
+
"grad_norm": 0.9925230741500854,
|
| 3092 |
+
"learning_rate": 4.456808753941205e-06,
|
| 3093 |
+
"loss": 0.4832,
|
| 3094 |
+
"step": 435
|
| 3095 |
+
},
|
| 3096 |
+
{
|
| 3097 |
+
"epoch": 0.07963470319634704,
|
| 3098 |
+
"grad_norm": 1.4724770784378052,
|
| 3099 |
+
"learning_rate": 4.322727117869951e-06,
|
| 3100 |
+
"loss": 0.5059,
|
| 3101 |
+
"step": 436
|
| 3102 |
+
},
|
| 3103 |
+
{
|
| 3104 |
+
"epoch": 0.07981735159817352,
|
| 3105 |
+
"grad_norm": 1.2720314264297485,
|
| 3106 |
+
"learning_rate": 4.190602144143207e-06,
|
| 3107 |
+
"loss": 0.5338,
|
| 3108 |
+
"step": 437
|
| 3109 |
+
},
|
| 3110 |
+
{
|
| 3111 |
+
"epoch": 0.08,
|
| 3112 |
+
"grad_norm": 1.396422028541565,
|
| 3113 |
+
"learning_rate": 4.06043949255509e-06,
|
| 3114 |
+
"loss": 0.4048,
|
| 3115 |
+
"step": 438
|
| 3116 |
+
},
|
| 3117 |
+
{
|
| 3118 |
+
"epoch": 0.08018264840182648,
|
| 3119 |
+
"grad_norm": 1.325544834136963,
|
| 3120 |
+
"learning_rate": 3.932244738840379e-06,
|
| 3121 |
+
"loss": 0.4413,
|
| 3122 |
+
"step": 439
|
| 3123 |
+
},
|
| 3124 |
+
{
|
| 3125 |
+
"epoch": 0.08036529680365297,
|
| 3126 |
+
"grad_norm": 1.644364833831787,
|
| 3127 |
+
"learning_rate": 3.8060233744356633e-06,
|
| 3128 |
+
"loss": 0.4839,
|
| 3129 |
+
"step": 440
|
| 3130 |
+
},
|
| 3131 |
+
{
|
| 3132 |
+
"epoch": 0.08054794520547945,
|
| 3133 |
+
"grad_norm": 1.1722346544265747,
|
| 3134 |
+
"learning_rate": 3.681780806244095e-06,
|
| 3135 |
+
"loss": 0.3067,
|
| 3136 |
+
"step": 441
|
| 3137 |
+
},
|
| 3138 |
+
{
|
| 3139 |
+
"epoch": 0.08073059360730593,
|
| 3140 |
+
"grad_norm": 1.310457706451416,
|
| 3141 |
+
"learning_rate": 3.5595223564037884e-06,
|
| 3142 |
+
"loss": 0.4886,
|
| 3143 |
+
"step": 442
|
| 3144 |
+
},
|
| 3145 |
+
{
|
| 3146 |
+
"epoch": 0.08091324200913241,
|
| 3147 |
+
"grad_norm": 1.7274094820022583,
|
| 3148 |
+
"learning_rate": 3.4392532620598216e-06,
|
| 3149 |
+
"loss": 0.484,
|
| 3150 |
+
"step": 443
|
| 3151 |
+
},
|
| 3152 |
+
{
|
| 3153 |
+
"epoch": 0.08109589041095891,
|
| 3154 |
+
"grad_norm": 1.5255731344223022,
|
| 3155 |
+
"learning_rate": 3.3209786751399187e-06,
|
| 3156 |
+
"loss": 0.3585,
|
| 3157 |
+
"step": 444
|
| 3158 |
+
},
|
| 3159 |
+
{
|
| 3160 |
+
"epoch": 0.08127853881278539,
|
| 3161 |
+
"grad_norm": 1.4306080341339111,
|
| 3162 |
+
"learning_rate": 3.2047036621337236e-06,
|
| 3163 |
+
"loss": 0.4051,
|
| 3164 |
+
"step": 445
|
| 3165 |
+
},
|
| 3166 |
+
{
|
| 3167 |
+
"epoch": 0.08146118721461187,
|
| 3168 |
+
"grad_norm": 1.2907383441925049,
|
| 3169 |
+
"learning_rate": 3.0904332038757977e-06,
|
| 3170 |
+
"loss": 0.2441,
|
| 3171 |
+
"step": 446
|
| 3172 |
+
},
|
| 3173 |
+
{
|
| 3174 |
+
"epoch": 0.08164383561643836,
|
| 3175 |
+
"grad_norm": 1.5992954969406128,
|
| 3176 |
+
"learning_rate": 2.978172195332263e-06,
|
| 3177 |
+
"loss": 0.4898,
|
| 3178 |
+
"step": 447
|
| 3179 |
+
},
|
| 3180 |
+
{
|
| 3181 |
+
"epoch": 0.08182648401826484,
|
| 3182 |
+
"grad_norm": 1.7128164768218994,
|
| 3183 |
+
"learning_rate": 2.8679254453910785e-06,
|
| 3184 |
+
"loss": 0.5083,
|
| 3185 |
+
"step": 448
|
| 3186 |
+
},
|
| 3187 |
+
{
|
| 3188 |
+
"epoch": 0.08200913242009132,
|
| 3189 |
+
"grad_norm": 1.7184181213378906,
|
| 3190 |
+
"learning_rate": 2.759697676656098e-06,
|
| 3191 |
+
"loss": 0.382,
|
| 3192 |
+
"step": 449
|
| 3193 |
+
},
|
| 3194 |
+
{
|
| 3195 |
+
"epoch": 0.0821917808219178,
|
| 3196 |
+
"grad_norm": 2.048771858215332,
|
| 3197 |
+
"learning_rate": 2.653493525244721e-06,
|
| 3198 |
+
"loss": 0.3999,
|
| 3199 |
+
"step": 450
|
| 3200 |
+
},
|
| 3201 |
+
{
|
| 3202 |
+
"epoch": 0.08237442922374429,
|
| 3203 |
+
"grad_norm": 1.1284714937210083,
|
| 3204 |
+
"learning_rate": 2.549317540589308e-06,
|
| 3205 |
+
"loss": 0.5719,
|
| 3206 |
+
"step": 451
|
| 3207 |
+
},
|
| 3208 |
+
{
|
| 3209 |
+
"epoch": 0.08255707762557078,
|
| 3210 |
+
"grad_norm": 1.186042308807373,
|
| 3211 |
+
"learning_rate": 2.4471741852423237e-06,
|
| 3212 |
+
"loss": 0.5015,
|
| 3213 |
+
"step": 452
|
| 3214 |
+
},
|
| 3215 |
+
{
|
| 3216 |
+
"epoch": 0.08273972602739726,
|
| 3217 |
+
"grad_norm": 1.1713098287582397,
|
| 3218 |
+
"learning_rate": 2.3470678346851518e-06,
|
| 3219 |
+
"loss": 0.4088,
|
| 3220 |
+
"step": 453
|
| 3221 |
+
},
|
| 3222 |
+
{
|
| 3223 |
+
"epoch": 0.08292237442922375,
|
| 3224 |
+
"grad_norm": 1.1227015256881714,
|
| 3225 |
+
"learning_rate": 2.2490027771406687e-06,
|
| 3226 |
+
"loss": 0.4691,
|
| 3227 |
+
"step": 454
|
| 3228 |
+
},
|
| 3229 |
+
{
|
| 3230 |
+
"epoch": 0.08310502283105023,
|
| 3231 |
+
"grad_norm": 1.1555074453353882,
|
| 3232 |
+
"learning_rate": 2.152983213389559e-06,
|
| 3233 |
+
"loss": 0.4275,
|
| 3234 |
+
"step": 455
|
| 3235 |
+
},
|
| 3236 |
+
{
|
| 3237 |
+
"epoch": 0.08328767123287671,
|
| 3238 |
+
"grad_norm": 1.7054256200790405,
|
| 3239 |
+
"learning_rate": 2.0590132565903476e-06,
|
| 3240 |
+
"loss": 0.4563,
|
| 3241 |
+
"step": 456
|
| 3242 |
+
},
|
| 3243 |
+
{
|
| 3244 |
+
"epoch": 0.0834703196347032,
|
| 3245 |
+
"grad_norm": 1.2316186428070068,
|
| 3246 |
+
"learning_rate": 1.9670969321032407e-06,
|
| 3247 |
+
"loss": 0.5705,
|
| 3248 |
+
"step": 457
|
| 3249 |
+
},
|
| 3250 |
+
{
|
| 3251 |
+
"epoch": 0.08365296803652968,
|
| 3252 |
+
"grad_norm": 1.1939345598220825,
|
| 3253 |
+
"learning_rate": 1.8772381773176417e-06,
|
| 3254 |
+
"loss": 0.5377,
|
| 3255 |
+
"step": 458
|
| 3256 |
+
},
|
| 3257 |
+
{
|
| 3258 |
+
"epoch": 0.08383561643835616,
|
| 3259 |
+
"grad_norm": 1.361952543258667,
|
| 3260 |
+
"learning_rate": 1.7894408414835362e-06,
|
| 3261 |
+
"loss": 0.599,
|
| 3262 |
+
"step": 459
|
| 3263 |
+
},
|
| 3264 |
+
{
|
| 3265 |
+
"epoch": 0.08401826484018265,
|
| 3266 |
+
"grad_norm": 1.2690627574920654,
|
| 3267 |
+
"learning_rate": 1.70370868554659e-06,
|
| 3268 |
+
"loss": 0.4931,
|
| 3269 |
+
"step": 460
|
| 3270 |
+
},
|
| 3271 |
+
{
|
| 3272 |
+
"epoch": 0.08420091324200914,
|
| 3273 |
+
"grad_norm": 1.0819649696350098,
|
| 3274 |
+
"learning_rate": 1.620045381987012e-06,
|
| 3275 |
+
"loss": 0.4172,
|
| 3276 |
+
"step": 461
|
| 3277 |
+
},
|
| 3278 |
+
{
|
| 3279 |
+
"epoch": 0.08438356164383562,
|
| 3280 |
+
"grad_norm": 1.047044038772583,
|
| 3281 |
+
"learning_rate": 1.5384545146622852e-06,
|
| 3282 |
+
"loss": 0.468,
|
| 3283 |
+
"step": 462
|
| 3284 |
+
},
|
| 3285 |
+
{
|
| 3286 |
+
"epoch": 0.0845662100456621,
|
| 3287 |
+
"grad_norm": 1.3348311185836792,
|
| 3288 |
+
"learning_rate": 1.4589395786535953e-06,
|
| 3289 |
+
"loss": 0.5959,
|
| 3290 |
+
"step": 463
|
| 3291 |
+
},
|
| 3292 |
+
{
|
| 3293 |
+
"epoch": 0.08474885844748858,
|
| 3294 |
+
"grad_norm": 1.244368076324463,
|
| 3295 |
+
"learning_rate": 1.3815039801161721e-06,
|
| 3296 |
+
"loss": 0.6091,
|
| 3297 |
+
"step": 464
|
| 3298 |
+
},
|
| 3299 |
+
{
|
| 3300 |
+
"epoch": 0.08493150684931507,
|
| 3301 |
+
"grad_norm": 1.2115455865859985,
|
| 3302 |
+
"learning_rate": 1.3061510361333185e-06,
|
| 3303 |
+
"loss": 0.5653,
|
| 3304 |
+
"step": 465
|
| 3305 |
+
},
|
| 3306 |
+
{
|
| 3307 |
+
"epoch": 0.08511415525114155,
|
| 3308 |
+
"grad_norm": 0.9911224842071533,
|
| 3309 |
+
"learning_rate": 1.232883974574367e-06,
|
| 3310 |
+
"loss": 0.422,
|
| 3311 |
+
"step": 466
|
| 3312 |
+
},
|
| 3313 |
+
{
|
| 3314 |
+
"epoch": 0.08529680365296803,
|
| 3315 |
+
"grad_norm": 1.102059006690979,
|
| 3316 |
+
"learning_rate": 1.1617059339563807e-06,
|
| 3317 |
+
"loss": 0.529,
|
| 3318 |
+
"step": 467
|
| 3319 |
+
},
|
| 3320 |
+
{
|
| 3321 |
+
"epoch": 0.08547945205479453,
|
| 3322 |
+
"grad_norm": 1.138898491859436,
|
| 3323 |
+
"learning_rate": 1.0926199633097157e-06,
|
| 3324 |
+
"loss": 0.5913,
|
| 3325 |
+
"step": 468
|
| 3326 |
+
},
|
| 3327 |
+
{
|
| 3328 |
+
"epoch": 0.08566210045662101,
|
| 3329 |
+
"grad_norm": 1.2904703617095947,
|
| 3330 |
+
"learning_rate": 1.0256290220474307e-06,
|
| 3331 |
+
"loss": 0.6278,
|
| 3332 |
+
"step": 469
|
| 3333 |
+
},
|
| 3334 |
+
{
|
| 3335 |
+
"epoch": 0.08584474885844749,
|
| 3336 |
+
"grad_norm": 0.8520990610122681,
|
| 3337 |
+
"learning_rate": 9.607359798384785e-07,
|
| 3338 |
+
"loss": 0.401,
|
| 3339 |
+
"step": 470
|
| 3340 |
+
},
|
| 3341 |
+
{
|
| 3342 |
+
"epoch": 0.08602739726027397,
|
| 3343 |
+
"grad_norm": 1.2092121839523315,
|
| 3344 |
+
"learning_rate": 8.979436164848088e-07,
|
| 3345 |
+
"loss": 0.5386,
|
| 3346 |
+
"step": 471
|
| 3347 |
+
},
|
| 3348 |
+
{
|
| 3349 |
+
"epoch": 0.08621004566210046,
|
| 3350 |
+
"grad_norm": 1.0581966638565063,
|
| 3351 |
+
"learning_rate": 8.372546218022747e-07,
|
| 3352 |
+
"loss": 0.4301,
|
| 3353 |
+
"step": 472
|
| 3354 |
+
},
|
| 3355 |
+
{
|
| 3356 |
+
"epoch": 0.08639269406392694,
|
| 3357 |
+
"grad_norm": 1.1138097047805786,
|
| 3358 |
+
"learning_rate": 7.786715955054203e-07,
|
| 3359 |
+
"loss": 0.552,
|
| 3360 |
+
"step": 473
|
| 3361 |
+
},
|
| 3362 |
+
{
|
| 3363 |
+
"epoch": 0.08657534246575342,
|
| 3364 |
+
"grad_norm": 1.2832486629486084,
|
| 3365 |
+
"learning_rate": 7.221970470961125e-07,
|
| 3366 |
+
"loss": 0.544,
|
| 3367 |
+
"step": 474
|
| 3368 |
+
},
|
| 3369 |
+
{
|
| 3370 |
+
"epoch": 0.0867579908675799,
|
| 3371 |
+
"grad_norm": 1.1421263217926025,
|
| 3372 |
+
"learning_rate": 6.678333957560512e-07,
|
| 3373 |
+
"loss": 0.4701,
|
| 3374 |
+
"step": 475
|
| 3375 |
+
},
|
| 3376 |
+
{
|
| 3377 |
+
"epoch": 0.08694063926940639,
|
| 3378 |
+
"grad_norm": 1.7712730169296265,
|
| 3379 |
+
"learning_rate": 6.15582970243117e-07,
|
| 3380 |
+
"loss": 0.4951,
|
| 3381 |
+
"step": 476
|
| 3382 |
+
},
|
| 3383 |
+
{
|
| 3384 |
+
"epoch": 0.08712328767123288,
|
| 3385 |
+
"grad_norm": 1.3618603944778442,
|
| 3386 |
+
"learning_rate": 5.654480087916303e-07,
|
| 3387 |
+
"loss": 0.7162,
|
| 3388 |
+
"step": 477
|
| 3389 |
+
},
|
| 3390 |
+
{
|
| 3391 |
+
"epoch": 0.08730593607305936,
|
| 3392 |
+
"grad_norm": 1.3067758083343506,
|
| 3393 |
+
"learning_rate": 5.174306590164879e-07,
|
| 3394 |
+
"loss": 0.5436,
|
| 3395 |
+
"step": 478
|
| 3396 |
+
},
|
| 3397 |
+
{
|
| 3398 |
+
"epoch": 0.08748858447488585,
|
| 3399 |
+
"grad_norm": 1.1023305654525757,
|
| 3400 |
+
"learning_rate": 4.715329778211375e-07,
|
| 3401 |
+
"loss": 0.5493,
|
| 3402 |
+
"step": 479
|
| 3403 |
+
},
|
| 3404 |
+
{
|
| 3405 |
+
"epoch": 0.08767123287671233,
|
| 3406 |
+
"grad_norm": 1.245082974433899,
|
| 3407 |
+
"learning_rate": 4.277569313094809e-07,
|
| 3408 |
+
"loss": 0.604,
|
| 3409 |
+
"step": 480
|
| 3410 |
+
},
|
| 3411 |
+
{
|
| 3412 |
+
"epoch": 0.08785388127853881,
|
| 3413 |
+
"grad_norm": 1.2294963598251343,
|
| 3414 |
+
"learning_rate": 3.8610439470164737e-07,
|
| 3415 |
+
"loss": 0.5481,
|
| 3416 |
+
"step": 481
|
| 3417 |
+
},
|
| 3418 |
+
{
|
| 3419 |
+
"epoch": 0.0880365296803653,
|
| 3420 |
+
"grad_norm": 1.3997677564620972,
|
| 3421 |
+
"learning_rate": 3.465771522536854e-07,
|
| 3422 |
+
"loss": 0.4414,
|
| 3423 |
+
"step": 482
|
| 3424 |
+
},
|
| 3425 |
+
{
|
| 3426 |
+
"epoch": 0.08821917808219178,
|
| 3427 |
+
"grad_norm": 1.1400816440582275,
|
| 3428 |
+
"learning_rate": 3.09176897181096e-07,
|
| 3429 |
+
"loss": 0.4481,
|
| 3430 |
+
"step": 483
|
| 3431 |
+
},
|
| 3432 |
+
{
|
| 3433 |
+
"epoch": 0.08840182648401826,
|
| 3434 |
+
"grad_norm": 1.1948268413543701,
|
| 3435 |
+
"learning_rate": 2.7390523158633554e-07,
|
| 3436 |
+
"loss": 0.456,
|
| 3437 |
+
"step": 484
|
| 3438 |
+
},
|
| 3439 |
+
{
|
| 3440 |
+
"epoch": 0.08858447488584476,
|
| 3441 |
+
"grad_norm": 1.2014074325561523,
|
| 3442 |
+
"learning_rate": 2.407636663901591e-07,
|
| 3443 |
+
"loss": 0.4554,
|
| 3444 |
+
"step": 485
|
| 3445 |
+
},
|
| 3446 |
+
{
|
| 3447 |
+
"epoch": 0.08876712328767124,
|
| 3448 |
+
"grad_norm": 1.2776867151260376,
|
| 3449 |
+
"learning_rate": 2.0975362126691712e-07,
|
| 3450 |
+
"loss": 0.4105,
|
| 3451 |
+
"step": 486
|
| 3452 |
+
},
|
| 3453 |
+
{
|
| 3454 |
+
"epoch": 0.08894977168949772,
|
| 3455 |
+
"grad_norm": 1.259320616722107,
|
| 3456 |
+
"learning_rate": 1.8087642458373134e-07,
|
| 3457 |
+
"loss": 0.4406,
|
| 3458 |
+
"step": 487
|
| 3459 |
+
},
|
| 3460 |
+
{
|
| 3461 |
+
"epoch": 0.0891324200913242,
|
| 3462 |
+
"grad_norm": 1.1561946868896484,
|
| 3463 |
+
"learning_rate": 1.5413331334360182e-07,
|
| 3464 |
+
"loss": 0.4482,
|
| 3465 |
+
"step": 488
|
| 3466 |
+
},
|
| 3467 |
+
{
|
| 3468 |
+
"epoch": 0.08931506849315068,
|
| 3469 |
+
"grad_norm": 1.4489840269088745,
|
| 3470 |
+
"learning_rate": 1.2952543313240472e-07,
|
| 3471 |
+
"loss": 0.4719,
|
| 3472 |
+
"step": 489
|
| 3473 |
+
},
|
| 3474 |
+
{
|
| 3475 |
+
"epoch": 0.08949771689497717,
|
| 3476 |
+
"grad_norm": 1.1808651685714722,
|
| 3477 |
+
"learning_rate": 1.0705383806982606e-07,
|
| 3478 |
+
"loss": 0.337,
|
| 3479 |
+
"step": 490
|
| 3480 |
+
},
|
| 3481 |
+
{
|
| 3482 |
+
"epoch": 0.08968036529680365,
|
| 3483 |
+
"grad_norm": 1.3887622356414795,
|
| 3484 |
+
"learning_rate": 8.671949076420882e-08,
|
| 3485 |
+
"loss": 0.4287,
|
| 3486 |
+
"step": 491
|
| 3487 |
+
},
|
| 3488 |
+
{
|
| 3489 |
+
"epoch": 0.08986301369863013,
|
| 3490 |
+
"grad_norm": 1.3295773267745972,
|
| 3491 |
+
"learning_rate": 6.852326227130834e-08,
|
| 3492 |
+
"loss": 0.4468,
|
| 3493 |
+
"step": 492
|
| 3494 |
+
},
|
| 3495 |
+
{
|
| 3496 |
+
"epoch": 0.09004566210045663,
|
| 3497 |
+
"grad_norm": 1.7511096000671387,
|
| 3498 |
+
"learning_rate": 5.246593205699424e-08,
|
| 3499 |
+
"loss": 0.4839,
|
| 3500 |
+
"step": 493
|
| 3501 |
+
},
|
| 3502 |
+
{
|
| 3503 |
+
"epoch": 0.09022831050228311,
|
| 3504 |
+
"grad_norm": 1.6349002122879028,
|
| 3505 |
+
"learning_rate": 3.8548187963854956e-08,
|
| 3506 |
+
"loss": 0.3704,
|
| 3507 |
+
"step": 494
|
| 3508 |
+
},
|
| 3509 |
+
{
|
| 3510 |
+
"epoch": 0.09041095890410959,
|
| 3511 |
+
"grad_norm": 1.719805121421814,
|
| 3512 |
+
"learning_rate": 2.6770626181715773e-08,
|
| 3513 |
+
"loss": 0.4897,
|
| 3514 |
+
"step": 495
|
| 3515 |
+
},
|
| 3516 |
+
{
|
| 3517 |
+
"epoch": 0.09059360730593607,
|
| 3518 |
+
"grad_norm": 1.7145909070968628,
|
| 3519 |
+
"learning_rate": 1.7133751222137007e-08,
|
| 3520 |
+
"loss": 0.3541,
|
| 3521 |
+
"step": 496
|
| 3522 |
+
},
|
| 3523 |
+
{
|
| 3524 |
+
"epoch": 0.09077625570776256,
|
| 3525 |
+
"grad_norm": 1.6819974184036255,
|
| 3526 |
+
"learning_rate": 9.637975896759077e-09,
|
| 3527 |
+
"loss": 0.3164,
|
| 3528 |
+
"step": 497
|
| 3529 |
+
},
|
| 3530 |
+
{
|
| 3531 |
+
"epoch": 0.09095890410958904,
|
| 3532 |
+
"grad_norm": 2.2097480297088623,
|
| 3533 |
+
"learning_rate": 4.2836212996499865e-09,
|
| 3534 |
+
"loss": 0.3985,
|
| 3535 |
+
"step": 498
|
| 3536 |
+
},
|
| 3537 |
+
{
|
| 3538 |
+
"epoch": 0.09114155251141552,
|
| 3539 |
+
"grad_norm": 1.636614441871643,
|
| 3540 |
+
"learning_rate": 1.0709167935385455e-09,
|
| 3541 |
+
"loss": 0.3184,
|
| 3542 |
+
"step": 499
|
| 3543 |
+
},
|
| 3544 |
+
{
|
| 3545 |
+
"epoch": 0.091324200913242,
|
| 3546 |
+
"grad_norm": 2.1838996410369873,
|
| 3547 |
+
"learning_rate": 0.0,
|
| 3548 |
+
"loss": 0.4692,
|
| 3549 |
+
"step": 500
|
| 3550 |
+
},
|
| 3551 |
+
{
|
| 3552 |
+
"epoch": 0.091324200913242,
|
| 3553 |
+
"eval_loss": 0.5021965503692627,
|
| 3554 |
+
"eval_runtime": 574.0387,
|
| 3555 |
+
"eval_samples_per_second": 16.063,
|
| 3556 |
+
"eval_steps_per_second": 4.017,
|
| 3557 |
+
"step": 500
|
| 3558 |
}
|
| 3559 |
],
|
| 3560 |
"logging_steps": 1,
|
|
|
|
| 3578 |
"should_evaluate": false,
|
| 3579 |
"should_log": false,
|
| 3580 |
"should_save": true,
|
| 3581 |
+
"should_training_stop": true
|
| 3582 |
},
|
| 3583 |
"attributes": {}
|
| 3584 |
}
|
| 3585 |
},
|
| 3586 |
+
"total_flos": 3.8642412158976e+16,
|
| 3587 |
"train_batch_size": 8,
|
| 3588 |
"trial_name": null,
|
| 3589 |
"trial_params": null
|