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 42487072
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:37d4cb892a10d02b24d61aedce704b107108385a83eafaf27cbf0a6a9a16b90c
|
3 |
size 42487072
|
last-checkpoint/optimizer.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 21735482
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0a50d5abcb2bc5c3cbf54bf226ede6fec91eb42ad81549176ce88135755b2186
|
3 |
size 21735482
|
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:f29f580bbba93af94f51d1cdb432af843d87cd4198761212d6136820b3ee42f3
|
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": 3.
|
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": 173.697,
|
2848 |
"eval_steps_per_second": 43.435,
|
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":
|
2879 |
"train_batch_size": 8,
|
2880 |
"trial_name": null,
|
2881 |
"trial_params": null
|
|
|
1 |
{
|
2 |
+
"best_metric": 3.130511999130249,
|
3 |
+
"best_model_checkpoint": "miner_id_24/checkpoint-500",
|
4 |
+
"epoch": 0.10231749117511639,
|
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": 173.697,
|
2848 |
"eval_steps_per_second": 43.435,
|
2849 |
"step": 400
|
2850 |
+
},
|
2851 |
+
{
|
2852 |
+
"epoch": 0.08205862792244334,
|
2853 |
+
"grad_norm": 2.140819549560547,
|
2854 |
+
"learning_rate": 1.013396731136465e-05,
|
2855 |
+
"loss": 12.6548,
|
2856 |
+
"step": 401
|
2857 |
+
},
|
2858 |
+
{
|
2859 |
+
"epoch": 0.08226326290479358,
|
2860 |
+
"grad_norm": 2.7390999794006348,
|
2861 |
+
"learning_rate": 9.937309365446973e-06,
|
2862 |
+
"loss": 12.7752,
|
2863 |
+
"step": 402
|
2864 |
+
},
|
2865 |
+
{
|
2866 |
+
"epoch": 0.0824678978871438,
|
2867 |
+
"grad_norm": 2.9735829830169678,
|
2868 |
+
"learning_rate": 9.742367571857091e-06,
|
2869 |
+
"loss": 13.2948,
|
2870 |
+
"step": 403
|
2871 |
+
},
|
2872 |
+
{
|
2873 |
+
"epoch": 0.08267253286949404,
|
2874 |
+
"grad_norm": 2.9806323051452637,
|
2875 |
+
"learning_rate": 9.549150281252633e-06,
|
2876 |
+
"loss": 12.9468,
|
2877 |
+
"step": 404
|
2878 |
+
},
|
2879 |
+
{
|
2880 |
+
"epoch": 0.08287716785184428,
|
2881 |
+
"grad_norm": 2.9686119556427,
|
2882 |
+
"learning_rate": 9.357665770419244e-06,
|
2883 |
+
"loss": 12.8879,
|
2884 |
+
"step": 405
|
2885 |
+
},
|
2886 |
+
{
|
2887 |
+
"epoch": 0.0830818028341945,
|
2888 |
+
"grad_norm": 3.127723217010498,
|
2889 |
+
"learning_rate": 9.167922241916055e-06,
|
2890 |
+
"loss": 12.6892,
|
2891 |
+
"step": 406
|
2892 |
+
},
|
2893 |
+
{
|
2894 |
+
"epoch": 0.08328643781654474,
|
2895 |
+
"grad_norm": 3.3917040824890137,
|
2896 |
+
"learning_rate": 8.97992782372432e-06,
|
2897 |
+
"loss": 13.2438,
|
2898 |
+
"step": 407
|
2899 |
+
},
|
2900 |
+
{
|
2901 |
+
"epoch": 0.08349107279889498,
|
2902 |
+
"grad_norm": 3.28285551071167,
|
2903 |
+
"learning_rate": 8.793690568899216e-06,
|
2904 |
+
"loss": 12.7453,
|
2905 |
+
"step": 408
|
2906 |
+
},
|
2907 |
+
{
|
2908 |
+
"epoch": 0.0836957077812452,
|
2909 |
+
"grad_norm": 3.364295482635498,
|
2910 |
+
"learning_rate": 8.609218455224893e-06,
|
2911 |
+
"loss": 13.0133,
|
2912 |
+
"step": 409
|
2913 |
+
},
|
2914 |
+
{
|
2915 |
+
"epoch": 0.08390034276359544,
|
2916 |
+
"grad_norm": 3.374210834503174,
|
2917 |
+
"learning_rate": 8.426519384872733e-06,
|
2918 |
+
"loss": 12.79,
|
2919 |
+
"step": 410
|
2920 |
+
},
|
2921 |
+
{
|
2922 |
+
"epoch": 0.08410497774594566,
|
2923 |
+
"grad_norm": 3.6572799682617188,
|
2924 |
+
"learning_rate": 8.245601184062852e-06,
|
2925 |
+
"loss": 12.6433,
|
2926 |
+
"step": 411
|
2927 |
+
},
|
2928 |
+
{
|
2929 |
+
"epoch": 0.0843096127282959,
|
2930 |
+
"grad_norm": 3.5379762649536133,
|
2931 |
+
"learning_rate": 8.066471602728803e-06,
|
2932 |
+
"loss": 12.6151,
|
2933 |
+
"step": 412
|
2934 |
+
},
|
2935 |
+
{
|
2936 |
+
"epoch": 0.08451424771064614,
|
2937 |
+
"grad_norm": 3.6328377723693848,
|
2938 |
+
"learning_rate": 7.889138314185678e-06,
|
2939 |
+
"loss": 13.0676,
|
2940 |
+
"step": 413
|
2941 |
+
},
|
2942 |
+
{
|
2943 |
+
"epoch": 0.08471888269299636,
|
2944 |
+
"grad_norm": 4.000665664672852,
|
2945 |
+
"learning_rate": 7.71360891480134e-06,
|
2946 |
+
"loss": 13.2594,
|
2947 |
+
"step": 414
|
2948 |
+
},
|
2949 |
+
{
|
2950 |
+
"epoch": 0.0849235176753466,
|
2951 |
+
"grad_norm": 4.031978130340576,
|
2952 |
+
"learning_rate": 7.539890923671062e-06,
|
2953 |
+
"loss": 12.738,
|
2954 |
+
"step": 415
|
2955 |
+
},
|
2956 |
+
{
|
2957 |
+
"epoch": 0.08512815265769684,
|
2958 |
+
"grad_norm": 3.724813461303711,
|
2959 |
+
"learning_rate": 7.367991782295391e-06,
|
2960 |
+
"loss": 12.2267,
|
2961 |
+
"step": 416
|
2962 |
+
},
|
2963 |
+
{
|
2964 |
+
"epoch": 0.08533278764004706,
|
2965 |
+
"grad_norm": 4.113027095794678,
|
2966 |
+
"learning_rate": 7.197918854261432e-06,
|
2967 |
+
"loss": 12.7985,
|
2968 |
+
"step": 417
|
2969 |
+
},
|
2970 |
+
{
|
2971 |
+
"epoch": 0.0855374226223973,
|
2972 |
+
"grad_norm": 4.147072792053223,
|
2973 |
+
"learning_rate": 7.029679424927365e-06,
|
2974 |
+
"loss": 13.3496,
|
2975 |
+
"step": 418
|
2976 |
+
},
|
2977 |
+
{
|
2978 |
+
"epoch": 0.08574205760474753,
|
2979 |
+
"grad_norm": 4.0234527587890625,
|
2980 |
+
"learning_rate": 6.863280701110408e-06,
|
2981 |
+
"loss": 13.023,
|
2982 |
+
"step": 419
|
2983 |
+
},
|
2984 |
+
{
|
2985 |
+
"epoch": 0.08594669258709776,
|
2986 |
+
"grad_norm": 4.267332077026367,
|
2987 |
+
"learning_rate": 6.698729810778065e-06,
|
2988 |
+
"loss": 13.0938,
|
2989 |
+
"step": 420
|
2990 |
+
},
|
2991 |
+
{
|
2992 |
+
"epoch": 0.086151327569448,
|
2993 |
+
"grad_norm": 4.35993766784668,
|
2994 |
+
"learning_rate": 6.536033802742813e-06,
|
2995 |
+
"loss": 12.7928,
|
2996 |
+
"step": 421
|
2997 |
+
},
|
2998 |
+
{
|
2999 |
+
"epoch": 0.08635596255179823,
|
3000 |
+
"grad_norm": 4.47703218460083,
|
3001 |
+
"learning_rate": 6.375199646360142e-06,
|
3002 |
+
"loss": 12.9274,
|
3003 |
+
"step": 422
|
3004 |
+
},
|
3005 |
+
{
|
3006 |
+
"epoch": 0.08656059753414846,
|
3007 |
+
"grad_norm": 4.716027736663818,
|
3008 |
+
"learning_rate": 6.216234231230012e-06,
|
3009 |
+
"loss": 13.2614,
|
3010 |
+
"step": 423
|
3011 |
+
},
|
3012 |
+
{
|
3013 |
+
"epoch": 0.08676523251649869,
|
3014 |
+
"grad_norm": 4.96610164642334,
|
3015 |
+
"learning_rate": 6.059144366901736e-06,
|
3016 |
+
"loss": 12.2165,
|
3017 |
+
"step": 424
|
3018 |
+
},
|
3019 |
+
{
|
3020 |
+
"epoch": 0.08696986749884893,
|
3021 |
+
"grad_norm": 4.611530780792236,
|
3022 |
+
"learning_rate": 5.903936782582253e-06,
|
3023 |
+
"loss": 12.7755,
|
3024 |
+
"step": 425
|
3025 |
+
},
|
3026 |
+
{
|
3027 |
+
"epoch": 0.08717450248119916,
|
3028 |
+
"grad_norm": 4.9565300941467285,
|
3029 |
+
"learning_rate": 5.750618126847912e-06,
|
3030 |
+
"loss": 13.0633,
|
3031 |
+
"step": 426
|
3032 |
+
},
|
3033 |
+
{
|
3034 |
+
"epoch": 0.08737913746354939,
|
3035 |
+
"grad_norm": 5.047351837158203,
|
3036 |
+
"learning_rate": 5.599194967359639e-06,
|
3037 |
+
"loss": 12.7034,
|
3038 |
+
"step": 427
|
3039 |
+
},
|
3040 |
+
{
|
3041 |
+
"epoch": 0.08758377244589963,
|
3042 |
+
"grad_norm": 4.904860019683838,
|
3043 |
+
"learning_rate": 5.449673790581611e-06,
|
3044 |
+
"loss": 12.5122,
|
3045 |
+
"step": 428
|
3046 |
+
},
|
3047 |
+
{
|
3048 |
+
"epoch": 0.08778840742824986,
|
3049 |
+
"grad_norm": 5.229506015777588,
|
3050 |
+
"learning_rate": 5.302061001503394e-06,
|
3051 |
+
"loss": 13.0031,
|
3052 |
+
"step": 429
|
3053 |
+
},
|
3054 |
+
{
|
3055 |
+
"epoch": 0.08799304241060009,
|
3056 |
+
"grad_norm": 5.458662509918213,
|
3057 |
+
"learning_rate": 5.156362923365588e-06,
|
3058 |
+
"loss": 12.8912,
|
3059 |
+
"step": 430
|
3060 |
+
},
|
3061 |
+
{
|
3062 |
+
"epoch": 0.08819767739295033,
|
3063 |
+
"grad_norm": 5.457494735717773,
|
3064 |
+
"learning_rate": 5.012585797388936e-06,
|
3065 |
+
"loss": 13.3934,
|
3066 |
+
"step": 431
|
3067 |
+
},
|
3068 |
+
{
|
3069 |
+
"epoch": 0.08840231237530055,
|
3070 |
+
"grad_norm": 5.259174823760986,
|
3071 |
+
"learning_rate": 4.87073578250698e-06,
|
3072 |
+
"loss": 12.2529,
|
3073 |
+
"step": 432
|
3074 |
+
},
|
3075 |
+
{
|
3076 |
+
"epoch": 0.08860694735765079,
|
3077 |
+
"grad_norm": 5.968120574951172,
|
3078 |
+
"learning_rate": 4.730818955102234e-06,
|
3079 |
+
"loss": 12.54,
|
3080 |
+
"step": 433
|
3081 |
+
},
|
3082 |
+
{
|
3083 |
+
"epoch": 0.08881158234000103,
|
3084 |
+
"grad_norm": 6.257950782775879,
|
3085 |
+
"learning_rate": 4.592841308745932e-06,
|
3086 |
+
"loss": 12.6848,
|
3087 |
+
"step": 434
|
3088 |
+
},
|
3089 |
+
{
|
3090 |
+
"epoch": 0.08901621732235125,
|
3091 |
+
"grad_norm": 6.009523868560791,
|
3092 |
+
"learning_rate": 4.456808753941205e-06,
|
3093 |
+
"loss": 13.0703,
|
3094 |
+
"step": 435
|
3095 |
+
},
|
3096 |
+
{
|
3097 |
+
"epoch": 0.08922085230470149,
|
3098 |
+
"grad_norm": 6.272739887237549,
|
3099 |
+
"learning_rate": 4.322727117869951e-06,
|
3100 |
+
"loss": 12.3788,
|
3101 |
+
"step": 436
|
3102 |
+
},
|
3103 |
+
{
|
3104 |
+
"epoch": 0.08942548728705173,
|
3105 |
+
"grad_norm": 6.675440788269043,
|
3106 |
+
"learning_rate": 4.190602144143207e-06,
|
3107 |
+
"loss": 13.0829,
|
3108 |
+
"step": 437
|
3109 |
+
},
|
3110 |
+
{
|
3111 |
+
"epoch": 0.08963012226940195,
|
3112 |
+
"grad_norm": 6.91643762588501,
|
3113 |
+
"learning_rate": 4.06043949255509e-06,
|
3114 |
+
"loss": 13.3085,
|
3115 |
+
"step": 438
|
3116 |
+
},
|
3117 |
+
{
|
3118 |
+
"epoch": 0.08983475725175219,
|
3119 |
+
"grad_norm": 7.155692100524902,
|
3120 |
+
"learning_rate": 3.932244738840379e-06,
|
3121 |
+
"loss": 12.7579,
|
3122 |
+
"step": 439
|
3123 |
+
},
|
3124 |
+
{
|
3125 |
+
"epoch": 0.09003939223410241,
|
3126 |
+
"grad_norm": 6.8719940185546875,
|
3127 |
+
"learning_rate": 3.8060233744356633e-06,
|
3128 |
+
"loss": 12.96,
|
3129 |
+
"step": 440
|
3130 |
+
},
|
3131 |
+
{
|
3132 |
+
"epoch": 0.09024402721645265,
|
3133 |
+
"grad_norm": 7.757195949554443,
|
3134 |
+
"learning_rate": 3.681780806244095e-06,
|
3135 |
+
"loss": 12.8733,
|
3136 |
+
"step": 441
|
3137 |
+
},
|
3138 |
+
{
|
3139 |
+
"epoch": 0.09044866219880289,
|
3140 |
+
"grad_norm": 7.0664215087890625,
|
3141 |
+
"learning_rate": 3.5595223564037884e-06,
|
3142 |
+
"loss": 12.6288,
|
3143 |
+
"step": 442
|
3144 |
+
},
|
3145 |
+
{
|
3146 |
+
"epoch": 0.09065329718115311,
|
3147 |
+
"grad_norm": 8.440971374511719,
|
3148 |
+
"learning_rate": 3.4392532620598216e-06,
|
3149 |
+
"loss": 13.3932,
|
3150 |
+
"step": 443
|
3151 |
+
},
|
3152 |
+
{
|
3153 |
+
"epoch": 0.09085793216350335,
|
3154 |
+
"grad_norm": 8.004168510437012,
|
3155 |
+
"learning_rate": 3.3209786751399187e-06,
|
3156 |
+
"loss": 12.6128,
|
3157 |
+
"step": 444
|
3158 |
+
},
|
3159 |
+
{
|
3160 |
+
"epoch": 0.09106256714585359,
|
3161 |
+
"grad_norm": 8.411291122436523,
|
3162 |
+
"learning_rate": 3.2047036621337236e-06,
|
3163 |
+
"loss": 13.4134,
|
3164 |
+
"step": 445
|
3165 |
+
},
|
3166 |
+
{
|
3167 |
+
"epoch": 0.09126720212820381,
|
3168 |
+
"grad_norm": 9.257599830627441,
|
3169 |
+
"learning_rate": 3.0904332038757977e-06,
|
3170 |
+
"loss": 12.8954,
|
3171 |
+
"step": 446
|
3172 |
+
},
|
3173 |
+
{
|
3174 |
+
"epoch": 0.09147183711055405,
|
3175 |
+
"grad_norm": 9.609465599060059,
|
3176 |
+
"learning_rate": 2.978172195332263e-06,
|
3177 |
+
"loss": 13.3849,
|
3178 |
+
"step": 447
|
3179 |
+
},
|
3180 |
+
{
|
3181 |
+
"epoch": 0.09167647209290428,
|
3182 |
+
"grad_norm": 11.43130111694336,
|
3183 |
+
"learning_rate": 2.8679254453910785e-06,
|
3184 |
+
"loss": 12.0155,
|
3185 |
+
"step": 448
|
3186 |
+
},
|
3187 |
+
{
|
3188 |
+
"epoch": 0.09188110707525451,
|
3189 |
+
"grad_norm": 12.757843971252441,
|
3190 |
+
"learning_rate": 2.759697676656098e-06,
|
3191 |
+
"loss": 12.183,
|
3192 |
+
"step": 449
|
3193 |
+
},
|
3194 |
+
{
|
3195 |
+
"epoch": 0.09208574205760475,
|
3196 |
+
"grad_norm": 17.703031539916992,
|
3197 |
+
"learning_rate": 2.653493525244721e-06,
|
3198 |
+
"loss": 13.8012,
|
3199 |
+
"step": 450
|
3200 |
+
},
|
3201 |
+
{
|
3202 |
+
"epoch": 0.09229037703995498,
|
3203 |
+
"grad_norm": 2.267075300216675,
|
3204 |
+
"learning_rate": 2.549317540589308e-06,
|
3205 |
+
"loss": 12.4868,
|
3206 |
+
"step": 451
|
3207 |
+
},
|
3208 |
+
{
|
3209 |
+
"epoch": 0.09249501202230521,
|
3210 |
+
"grad_norm": 2.5427181720733643,
|
3211 |
+
"learning_rate": 2.4471741852423237e-06,
|
3212 |
+
"loss": 12.5128,
|
3213 |
+
"step": 452
|
3214 |
+
},
|
3215 |
+
{
|
3216 |
+
"epoch": 0.09269964700465545,
|
3217 |
+
"grad_norm": 2.599229097366333,
|
3218 |
+
"learning_rate": 2.3470678346851518e-06,
|
3219 |
+
"loss": 12.9201,
|
3220 |
+
"step": 453
|
3221 |
+
},
|
3222 |
+
{
|
3223 |
+
"epoch": 0.09290428198700568,
|
3224 |
+
"grad_norm": 2.6054906845092773,
|
3225 |
+
"learning_rate": 2.2490027771406687e-06,
|
3226 |
+
"loss": 12.6195,
|
3227 |
+
"step": 454
|
3228 |
+
},
|
3229 |
+
{
|
3230 |
+
"epoch": 0.09310891696935591,
|
3231 |
+
"grad_norm": 2.9035027027130127,
|
3232 |
+
"learning_rate": 2.152983213389559e-06,
|
3233 |
+
"loss": 12.7726,
|
3234 |
+
"step": 455
|
3235 |
+
},
|
3236 |
+
{
|
3237 |
+
"epoch": 0.09331355195170614,
|
3238 |
+
"grad_norm": 3.060668468475342,
|
3239 |
+
"learning_rate": 2.0590132565903476e-06,
|
3240 |
+
"loss": 13.0542,
|
3241 |
+
"step": 456
|
3242 |
+
},
|
3243 |
+
{
|
3244 |
+
"epoch": 0.09351818693405638,
|
3245 |
+
"grad_norm": 3.1342105865478516,
|
3246 |
+
"learning_rate": 1.9670969321032407e-06,
|
3247 |
+
"loss": 12.9032,
|
3248 |
+
"step": 457
|
3249 |
+
},
|
3250 |
+
{
|
3251 |
+
"epoch": 0.09372282191640662,
|
3252 |
+
"grad_norm": 3.196115493774414,
|
3253 |
+
"learning_rate": 1.8772381773176417e-06,
|
3254 |
+
"loss": 13.1882,
|
3255 |
+
"step": 458
|
3256 |
+
},
|
3257 |
+
{
|
3258 |
+
"epoch": 0.09392745689875684,
|
3259 |
+
"grad_norm": 3.232515335083008,
|
3260 |
+
"learning_rate": 1.7894408414835362e-06,
|
3261 |
+
"loss": 12.8594,
|
3262 |
+
"step": 459
|
3263 |
+
},
|
3264 |
+
{
|
3265 |
+
"epoch": 0.09413209188110708,
|
3266 |
+
"grad_norm": 3.358297348022461,
|
3267 |
+
"learning_rate": 1.70370868554659e-06,
|
3268 |
+
"loss": 12.909,
|
3269 |
+
"step": 460
|
3270 |
+
},
|
3271 |
+
{
|
3272 |
+
"epoch": 0.0943367268634573,
|
3273 |
+
"grad_norm": 3.49599027633667,
|
3274 |
+
"learning_rate": 1.620045381987012e-06,
|
3275 |
+
"loss": 12.7092,
|
3276 |
+
"step": 461
|
3277 |
+
},
|
3278 |
+
{
|
3279 |
+
"epoch": 0.09454136184580754,
|
3280 |
+
"grad_norm": 3.4245402812957764,
|
3281 |
+
"learning_rate": 1.5384545146622852e-06,
|
3282 |
+
"loss": 12.656,
|
3283 |
+
"step": 462
|
3284 |
+
},
|
3285 |
+
{
|
3286 |
+
"epoch": 0.09474599682815778,
|
3287 |
+
"grad_norm": 3.7484071254730225,
|
3288 |
+
"learning_rate": 1.4589395786535953e-06,
|
3289 |
+
"loss": 12.6428,
|
3290 |
+
"step": 463
|
3291 |
+
},
|
3292 |
+
{
|
3293 |
+
"epoch": 0.094950631810508,
|
3294 |
+
"grad_norm": 3.6963798999786377,
|
3295 |
+
"learning_rate": 1.3815039801161721e-06,
|
3296 |
+
"loss": 12.75,
|
3297 |
+
"step": 464
|
3298 |
+
},
|
3299 |
+
{
|
3300 |
+
"epoch": 0.09515526679285824,
|
3301 |
+
"grad_norm": 3.9733211994171143,
|
3302 |
+
"learning_rate": 1.3061510361333185e-06,
|
3303 |
+
"loss": 12.9671,
|
3304 |
+
"step": 465
|
3305 |
+
},
|
3306 |
+
{
|
3307 |
+
"epoch": 0.09535990177520848,
|
3308 |
+
"grad_norm": 3.8950541019439697,
|
3309 |
+
"learning_rate": 1.232883974574367e-06,
|
3310 |
+
"loss": 12.7971,
|
3311 |
+
"step": 466
|
3312 |
+
},
|
3313 |
+
{
|
3314 |
+
"epoch": 0.0955645367575587,
|
3315 |
+
"grad_norm": 4.091780185699463,
|
3316 |
+
"learning_rate": 1.1617059339563807e-06,
|
3317 |
+
"loss": 12.4321,
|
3318 |
+
"step": 467
|
3319 |
+
},
|
3320 |
+
{
|
3321 |
+
"epoch": 0.09576917173990894,
|
3322 |
+
"grad_norm": 4.282808303833008,
|
3323 |
+
"learning_rate": 1.0926199633097157e-06,
|
3324 |
+
"loss": 13.4677,
|
3325 |
+
"step": 468
|
3326 |
+
},
|
3327 |
+
{
|
3328 |
+
"epoch": 0.09597380672225916,
|
3329 |
+
"grad_norm": 4.246225357055664,
|
3330 |
+
"learning_rate": 1.0256290220474307e-06,
|
3331 |
+
"loss": 12.7789,
|
3332 |
+
"step": 469
|
3333 |
+
},
|
3334 |
+
{
|
3335 |
+
"epoch": 0.0961784417046094,
|
3336 |
+
"grad_norm": 4.31233549118042,
|
3337 |
+
"learning_rate": 9.607359798384785e-07,
|
3338 |
+
"loss": 12.7813,
|
3339 |
+
"step": 470
|
3340 |
+
},
|
3341 |
+
{
|
3342 |
+
"epoch": 0.09638307668695964,
|
3343 |
+
"grad_norm": 4.715573787689209,
|
3344 |
+
"learning_rate": 8.979436164848088e-07,
|
3345 |
+
"loss": 13.1816,
|
3346 |
+
"step": 471
|
3347 |
+
},
|
3348 |
+
{
|
3349 |
+
"epoch": 0.09658771166930986,
|
3350 |
+
"grad_norm": 4.476752281188965,
|
3351 |
+
"learning_rate": 8.372546218022747e-07,
|
3352 |
+
"loss": 13.1693,
|
3353 |
+
"step": 472
|
3354 |
+
},
|
3355 |
+
{
|
3356 |
+
"epoch": 0.0967923466516601,
|
3357 |
+
"grad_norm": 4.613335132598877,
|
3358 |
+
"learning_rate": 7.786715955054203e-07,
|
3359 |
+
"loss": 12.6692,
|
3360 |
+
"step": 473
|
3361 |
+
},
|
3362 |
+
{
|
3363 |
+
"epoch": 0.09699698163401034,
|
3364 |
+
"grad_norm": 4.564615249633789,
|
3365 |
+
"learning_rate": 7.221970470961125e-07,
|
3366 |
+
"loss": 12.6351,
|
3367 |
+
"step": 474
|
3368 |
+
},
|
3369 |
+
{
|
3370 |
+
"epoch": 0.09720161661636056,
|
3371 |
+
"grad_norm": 4.650303840637207,
|
3372 |
+
"learning_rate": 6.678333957560512e-07,
|
3373 |
+
"loss": 13.1352,
|
3374 |
+
"step": 475
|
3375 |
+
},
|
3376 |
+
{
|
3377 |
+
"epoch": 0.0974062515987108,
|
3378 |
+
"grad_norm": 4.742563724517822,
|
3379 |
+
"learning_rate": 6.15582970243117e-07,
|
3380 |
+
"loss": 12.8478,
|
3381 |
+
"step": 476
|
3382 |
+
},
|
3383 |
+
{
|
3384 |
+
"epoch": 0.09761088658106103,
|
3385 |
+
"grad_norm": 5.02756929397583,
|
3386 |
+
"learning_rate": 5.654480087916303e-07,
|
3387 |
+
"loss": 12.4434,
|
3388 |
+
"step": 477
|
3389 |
+
},
|
3390 |
+
{
|
3391 |
+
"epoch": 0.09781552156341126,
|
3392 |
+
"grad_norm": 5.11809778213501,
|
3393 |
+
"learning_rate": 5.174306590164879e-07,
|
3394 |
+
"loss": 12.5754,
|
3395 |
+
"step": 478
|
3396 |
+
},
|
3397 |
+
{
|
3398 |
+
"epoch": 0.0980201565457615,
|
3399 |
+
"grad_norm": 5.026023864746094,
|
3400 |
+
"learning_rate": 4.715329778211375e-07,
|
3401 |
+
"loss": 12.7589,
|
3402 |
+
"step": 479
|
3403 |
+
},
|
3404 |
+
{
|
3405 |
+
"epoch": 0.09822479152811173,
|
3406 |
+
"grad_norm": 5.174000263214111,
|
3407 |
+
"learning_rate": 4.277569313094809e-07,
|
3408 |
+
"loss": 12.4781,
|
3409 |
+
"step": 480
|
3410 |
+
},
|
3411 |
+
{
|
3412 |
+
"epoch": 0.09842942651046196,
|
3413 |
+
"grad_norm": 5.73392391204834,
|
3414 |
+
"learning_rate": 3.8610439470164737e-07,
|
3415 |
+
"loss": 13.2662,
|
3416 |
+
"step": 481
|
3417 |
+
},
|
3418 |
+
{
|
3419 |
+
"epoch": 0.0986340614928122,
|
3420 |
+
"grad_norm": 5.5650224685668945,
|
3421 |
+
"learning_rate": 3.465771522536854e-07,
|
3422 |
+
"loss": 12.698,
|
3423 |
+
"step": 482
|
3424 |
+
},
|
3425 |
+
{
|
3426 |
+
"epoch": 0.09883869647516243,
|
3427 |
+
"grad_norm": 5.415269374847412,
|
3428 |
+
"learning_rate": 3.09176897181096e-07,
|
3429 |
+
"loss": 12.4908,
|
3430 |
+
"step": 483
|
3431 |
+
},
|
3432 |
+
{
|
3433 |
+
"epoch": 0.09904333145751266,
|
3434 |
+
"grad_norm": 5.8570170402526855,
|
3435 |
+
"learning_rate": 2.7390523158633554e-07,
|
3436 |
+
"loss": 12.5447,
|
3437 |
+
"step": 484
|
3438 |
+
},
|
3439 |
+
{
|
3440 |
+
"epoch": 0.09924796643986289,
|
3441 |
+
"grad_norm": 6.1962361335754395,
|
3442 |
+
"learning_rate": 2.407636663901591e-07,
|
3443 |
+
"loss": 13.4094,
|
3444 |
+
"step": 485
|
3445 |
+
},
|
3446 |
+
{
|
3447 |
+
"epoch": 0.09945260142221313,
|
3448 |
+
"grad_norm": 6.143553256988525,
|
3449 |
+
"learning_rate": 2.0975362126691712e-07,
|
3450 |
+
"loss": 12.9071,
|
3451 |
+
"step": 486
|
3452 |
+
},
|
3453 |
+
{
|
3454 |
+
"epoch": 0.09965723640456337,
|
3455 |
+
"grad_norm": 6.178109645843506,
|
3456 |
+
"learning_rate": 1.8087642458373134e-07,
|
3457 |
+
"loss": 12.7015,
|
3458 |
+
"step": 487
|
3459 |
+
},
|
3460 |
+
{
|
3461 |
+
"epoch": 0.09986187138691359,
|
3462 |
+
"grad_norm": 6.477899551391602,
|
3463 |
+
"learning_rate": 1.5413331334360182e-07,
|
3464 |
+
"loss": 12.9638,
|
3465 |
+
"step": 488
|
3466 |
+
},
|
3467 |
+
{
|
3468 |
+
"epoch": 0.10006650636926383,
|
3469 |
+
"grad_norm": 6.375460624694824,
|
3470 |
+
"learning_rate": 1.2952543313240472e-07,
|
3471 |
+
"loss": 12.3259,
|
3472 |
+
"step": 489
|
3473 |
+
},
|
3474 |
+
{
|
3475 |
+
"epoch": 0.10027114135161405,
|
3476 |
+
"grad_norm": 7.033268928527832,
|
3477 |
+
"learning_rate": 1.0705383806982606e-07,
|
3478 |
+
"loss": 12.8778,
|
3479 |
+
"step": 490
|
3480 |
+
},
|
3481 |
+
{
|
3482 |
+
"epoch": 0.10047577633396429,
|
3483 |
+
"grad_norm": 6.597715377807617,
|
3484 |
+
"learning_rate": 8.671949076420882e-08,
|
3485 |
+
"loss": 12.0536,
|
3486 |
+
"step": 491
|
3487 |
+
},
|
3488 |
+
{
|
3489 |
+
"epoch": 0.10068041131631453,
|
3490 |
+
"grad_norm": 7.840211868286133,
|
3491 |
+
"learning_rate": 6.852326227130834e-08,
|
3492 |
+
"loss": 13.5376,
|
3493 |
+
"step": 492
|
3494 |
+
},
|
3495 |
+
{
|
3496 |
+
"epoch": 0.10088504629866475,
|
3497 |
+
"grad_norm": 7.423345565795898,
|
3498 |
+
"learning_rate": 5.246593205699424e-08,
|
3499 |
+
"loss": 12.8476,
|
3500 |
+
"step": 493
|
3501 |
+
},
|
3502 |
+
{
|
3503 |
+
"epoch": 0.10108968128101499,
|
3504 |
+
"grad_norm": 8.046225547790527,
|
3505 |
+
"learning_rate": 3.8548187963854956e-08,
|
3506 |
+
"loss": 12.5719,
|
3507 |
+
"step": 494
|
3508 |
+
},
|
3509 |
+
{
|
3510 |
+
"epoch": 0.10129431626336523,
|
3511 |
+
"grad_norm": 8.249234199523926,
|
3512 |
+
"learning_rate": 2.6770626181715773e-08,
|
3513 |
+
"loss": 12.1992,
|
3514 |
+
"step": 495
|
3515 |
+
},
|
3516 |
+
{
|
3517 |
+
"epoch": 0.10149895124571545,
|
3518 |
+
"grad_norm": 9.472099304199219,
|
3519 |
+
"learning_rate": 1.7133751222137007e-08,
|
3520 |
+
"loss": 12.8579,
|
3521 |
+
"step": 496
|
3522 |
+
},
|
3523 |
+
{
|
3524 |
+
"epoch": 0.10170358622806569,
|
3525 |
+
"grad_norm": 9.204952239990234,
|
3526 |
+
"learning_rate": 9.637975896759077e-09,
|
3527 |
+
"loss": 12.0369,
|
3528 |
+
"step": 497
|
3529 |
+
},
|
3530 |
+
{
|
3531 |
+
"epoch": 0.10190822121041591,
|
3532 |
+
"grad_norm": 11.747493743896484,
|
3533 |
+
"learning_rate": 4.2836212996499865e-09,
|
3534 |
+
"loss": 13.0472,
|
3535 |
+
"step": 498
|
3536 |
+
},
|
3537 |
+
{
|
3538 |
+
"epoch": 0.10211285619276615,
|
3539 |
+
"grad_norm": 12.529196739196777,
|
3540 |
+
"learning_rate": 1.0709167935385455e-09,
|
3541 |
+
"loss": 12.7862,
|
3542 |
+
"step": 499
|
3543 |
+
},
|
3544 |
+
{
|
3545 |
+
"epoch": 0.10231749117511639,
|
3546 |
+
"grad_norm": 19.44767951965332,
|
3547 |
+
"learning_rate": 0.0,
|
3548 |
+
"loss": 13.4285,
|
3549 |
+
"step": 500
|
3550 |
+
},
|
3551 |
+
{
|
3552 |
+
"epoch": 0.10231749117511639,
|
3553 |
+
"eval_loss": 3.130511999130249,
|
3554 |
+
"eval_runtime": 47.3503,
|
3555 |
+
"eval_samples_per_second": 173.811,
|
3556 |
+
"eval_steps_per_second": 43.463,
|
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": 9442642165235712.0,
|
3587 |
"train_batch_size": 8,
|
3588 |
"trial_name": null,
|
3589 |
"trial_params": null
|