Training in progress, step 488, 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:da6cb8e3ceb59865787e3be19146ecfe8f090e00a5e13cdf927c08d575198fae
|
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:b4957b732848f725fe8e97d4887a86c7374d04e7a61a5f57433336b19c5f05d6
|
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 sha256:41b8611fadb80f1bdbbbba5e4fc8c3638269de0aed2e0197a291cadd4ef1f1ad
|
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 sha256:1e7fb2e4e3f7c67f2979e17d634f1bb48e2792b76f31dccbd7feec15021f43a0
|
3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
@@ -1,9 +1,9 @@
|
|
1 |
{
|
2 |
"best_metric": 0.813357412815094,
|
3 |
"best_model_checkpoint": "miner_id_24/checkpoint-400",
|
4 |
-
"epoch":
|
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 +2847,622 @@
|
|
2847 |
"eval_samples_per_second": 58.256,
|
2848 |
"eval_steps_per_second": 14.599,
|
2849 |
"step": 400
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2850 |
}
|
2851 |
],
|
2852 |
"logging_steps": 1,
|
@@ -2870,12 +3486,12 @@
|
|
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": 0.813357412815094,
|
3 |
"best_model_checkpoint": "miner_id_24/checkpoint-400",
|
4 |
+
"epoch": 1.001025641025641,
|
5 |
"eval_steps": 100,
|
6 |
+
"global_step": 488,
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
|
|
2847 |
"eval_samples_per_second": 58.256,
|
2848 |
"eval_steps_per_second": 14.599,
|
2849 |
"step": 400
|
2850 |
+
},
|
2851 |
+
{
|
2852 |
+
"epoch": 0.8225641025641026,
|
2853 |
+
"grad_norm": 2.1067893505096436,
|
2854 |
+
"learning_rate": 8.287193065856935e-06,
|
2855 |
+
"loss": 0.857,
|
2856 |
+
"step": 401
|
2857 |
+
},
|
2858 |
+
{
|
2859 |
+
"epoch": 0.8246153846153846,
|
2860 |
+
"grad_norm": 2.872138261795044,
|
2861 |
+
"learning_rate": 8.103069939070945e-06,
|
2862 |
+
"loss": 0.9339,
|
2863 |
+
"step": 402
|
2864 |
+
},
|
2865 |
+
{
|
2866 |
+
"epoch": 0.8266666666666667,
|
2867 |
+
"grad_norm": 2.640960216522217,
|
2868 |
+
"learning_rate": 7.920834754120304e-06,
|
2869 |
+
"loss": 0.7526,
|
2870 |
+
"step": 403
|
2871 |
+
},
|
2872 |
+
{
|
2873 |
+
"epoch": 0.8287179487179487,
|
2874 |
+
"grad_norm": 2.455273389816284,
|
2875 |
+
"learning_rate": 7.740495722810271e-06,
|
2876 |
+
"loss": 0.8467,
|
2877 |
+
"step": 404
|
2878 |
+
},
|
2879 |
+
{
|
2880 |
+
"epoch": 0.8307692307692308,
|
2881 |
+
"grad_norm": 3.302844285964966,
|
2882 |
+
"learning_rate": 7.562060971502383e-06,
|
2883 |
+
"loss": 0.7286,
|
2884 |
+
"step": 405
|
2885 |
+
},
|
2886 |
+
{
|
2887 |
+
"epoch": 0.8328205128205128,
|
2888 |
+
"grad_norm": 2.6610546112060547,
|
2889 |
+
"learning_rate": 7.385538540748327e-06,
|
2890 |
+
"loss": 0.8152,
|
2891 |
+
"step": 406
|
2892 |
+
},
|
2893 |
+
{
|
2894 |
+
"epoch": 0.8348717948717949,
|
2895 |
+
"grad_norm": 2.261064291000366,
|
2896 |
+
"learning_rate": 7.21093638492763e-06,
|
2897 |
+
"loss": 0.6278,
|
2898 |
+
"step": 407
|
2899 |
+
},
|
2900 |
+
{
|
2901 |
+
"epoch": 0.8369230769230769,
|
2902 |
+
"grad_norm": 3.7655062675476074,
|
2903 |
+
"learning_rate": 7.038262371889159e-06,
|
2904 |
+
"loss": 0.8202,
|
2905 |
+
"step": 408
|
2906 |
+
},
|
2907 |
+
{
|
2908 |
+
"epoch": 0.838974358974359,
|
2909 |
+
"grad_norm": 2.8197319507598877,
|
2910 |
+
"learning_rate": 6.867524282596655e-06,
|
2911 |
+
"loss": 0.8749,
|
2912 |
+
"step": 409
|
2913 |
+
},
|
2914 |
+
{
|
2915 |
+
"epoch": 0.841025641025641,
|
2916 |
+
"grad_norm": 2.7082679271698,
|
2917 |
+
"learning_rate": 6.698729810778065e-06,
|
2918 |
+
"loss": 0.7821,
|
2919 |
+
"step": 410
|
2920 |
+
},
|
2921 |
+
{
|
2922 |
+
"epoch": 0.8430769230769231,
|
2923 |
+
"grad_norm": 2.2520222663879395,
|
2924 |
+
"learning_rate": 6.531886562578859e-06,
|
2925 |
+
"loss": 0.5002,
|
2926 |
+
"step": 411
|
2927 |
+
},
|
2928 |
+
{
|
2929 |
+
"epoch": 0.8451282051282051,
|
2930 |
+
"grad_norm": 2.5191712379455566,
|
2931 |
+
"learning_rate": 6.367002056219284e-06,
|
2932 |
+
"loss": 0.7014,
|
2933 |
+
"step": 412
|
2934 |
+
},
|
2935 |
+
{
|
2936 |
+
"epoch": 0.8471794871794872,
|
2937 |
+
"grad_norm": 5.02801513671875,
|
2938 |
+
"learning_rate": 6.204083721655607e-06,
|
2939 |
+
"loss": 0.8951,
|
2940 |
+
"step": 413
|
2941 |
+
},
|
2942 |
+
{
|
2943 |
+
"epoch": 0.8492307692307692,
|
2944 |
+
"grad_norm": 2.9786057472229004,
|
2945 |
+
"learning_rate": 6.043138900245277e-06,
|
2946 |
+
"loss": 0.9056,
|
2947 |
+
"step": 414
|
2948 |
+
},
|
2949 |
+
{
|
2950 |
+
"epoch": 0.8512820512820513,
|
2951 |
+
"grad_norm": 2.791360378265381,
|
2952 |
+
"learning_rate": 5.884174844416102e-06,
|
2953 |
+
"loss": 0.7512,
|
2954 |
+
"step": 415
|
2955 |
+
},
|
2956 |
+
{
|
2957 |
+
"epoch": 0.8533333333333334,
|
2958 |
+
"grad_norm": 2.5517685413360596,
|
2959 |
+
"learning_rate": 5.727198717339511e-06,
|
2960 |
+
"loss": 0.7028,
|
2961 |
+
"step": 416
|
2962 |
+
},
|
2963 |
+
{
|
2964 |
+
"epoch": 0.8553846153846154,
|
2965 |
+
"grad_norm": 2.8642799854278564,
|
2966 |
+
"learning_rate": 5.572217592607687e-06,
|
2967 |
+
"loss": 0.8472,
|
2968 |
+
"step": 417
|
2969 |
+
},
|
2970 |
+
{
|
2971 |
+
"epoch": 0.8574358974358974,
|
2972 |
+
"grad_norm": 2.9429116249084473,
|
2973 |
+
"learning_rate": 5.41923845391486e-06,
|
2974 |
+
"loss": 0.5711,
|
2975 |
+
"step": 418
|
2976 |
+
},
|
2977 |
+
{
|
2978 |
+
"epoch": 0.8594871794871795,
|
2979 |
+
"grad_norm": 2.486117124557495,
|
2980 |
+
"learning_rate": 5.2682681947426375e-06,
|
2981 |
+
"loss": 0.5636,
|
2982 |
+
"step": 419
|
2983 |
+
},
|
2984 |
+
{
|
2985 |
+
"epoch": 0.8615384615384616,
|
2986 |
+
"grad_norm": 3.1933610439300537,
|
2987 |
+
"learning_rate": 5.1193136180493095e-06,
|
2988 |
+
"loss": 0.6232,
|
2989 |
+
"step": 420
|
2990 |
+
},
|
2991 |
+
{
|
2992 |
+
"epoch": 0.8635897435897436,
|
2993 |
+
"grad_norm": 3.0535292625427246,
|
2994 |
+
"learning_rate": 4.97238143596333e-06,
|
2995 |
+
"loss": 0.7645,
|
2996 |
+
"step": 421
|
2997 |
+
},
|
2998 |
+
{
|
2999 |
+
"epoch": 0.8656410256410256,
|
3000 |
+
"grad_norm": 2.3928494453430176,
|
3001 |
+
"learning_rate": 4.827478269480895e-06,
|
3002 |
+
"loss": 0.4642,
|
3003 |
+
"step": 422
|
3004 |
+
},
|
3005 |
+
{
|
3006 |
+
"epoch": 0.8676923076923077,
|
3007 |
+
"grad_norm": 3.066641330718994,
|
3008 |
+
"learning_rate": 4.684610648167503e-06,
|
3009 |
+
"loss": 0.8899,
|
3010 |
+
"step": 423
|
3011 |
+
},
|
3012 |
+
{
|
3013 |
+
"epoch": 0.8697435897435898,
|
3014 |
+
"grad_norm": 2.6845340728759766,
|
3015 |
+
"learning_rate": 4.54378500986381e-06,
|
3016 |
+
"loss": 0.6358,
|
3017 |
+
"step": 424
|
3018 |
+
},
|
3019 |
+
{
|
3020 |
+
"epoch": 0.8717948717948718,
|
3021 |
+
"grad_norm": 3.794255495071411,
|
3022 |
+
"learning_rate": 4.405007700395497e-06,
|
3023 |
+
"loss": 0.7231,
|
3024 |
+
"step": 425
|
3025 |
+
},
|
3026 |
+
{
|
3027 |
+
"epoch": 0.8738461538461538,
|
3028 |
+
"grad_norm": 3.2792623043060303,
|
3029 |
+
"learning_rate": 4.268284973287273e-06,
|
3030 |
+
"loss": 0.6754,
|
3031 |
+
"step": 426
|
3032 |
+
},
|
3033 |
+
{
|
3034 |
+
"epoch": 0.8758974358974358,
|
3035 |
+
"grad_norm": 4.019685745239258,
|
3036 |
+
"learning_rate": 4.133622989481145e-06,
|
3037 |
+
"loss": 0.8601,
|
3038 |
+
"step": 427
|
3039 |
+
},
|
3040 |
+
{
|
3041 |
+
"epoch": 0.877948717948718,
|
3042 |
+
"grad_norm": 5.093227386474609,
|
3043 |
+
"learning_rate": 4.001027817058789e-06,
|
3044 |
+
"loss": 0.9993,
|
3045 |
+
"step": 428
|
3046 |
+
},
|
3047 |
+
{
|
3048 |
+
"epoch": 0.88,
|
3049 |
+
"grad_norm": 3.0464017391204834,
|
3050 |
+
"learning_rate": 3.870505430968069e-06,
|
3051 |
+
"loss": 0.6586,
|
3052 |
+
"step": 429
|
3053 |
+
},
|
3054 |
+
{
|
3055 |
+
"epoch": 0.882051282051282,
|
3056 |
+
"grad_norm": 3.241689920425415,
|
3057 |
+
"learning_rate": 3.7420617127538248e-06,
|
3058 |
+
"loss": 0.7764,
|
3059 |
+
"step": 430
|
3060 |
+
},
|
3061 |
+
{
|
3062 |
+
"epoch": 0.884102564102564,
|
3063 |
+
"grad_norm": 3.573296546936035,
|
3064 |
+
"learning_rate": 3.615702450292857e-06,
|
3065 |
+
"loss": 0.7413,
|
3066 |
+
"step": 431
|
3067 |
+
},
|
3068 |
+
{
|
3069 |
+
"epoch": 0.8861538461538462,
|
3070 |
+
"grad_norm": 3.759922742843628,
|
3071 |
+
"learning_rate": 3.4914333375330898e-06,
|
3072 |
+
"loss": 0.7581,
|
3073 |
+
"step": 432
|
3074 |
+
},
|
3075 |
+
{
|
3076 |
+
"epoch": 0.8882051282051282,
|
3077 |
+
"grad_norm": 3.483435869216919,
|
3078 |
+
"learning_rate": 3.369259974236988e-06,
|
3079 |
+
"loss": 0.5884,
|
3080 |
+
"step": 433
|
3081 |
+
},
|
3082 |
+
{
|
3083 |
+
"epoch": 0.8902564102564102,
|
3084 |
+
"grad_norm": 3.6014602184295654,
|
3085 |
+
"learning_rate": 3.249187865729264e-06,
|
3086 |
+
"loss": 0.7147,
|
3087 |
+
"step": 434
|
3088 |
+
},
|
3089 |
+
{
|
3090 |
+
"epoch": 0.8923076923076924,
|
3091 |
+
"grad_norm": 3.7031497955322266,
|
3092 |
+
"learning_rate": 3.1312224226487442e-06,
|
3093 |
+
"loss": 0.7788,
|
3094 |
+
"step": 435
|
3095 |
+
},
|
3096 |
+
{
|
3097 |
+
"epoch": 0.8943589743589744,
|
3098 |
+
"grad_norm": 3.692850351333618,
|
3099 |
+
"learning_rate": 3.0153689607045845e-06,
|
3100 |
+
"loss": 0.7269,
|
3101 |
+
"step": 436
|
3102 |
+
},
|
3103 |
+
{
|
3104 |
+
"epoch": 0.8964102564102564,
|
3105 |
+
"grad_norm": 3.5270211696624756,
|
3106 |
+
"learning_rate": 2.901632700436757e-06,
|
3107 |
+
"loss": 0.7998,
|
3108 |
+
"step": 437
|
3109 |
+
},
|
3110 |
+
{
|
3111 |
+
"epoch": 0.8984615384615384,
|
3112 |
+
"grad_norm": 3.858431577682495,
|
3113 |
+
"learning_rate": 2.790018766980773e-06,
|
3114 |
+
"loss": 0.8085,
|
3115 |
+
"step": 438
|
3116 |
+
},
|
3117 |
+
{
|
3118 |
+
"epoch": 0.9005128205128206,
|
3119 |
+
"grad_norm": 3.0940706729888916,
|
3120 |
+
"learning_rate": 2.680532189836732e-06,
|
3121 |
+
"loss": 0.6352,
|
3122 |
+
"step": 439
|
3123 |
+
},
|
3124 |
+
{
|
3125 |
+
"epoch": 0.9025641025641026,
|
3126 |
+
"grad_norm": 4.406203269958496,
|
3127 |
+
"learning_rate": 2.573177902642726e-06,
|
3128 |
+
"loss": 0.8474,
|
3129 |
+
"step": 440
|
3130 |
+
},
|
3131 |
+
{
|
3132 |
+
"epoch": 0.9046153846153846,
|
3133 |
+
"grad_norm": 3.506132125854492,
|
3134 |
+
"learning_rate": 2.467960742952463e-06,
|
3135 |
+
"loss": 0.5986,
|
3136 |
+
"step": 441
|
3137 |
+
},
|
3138 |
+
{
|
3139 |
+
"epoch": 0.9066666666666666,
|
3140 |
+
"grad_norm": 4.049818992614746,
|
3141 |
+
"learning_rate": 2.3648854520173237e-06,
|
3142 |
+
"loss": 0.821,
|
3143 |
+
"step": 442
|
3144 |
+
},
|
3145 |
+
{
|
3146 |
+
"epoch": 0.9087179487179488,
|
3147 |
+
"grad_norm": 3.5991644859313965,
|
3148 |
+
"learning_rate": 2.2639566745727205e-06,
|
3149 |
+
"loss": 0.8375,
|
3150 |
+
"step": 443
|
3151 |
+
},
|
3152 |
+
{
|
3153 |
+
"epoch": 0.9107692307692308,
|
3154 |
+
"grad_norm": 4.406035900115967,
|
3155 |
+
"learning_rate": 2.1651789586287442e-06,
|
3156 |
+
"loss": 0.797,
|
3157 |
+
"step": 444
|
3158 |
+
},
|
3159 |
+
{
|
3160 |
+
"epoch": 0.9128205128205128,
|
3161 |
+
"grad_norm": 5.088395595550537,
|
3162 |
+
"learning_rate": 2.068556755265272e-06,
|
3163 |
+
"loss": 0.8375,
|
3164 |
+
"step": 445
|
3165 |
+
},
|
3166 |
+
{
|
3167 |
+
"epoch": 0.9148717948717948,
|
3168 |
+
"grad_norm": 7.348681926727295,
|
3169 |
+
"learning_rate": 1.974094418431388e-06,
|
3170 |
+
"loss": 0.7428,
|
3171 |
+
"step": 446
|
3172 |
+
},
|
3173 |
+
{
|
3174 |
+
"epoch": 0.916923076923077,
|
3175 |
+
"grad_norm": 5.225501537322998,
|
3176 |
+
"learning_rate": 1.8817962047491699e-06,
|
3177 |
+
"loss": 0.8446,
|
3178 |
+
"step": 447
|
3179 |
+
},
|
3180 |
+
{
|
3181 |
+
"epoch": 0.918974358974359,
|
3182 |
+
"grad_norm": 5.816100120544434,
|
3183 |
+
"learning_rate": 1.7916662733218847e-06,
|
3184 |
+
"loss": 1.1185,
|
3185 |
+
"step": 448
|
3186 |
+
},
|
3187 |
+
{
|
3188 |
+
"epoch": 0.921025641025641,
|
3189 |
+
"grad_norm": 6.379942417144775,
|
3190 |
+
"learning_rate": 1.70370868554659e-06,
|
3191 |
+
"loss": 1.1595,
|
3192 |
+
"step": 449
|
3193 |
+
},
|
3194 |
+
{
|
3195 |
+
"epoch": 0.9230769230769231,
|
3196 |
+
"grad_norm": 11.633496284484863,
|
3197 |
+
"learning_rate": 1.6179274049310966e-06,
|
3198 |
+
"loss": 1.8986,
|
3199 |
+
"step": 450
|
3200 |
+
},
|
3201 |
+
{
|
3202 |
+
"epoch": 0.9251282051282051,
|
3203 |
+
"grad_norm": 2.403395175933838,
|
3204 |
+
"learning_rate": 1.5343262969153783e-06,
|
3205 |
+
"loss": 0.8987,
|
3206 |
+
"step": 451
|
3207 |
+
},
|
3208 |
+
{
|
3209 |
+
"epoch": 0.9271794871794872,
|
3210 |
+
"grad_norm": 2.5767834186553955,
|
3211 |
+
"learning_rate": 1.4529091286973995e-06,
|
3212 |
+
"loss": 0.7343,
|
3213 |
+
"step": 452
|
3214 |
+
},
|
3215 |
+
{
|
3216 |
+
"epoch": 0.9292307692307692,
|
3217 |
+
"grad_norm": 2.610196352005005,
|
3218 |
+
"learning_rate": 1.3736795690633354e-06,
|
3219 |
+
"loss": 0.808,
|
3220 |
+
"step": 453
|
3221 |
+
},
|
3222 |
+
{
|
3223 |
+
"epoch": 0.9312820512820513,
|
3224 |
+
"grad_norm": 3.0826358795166016,
|
3225 |
+
"learning_rate": 1.2966411882222696e-06,
|
3226 |
+
"loss": 0.8493,
|
3227 |
+
"step": 454
|
3228 |
+
},
|
3229 |
+
{
|
3230 |
+
"epoch": 0.9333333333333333,
|
3231 |
+
"grad_norm": 2.933225154876709,
|
3232 |
+
"learning_rate": 1.2217974576453073e-06,
|
3233 |
+
"loss": 0.7868,
|
3234 |
+
"step": 455
|
3235 |
+
},
|
3236 |
+
{
|
3237 |
+
"epoch": 0.9353846153846154,
|
3238 |
+
"grad_norm": 2.5828311443328857,
|
3239 |
+
"learning_rate": 1.1491517499091498e-06,
|
3240 |
+
"loss": 0.6543,
|
3241 |
+
"step": 456
|
3242 |
+
},
|
3243 |
+
{
|
3244 |
+
"epoch": 0.9374358974358974,
|
3245 |
+
"grad_norm": 2.710909605026245,
|
3246 |
+
"learning_rate": 1.0787073385441048e-06,
|
3247 |
+
"loss": 0.6686,
|
3248 |
+
"step": 457
|
3249 |
+
},
|
3250 |
+
{
|
3251 |
+
"epoch": 0.9394871794871795,
|
3252 |
+
"grad_norm": 2.428332567214966,
|
3253 |
+
"learning_rate": 1.0104673978866164e-06,
|
3254 |
+
"loss": 0.6812,
|
3255 |
+
"step": 458
|
3256 |
+
},
|
3257 |
+
{
|
3258 |
+
"epoch": 0.9415384615384615,
|
3259 |
+
"grad_norm": 3.3245909214019775,
|
3260 |
+
"learning_rate": 9.44435002936167e-07,
|
3261 |
+
"loss": 0.8798,
|
3262 |
+
"step": 459
|
3263 |
+
},
|
3264 |
+
{
|
3265 |
+
"epoch": 0.9435897435897436,
|
3266 |
+
"grad_norm": 3.67254376411438,
|
3267 |
+
"learning_rate": 8.806131292167618e-07,
|
3268 |
+
"loss": 0.8829,
|
3269 |
+
"step": 460
|
3270 |
+
},
|
3271 |
+
{
|
3272 |
+
"epoch": 0.9456410256410256,
|
3273 |
+
"grad_norm": 2.8118197917938232,
|
3274 |
+
"learning_rate": 8.190046526428242e-07,
|
3275 |
+
"loss": 0.6889,
|
3276 |
+
"step": 461
|
3277 |
+
},
|
3278 |
+
{
|
3279 |
+
"epoch": 0.9476923076923077,
|
3280 |
+
"grad_norm": 2.5844786167144775,
|
3281 |
+
"learning_rate": 7.596123493895991e-07,
|
3282 |
+
"loss": 0.8317,
|
3283 |
+
"step": 462
|
3284 |
+
},
|
3285 |
+
{
|
3286 |
+
"epoch": 0.9497435897435897,
|
3287 |
+
"grad_norm": 3.378085136413574,
|
3288 |
+
"learning_rate": 7.024388957680705e-07,
|
3289 |
+
"loss": 0.7873,
|
3290 |
+
"step": 463
|
3291 |
+
},
|
3292 |
+
{
|
3293 |
+
"epoch": 0.9517948717948718,
|
3294 |
+
"grad_norm": 3.895148992538452,
|
3295 |
+
"learning_rate": 6.474868681043578e-07,
|
3296 |
+
"loss": 0.9465,
|
3297 |
+
"step": 464
|
3298 |
+
},
|
3299 |
+
{
|
3300 |
+
"epoch": 0.9538461538461539,
|
3301 |
+
"grad_norm": 2.4344851970672607,
|
3302 |
+
"learning_rate": 5.947587426236078e-07,
|
3303 |
+
"loss": 0.5724,
|
3304 |
+
"step": 465
|
3305 |
+
},
|
3306 |
+
{
|
3307 |
+
"epoch": 0.9558974358974359,
|
3308 |
+
"grad_norm": 3.014648199081421,
|
3309 |
+
"learning_rate": 5.442568953384186e-07,
|
3310 |
+
"loss": 0.72,
|
3311 |
+
"step": 466
|
3312 |
+
},
|
3313 |
+
{
|
3314 |
+
"epoch": 0.9579487179487179,
|
3315 |
+
"grad_norm": 3.841538429260254,
|
3316 |
+
"learning_rate": 4.959836019417963e-07,
|
3317 |
+
"loss": 0.7759,
|
3318 |
+
"step": 467
|
3319 |
+
},
|
3320 |
+
{
|
3321 |
+
"epoch": 0.96,
|
3322 |
+
"grad_norm": 2.5754928588867188,
|
3323 |
+
"learning_rate": 4.4994103770457653e-07,
|
3324 |
+
"loss": 0.579,
|
3325 |
+
"step": 468
|
3326 |
+
},
|
3327 |
+
{
|
3328 |
+
"epoch": 0.9620512820512821,
|
3329 |
+
"grad_norm": 3.4755969047546387,
|
3330 |
+
"learning_rate": 4.06131277377414e-07,
|
3331 |
+
"loss": 0.8353,
|
3332 |
+
"step": 469
|
3333 |
+
},
|
3334 |
+
{
|
3335 |
+
"epoch": 0.9641025641025641,
|
3336 |
+
"grad_norm": 3.1296160221099854,
|
3337 |
+
"learning_rate": 3.6455629509730136e-07,
|
3338 |
+
"loss": 0.8424,
|
3339 |
+
"step": 470
|
3340 |
+
},
|
3341 |
+
{
|
3342 |
+
"epoch": 0.9661538461538461,
|
3343 |
+
"grad_norm": 2.8293471336364746,
|
3344 |
+
"learning_rate": 3.2521796429859084e-07,
|
3345 |
+
"loss": 0.6852,
|
3346 |
+
"step": 471
|
3347 |
+
},
|
3348 |
+
{
|
3349 |
+
"epoch": 0.9682051282051282,
|
3350 |
+
"grad_norm": 3.9338464736938477,
|
3351 |
+
"learning_rate": 2.8811805762860576e-07,
|
3352 |
+
"loss": 0.7015,
|
3353 |
+
"step": 472
|
3354 |
+
},
|
3355 |
+
{
|
3356 |
+
"epoch": 0.9702564102564103,
|
3357 |
+
"grad_norm": 3.2302651405334473,
|
3358 |
+
"learning_rate": 2.532582468677214e-07,
|
3359 |
+
"loss": 0.7174,
|
3360 |
+
"step": 473
|
3361 |
+
},
|
3362 |
+
{
|
3363 |
+
"epoch": 0.9723076923076923,
|
3364 |
+
"grad_norm": 3.1608636379241943,
|
3365 |
+
"learning_rate": 2.206401028540639e-07,
|
3366 |
+
"loss": 0.6316,
|
3367 |
+
"step": 474
|
3368 |
+
},
|
3369 |
+
{
|
3370 |
+
"epoch": 0.9743589743589743,
|
3371 |
+
"grad_norm": 3.069204807281494,
|
3372 |
+
"learning_rate": 1.9026509541272275e-07,
|
3373 |
+
"loss": 0.6748,
|
3374 |
+
"step": 475
|
3375 |
+
},
|
3376 |
+
{
|
3377 |
+
"epoch": 0.9764102564102564,
|
3378 |
+
"grad_norm": 3.582129955291748,
|
3379 |
+
"learning_rate": 1.6213459328950352e-07,
|
3380 |
+
"loss": 0.7998,
|
3381 |
+
"step": 476
|
3382 |
+
},
|
3383 |
+
{
|
3384 |
+
"epoch": 0.9784615384615385,
|
3385 |
+
"grad_norm": 3.0587315559387207,
|
3386 |
+
"learning_rate": 1.3624986408924956e-07,
|
3387 |
+
"loss": 0.5679,
|
3388 |
+
"step": 477
|
3389 |
+
},
|
3390 |
+
{
|
3391 |
+
"epoch": 0.9805128205128205,
|
3392 |
+
"grad_norm": 4.51352071762085,
|
3393 |
+
"learning_rate": 1.1261207421874309e-07,
|
3394 |
+
"loss": 0.7706,
|
3395 |
+
"step": 478
|
3396 |
+
},
|
3397 |
+
{
|
3398 |
+
"epoch": 0.9825641025641025,
|
3399 |
+
"grad_norm": 3.1249945163726807,
|
3400 |
+
"learning_rate": 9.12222888341252e-08,
|
3401 |
+
"loss": 0.64,
|
3402 |
+
"step": 479
|
3403 |
+
},
|
3404 |
+
{
|
3405 |
+
"epoch": 0.9846153846153847,
|
3406 |
+
"grad_norm": 3.75665020942688,
|
3407 |
+
"learning_rate": 7.208147179291192e-08,
|
3408 |
+
"loss": 0.9884,
|
3409 |
+
"step": 480
|
3410 |
+
},
|
3411 |
+
{
|
3412 |
+
"epoch": 0.9866666666666667,
|
3413 |
+
"grad_norm": 3.27207088470459,
|
3414 |
+
"learning_rate": 5.5190485610534525e-08,
|
3415 |
+
"loss": 0.4883,
|
3416 |
+
"step": 481
|
3417 |
+
},
|
3418 |
+
{
|
3419 |
+
"epoch": 0.9887179487179487,
|
3420 |
+
"grad_norm": 4.946873664855957,
|
3421 |
+
"learning_rate": 4.055009142152067e-08,
|
3422 |
+
"loss": 0.8809,
|
3423 |
+
"step": 482
|
3424 |
+
},
|
3425 |
+
{
|
3426 |
+
"epoch": 0.9907692307692307,
|
3427 |
+
"grad_norm": 3.621314287185669,
|
3428 |
+
"learning_rate": 2.8160948945138434e-08,
|
3429 |
+
"loss": 0.8451,
|
3430 |
+
"step": 483
|
3431 |
+
},
|
3432 |
+
{
|
3433 |
+
"epoch": 0.9928205128205129,
|
3434 |
+
"grad_norm": 3.5091235637664795,
|
3435 |
+
"learning_rate": 1.802361645573125e-08,
|
3436 |
+
"loss": 0.7231,
|
3437 |
+
"step": 484
|
3438 |
+
},
|
3439 |
+
{
|
3440 |
+
"epoch": 0.9948717948717949,
|
3441 |
+
"grad_norm": 3.9179797172546387,
|
3442 |
+
"learning_rate": 1.0138550757493592e-08,
|
3443 |
+
"loss": 0.6897,
|
3444 |
+
"step": 485
|
3445 |
+
},
|
3446 |
+
{
|
3447 |
+
"epoch": 0.9969230769230769,
|
3448 |
+
"grad_norm": 5.395089149475098,
|
3449 |
+
"learning_rate": 4.506107163948503e-09,
|
3450 |
+
"loss": 1.0574,
|
3451 |
+
"step": 486
|
3452 |
+
},
|
3453 |
+
{
|
3454 |
+
"epoch": 0.9989743589743589,
|
3455 |
+
"grad_norm": 6.939630508422852,
|
3456 |
+
"learning_rate": 1.1265394818993358e-09,
|
3457 |
+
"loss": 1.0886,
|
3458 |
+
"step": 487
|
3459 |
+
},
|
3460 |
+
{
|
3461 |
+
"epoch": 1.001025641025641,
|
3462 |
+
"grad_norm": 24.829050064086914,
|
3463 |
+
"learning_rate": 0.0,
|
3464 |
+
"loss": 3.5024,
|
3465 |
+
"step": 488
|
3466 |
}
|
3467 |
],
|
3468 |
"logging_steps": 1,
|
|
|
3486 |
"should_evaluate": false,
|
3487 |
"should_log": false,
|
3488 |
"should_save": true,
|
3489 |
+
"should_training_stop": true
|
3490 |
},
|
3491 |
"attributes": {}
|
3492 |
}
|
3493 |
},
|
3494 |
+
"total_flos": 3.771257911640064e+16,
|
3495 |
"train_batch_size": 8,
|
3496 |
"trial_name": null,
|
3497 |
"trial_params": null
|