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 144805440
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:7a5302ed989b61bbf3c5a89d0ce9797280641e435ec2e9f80f205c1902119adb
|
3 |
size 144805440
|
last-checkpoint/optimizer.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 74292308
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:2280588a779f8600beb802cf022367cb692da9b34ac083f171a98b0ae124ae64
|
3 |
size 74292308
|
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 sha256:79d41988296ff7641dd7197ea779bba77e30e921990ba55a18499232a9f57fb0
|
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 sha256:9dc8764f26b1f60541ddcf4504ff4cf226063bff9c4b473f4392b732534a0b84
|
3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
@@ -1,9 +1,9 @@
|
|
1 |
{
|
2 |
-
"best_metric": 1.
|
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": 3.975,
|
2848 |
"eval_steps_per_second": 0.994,
|
2849 |
"step": 400
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2850 |
}
|
2851 |
],
|
2852 |
"logging_steps": 1,
|
@@ -2875,7 +3583,7 @@
|
|
2875 |
"attributes": {}
|
2876 |
}
|
2877 |
},
|
2878 |
-
"total_flos":
|
2879 |
"train_batch_size": 4,
|
2880 |
"trial_name": null,
|
2881 |
"trial_params": null
|
|
|
1 |
{
|
2 |
+
"best_metric": 1.22141695022583,
|
3 |
+
"best_model_checkpoint": "miner_id_24/checkpoint-500",
|
4 |
+
"epoch": 0.06356521393667315,
|
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": 3.975,
|
2848 |
"eval_steps_per_second": 0.994,
|
2849 |
"step": 400
|
2850 |
+
},
|
2851 |
+
{
|
2852 |
+
"epoch": 0.05097930157721187,
|
2853 |
+
"grad_norm": 0.19073964655399323,
|
2854 |
+
"learning_rate": 3.5918795244702396e-05,
|
2855 |
+
"loss": 1.1798,
|
2856 |
+
"step": 401
|
2857 |
+
},
|
2858 |
+
{
|
2859 |
+
"epoch": 0.05110643200508522,
|
2860 |
+
"grad_norm": 0.19348806142807007,
|
2861 |
+
"learning_rate": 3.547489380015701e-05,
|
2862 |
+
"loss": 1.2429,
|
2863 |
+
"step": 402
|
2864 |
+
},
|
2865 |
+
{
|
2866 |
+
"epoch": 0.05123356243295856,
|
2867 |
+
"grad_norm": 0.201893150806427,
|
2868 |
+
"learning_rate": 3.503316020289429e-05,
|
2869 |
+
"loss": 1.2302,
|
2870 |
+
"step": 403
|
2871 |
+
},
|
2872 |
+
{
|
2873 |
+
"epoch": 0.05136069286083191,
|
2874 |
+
"grad_norm": 0.20248207449913025,
|
2875 |
+
"learning_rate": 3.459360929381931e-05,
|
2876 |
+
"loss": 1.2295,
|
2877 |
+
"step": 404
|
2878 |
+
},
|
2879 |
+
{
|
2880 |
+
"epoch": 0.051487823288705256,
|
2881 |
+
"grad_norm": 0.20291946828365326,
|
2882 |
+
"learning_rate": 3.415625584050557e-05,
|
2883 |
+
"loss": 1.2925,
|
2884 |
+
"step": 405
|
2885 |
+
},
|
2886 |
+
{
|
2887 |
+
"epoch": 0.0516149537165786,
|
2888 |
+
"grad_norm": 0.19170844554901123,
|
2889 |
+
"learning_rate": 3.372111453669864e-05,
|
2890 |
+
"loss": 1.1825,
|
2891 |
+
"step": 406
|
2892 |
+
},
|
2893 |
+
{
|
2894 |
+
"epoch": 0.05174208414445195,
|
2895 |
+
"grad_norm": 0.1890149712562561,
|
2896 |
+
"learning_rate": 3.328820000182262e-05,
|
2897 |
+
"loss": 1.149,
|
2898 |
+
"step": 407
|
2899 |
+
},
|
2900 |
+
{
|
2901 |
+
"epoch": 0.051869214572325295,
|
2902 |
+
"grad_norm": 0.20486074686050415,
|
2903 |
+
"learning_rate": 3.285752678048892e-05,
|
2904 |
+
"loss": 1.1458,
|
2905 |
+
"step": 408
|
2906 |
+
},
|
2907 |
+
{
|
2908 |
+
"epoch": 0.05199634500019864,
|
2909 |
+
"grad_norm": 0.20006342232227325,
|
2910 |
+
"learning_rate": 3.242910934200775e-05,
|
2911 |
+
"loss": 1.2031,
|
2912 |
+
"step": 409
|
2913 |
+
},
|
2914 |
+
{
|
2915 |
+
"epoch": 0.05212347542807199,
|
2916 |
+
"grad_norm": 0.20113137364387512,
|
2917 |
+
"learning_rate": 3.2002962079901744e-05,
|
2918 |
+
"loss": 1.2474,
|
2919 |
+
"step": 410
|
2920 |
+
},
|
2921 |
+
{
|
2922 |
+
"epoch": 0.052250605855945334,
|
2923 |
+
"grad_norm": 0.1928662657737732,
|
2924 |
+
"learning_rate": 3.157909931142257e-05,
|
2925 |
+
"loss": 1.2189,
|
2926 |
+
"step": 411
|
2927 |
+
},
|
2928 |
+
{
|
2929 |
+
"epoch": 0.05237773628381868,
|
2930 |
+
"grad_norm": 0.19416049122810364,
|
2931 |
+
"learning_rate": 3.115753527706986e-05,
|
2932 |
+
"loss": 1.2506,
|
2933 |
+
"step": 412
|
2934 |
+
},
|
2935 |
+
{
|
2936 |
+
"epoch": 0.05250486671169203,
|
2937 |
+
"grad_norm": 0.19863539934158325,
|
2938 |
+
"learning_rate": 3.073828414011274e-05,
|
2939 |
+
"loss": 1.2019,
|
2940 |
+
"step": 413
|
2941 |
+
},
|
2942 |
+
{
|
2943 |
+
"epoch": 0.05263199713956537,
|
2944 |
+
"grad_norm": 0.19736243784427643,
|
2945 |
+
"learning_rate": 3.0321359986114096e-05,
|
2946 |
+
"loss": 1.2718,
|
2947 |
+
"step": 414
|
2948 |
+
},
|
2949 |
+
{
|
2950 |
+
"epoch": 0.05275912756743872,
|
2951 |
+
"grad_norm": 0.200786292552948,
|
2952 |
+
"learning_rate": 2.9906776822457205e-05,
|
2953 |
+
"loss": 1.2523,
|
2954 |
+
"step": 415
|
2955 |
+
},
|
2956 |
+
{
|
2957 |
+
"epoch": 0.05288625799531207,
|
2958 |
+
"grad_norm": 0.19629159569740295,
|
2959 |
+
"learning_rate": 2.9494548577875192e-05,
|
2960 |
+
"loss": 1.2156,
|
2961 |
+
"step": 416
|
2962 |
+
},
|
2963 |
+
{
|
2964 |
+
"epoch": 0.05301338842318541,
|
2965 |
+
"grad_norm": 0.19293853640556335,
|
2966 |
+
"learning_rate": 2.9084689101983075e-05,
|
2967 |
+
"loss": 1.2422,
|
2968 |
+
"step": 417
|
2969 |
+
},
|
2970 |
+
{
|
2971 |
+
"epoch": 0.053140518851058756,
|
2972 |
+
"grad_norm": 0.19834856688976288,
|
2973 |
+
"learning_rate": 2.8677212164812462e-05,
|
2974 |
+
"loss": 1.1569,
|
2975 |
+
"step": 418
|
2976 |
+
},
|
2977 |
+
{
|
2978 |
+
"epoch": 0.05326764927893211,
|
2979 |
+
"grad_norm": 0.19939422607421875,
|
2980 |
+
"learning_rate": 2.827213145634887e-05,
|
2981 |
+
"loss": 1.18,
|
2982 |
+
"step": 419
|
2983 |
+
},
|
2984 |
+
{
|
2985 |
+
"epoch": 0.05339477970680545,
|
2986 |
+
"grad_norm": 0.20803573727607727,
|
2987 |
+
"learning_rate": 2.7869460586071873e-05,
|
2988 |
+
"loss": 1.2536,
|
2989 |
+
"step": 420
|
2990 |
+
},
|
2991 |
+
{
|
2992 |
+
"epoch": 0.053521910134678795,
|
2993 |
+
"grad_norm": 0.20102088153362274,
|
2994 |
+
"learning_rate": 2.7469213082497736e-05,
|
2995 |
+
"loss": 1.2664,
|
2996 |
+
"step": 421
|
2997 |
+
},
|
2998 |
+
{
|
2999 |
+
"epoch": 0.053649040562552146,
|
3000 |
+
"grad_norm": 0.20080140233039856,
|
3001 |
+
"learning_rate": 2.7071402392725097e-05,
|
3002 |
+
"loss": 1.2931,
|
3003 |
+
"step": 422
|
3004 |
+
},
|
3005 |
+
{
|
3006 |
+
"epoch": 0.05377617099042549,
|
3007 |
+
"grad_norm": 0.19506552815437317,
|
3008 |
+
"learning_rate": 2.6676041881982962e-05,
|
3009 |
+
"loss": 1.2461,
|
3010 |
+
"step": 423
|
3011 |
+
},
|
3012 |
+
{
|
3013 |
+
"epoch": 0.053903301418298834,
|
3014 |
+
"grad_norm": 0.19534236192703247,
|
3015 |
+
"learning_rate": 2.6283144833181783e-05,
|
3016 |
+
"loss": 1.1971,
|
3017 |
+
"step": 424
|
3018 |
+
},
|
3019 |
+
{
|
3020 |
+
"epoch": 0.054030431846172185,
|
3021 |
+
"grad_norm": 0.20050783455371857,
|
3022 |
+
"learning_rate": 2.589272444646723e-05,
|
3023 |
+
"loss": 1.169,
|
3024 |
+
"step": 425
|
3025 |
+
},
|
3026 |
+
{
|
3027 |
+
"epoch": 0.05415756227404553,
|
3028 |
+
"grad_norm": 0.201995849609375,
|
3029 |
+
"learning_rate": 2.5504793838776586e-05,
|
3030 |
+
"loss": 1.2758,
|
3031 |
+
"step": 426
|
3032 |
+
},
|
3033 |
+
{
|
3034 |
+
"epoch": 0.05428469270191887,
|
3035 |
+
"grad_norm": 0.20075969398021698,
|
3036 |
+
"learning_rate": 2.5119366043398264e-05,
|
3037 |
+
"loss": 1.2455,
|
3038 |
+
"step": 427
|
3039 |
+
},
|
3040 |
+
{
|
3041 |
+
"epoch": 0.054411823129792224,
|
3042 |
+
"grad_norm": 0.19450737535953522,
|
3043 |
+
"learning_rate": 2.473645400953366e-05,
|
3044 |
+
"loss": 1.2319,
|
3045 |
+
"step": 428
|
3046 |
+
},
|
3047 |
+
{
|
3048 |
+
"epoch": 0.05453895355766557,
|
3049 |
+
"grad_norm": 0.19657433032989502,
|
3050 |
+
"learning_rate": 2.4356070601862324e-05,
|
3051 |
+
"loss": 1.2031,
|
3052 |
+
"step": 429
|
3053 |
+
},
|
3054 |
+
{
|
3055 |
+
"epoch": 0.05466608398553891,
|
3056 |
+
"grad_norm": 0.1956198513507843,
|
3057 |
+
"learning_rate": 2.3978228600109565e-05,
|
3058 |
+
"loss": 1.2345,
|
3059 |
+
"step": 430
|
3060 |
+
},
|
3061 |
+
{
|
3062 |
+
"epoch": 0.05479321441341226,
|
3063 |
+
"grad_norm": 0.19880710542201996,
|
3064 |
+
"learning_rate": 2.3602940698617325e-05,
|
3065 |
+
"loss": 1.3102,
|
3066 |
+
"step": 431
|
3067 |
+
},
|
3068 |
+
{
|
3069 |
+
"epoch": 0.05492034484128561,
|
3070 |
+
"grad_norm": 0.19965411722660065,
|
3071 |
+
"learning_rate": 2.3230219505917426e-05,
|
3072 |
+
"loss": 1.2873,
|
3073 |
+
"step": 432
|
3074 |
+
},
|
3075 |
+
{
|
3076 |
+
"epoch": 0.05504747526915895,
|
3077 |
+
"grad_norm": 0.19630952179431915,
|
3078 |
+
"learning_rate": 2.2860077544308124e-05,
|
3079 |
+
"loss": 1.272,
|
3080 |
+
"step": 433
|
3081 |
+
},
|
3082 |
+
{
|
3083 |
+
"epoch": 0.0551746056970323,
|
3084 |
+
"grad_norm": 0.19560639560222626,
|
3085 |
+
"learning_rate": 2.249252724943336e-05,
|
3086 |
+
"loss": 1.2593,
|
3087 |
+
"step": 434
|
3088 |
+
},
|
3089 |
+
{
|
3090 |
+
"epoch": 0.055301736124905645,
|
3091 |
+
"grad_norm": 0.20019298791885376,
|
3092 |
+
"learning_rate": 2.2127580969864925e-05,
|
3093 |
+
"loss": 1.2723,
|
3094 |
+
"step": 435
|
3095 |
+
},
|
3096 |
+
{
|
3097 |
+
"epoch": 0.05542886655277899,
|
3098 |
+
"grad_norm": 0.20072917640209198,
|
3099 |
+
"learning_rate": 2.176525096668769e-05,
|
3100 |
+
"loss": 1.2069,
|
3101 |
+
"step": 436
|
3102 |
+
},
|
3103 |
+
{
|
3104 |
+
"epoch": 0.05555599698065234,
|
3105 |
+
"grad_norm": 0.20204520225524902,
|
3106 |
+
"learning_rate": 2.1405549413087544e-05,
|
3107 |
+
"loss": 1.2361,
|
3108 |
+
"step": 437
|
3109 |
+
},
|
3110 |
+
{
|
3111 |
+
"epoch": 0.055683127408525684,
|
3112 |
+
"grad_norm": 0.20441173017024994,
|
3113 |
+
"learning_rate": 2.1048488393942454e-05,
|
3114 |
+
"loss": 1.1882,
|
3115 |
+
"step": 438
|
3116 |
+
},
|
3117 |
+
{
|
3118 |
+
"epoch": 0.05581025783639903,
|
3119 |
+
"grad_norm": 0.1971142739057541,
|
3120 |
+
"learning_rate": 2.0694079905416473e-05,
|
3121 |
+
"loss": 1.2168,
|
3122 |
+
"step": 439
|
3123 |
+
},
|
3124 |
+
{
|
3125 |
+
"epoch": 0.05593738826427238,
|
3126 |
+
"grad_norm": 0.20565344393253326,
|
3127 |
+
"learning_rate": 2.0342335854556737e-05,
|
3128 |
+
"loss": 1.2486,
|
3129 |
+
"step": 440
|
3130 |
+
},
|
3131 |
+
{
|
3132 |
+
"epoch": 0.05606451869214572,
|
3133 |
+
"grad_norm": 0.19731251895427704,
|
3134 |
+
"learning_rate": 1.9993268058893343e-05,
|
3135 |
+
"loss": 1.116,
|
3136 |
+
"step": 441
|
3137 |
+
},
|
3138 |
+
{
|
3139 |
+
"epoch": 0.05619164912001907,
|
3140 |
+
"grad_norm": 0.19447311758995056,
|
3141 |
+
"learning_rate": 1.964688824604234e-05,
|
3142 |
+
"loss": 1.2497,
|
3143 |
+
"step": 442
|
3144 |
+
},
|
3145 |
+
{
|
3146 |
+
"epoch": 0.05631877954789242,
|
3147 |
+
"grad_norm": 0.2041405290365219,
|
3148 |
+
"learning_rate": 1.930320805331176e-05,
|
3149 |
+
"loss": 1.2843,
|
3150 |
+
"step": 443
|
3151 |
+
},
|
3152 |
+
{
|
3153 |
+
"epoch": 0.05644590997576576,
|
3154 |
+
"grad_norm": 0.20058204233646393,
|
3155 |
+
"learning_rate": 1.896223902731058e-05,
|
3156 |
+
"loss": 1.2685,
|
3157 |
+
"step": 444
|
3158 |
+
},
|
3159 |
+
{
|
3160 |
+
"epoch": 0.056573040403639106,
|
3161 |
+
"grad_norm": 0.20194143056869507,
|
3162 |
+
"learning_rate": 1.8623992623560893e-05,
|
3163 |
+
"loss": 1.2751,
|
3164 |
+
"step": 445
|
3165 |
+
},
|
3166 |
+
{
|
3167 |
+
"epoch": 0.05670017083151246,
|
3168 |
+
"grad_norm": 0.20336030423641205,
|
3169 |
+
"learning_rate": 1.828848020611288e-05,
|
3170 |
+
"loss": 1.2362,
|
3171 |
+
"step": 446
|
3172 |
+
},
|
3173 |
+
{
|
3174 |
+
"epoch": 0.0568273012593858,
|
3175 |
+
"grad_norm": 0.2054579108953476,
|
3176 |
+
"learning_rate": 1.7955713047163157e-05,
|
3177 |
+
"loss": 1.2254,
|
3178 |
+
"step": 447
|
3179 |
+
},
|
3180 |
+
{
|
3181 |
+
"epoch": 0.056954431687259145,
|
3182 |
+
"grad_norm": 0.20549102127552032,
|
3183 |
+
"learning_rate": 1.762570232667595e-05,
|
3184 |
+
"loss": 1.2574,
|
3185 |
+
"step": 448
|
3186 |
+
},
|
3187 |
+
{
|
3188 |
+
"epoch": 0.057081562115132496,
|
3189 |
+
"grad_norm": 0.2029975950717926,
|
3190 |
+
"learning_rate": 1.7298459132007627e-05,
|
3191 |
+
"loss": 1.2066,
|
3192 |
+
"step": 449
|
3193 |
+
},
|
3194 |
+
{
|
3195 |
+
"epoch": 0.05720869254300584,
|
3196 |
+
"grad_norm": 0.20551042258739471,
|
3197 |
+
"learning_rate": 1.6973994457534026e-05,
|
3198 |
+
"loss": 1.2384,
|
3199 |
+
"step": 450
|
3200 |
+
},
|
3201 |
+
{
|
3202 |
+
"epoch": 0.057335822970879184,
|
3203 |
+
"grad_norm": 0.20290708541870117,
|
3204 |
+
"learning_rate": 1.6652319204281187e-05,
|
3205 |
+
"loss": 1.2257,
|
3206 |
+
"step": 451
|
3207 |
+
},
|
3208 |
+
{
|
3209 |
+
"epoch": 0.057462953398752535,
|
3210 |
+
"grad_norm": 0.202660471200943,
|
3211 |
+
"learning_rate": 1.6333444179559078e-05,
|
3212 |
+
"loss": 1.1865,
|
3213 |
+
"step": 452
|
3214 |
+
},
|
3215 |
+
{
|
3216 |
+
"epoch": 0.05759008382662588,
|
3217 |
+
"grad_norm": 0.19935418665409088,
|
3218 |
+
"learning_rate": 1.601738009659849e-05,
|
3219 |
+
"loss": 1.2445,
|
3220 |
+
"step": 453
|
3221 |
+
},
|
3222 |
+
{
|
3223 |
+
"epoch": 0.05771721425449922,
|
3224 |
+
"grad_norm": 0.19397136569023132,
|
3225 |
+
"learning_rate": 1.5704137574191203e-05,
|
3226 |
+
"loss": 1.2246,
|
3227 |
+
"step": 454
|
3228 |
+
},
|
3229 |
+
{
|
3230 |
+
"epoch": 0.057844344682372574,
|
3231 |
+
"grad_norm": 0.20033963024616241,
|
3232 |
+
"learning_rate": 1.5393727136333035e-05,
|
3233 |
+
"loss": 1.2452,
|
3234 |
+
"step": 455
|
3235 |
+
},
|
3236 |
+
{
|
3237 |
+
"epoch": 0.05797147511024592,
|
3238 |
+
"grad_norm": 0.19725894927978516,
|
3239 |
+
"learning_rate": 1.5086159211870442e-05,
|
3240 |
+
"loss": 1.1902,
|
3241 |
+
"step": 456
|
3242 |
+
},
|
3243 |
+
{
|
3244 |
+
"epoch": 0.05809860553811926,
|
3245 |
+
"grad_norm": 0.21174763143062592,
|
3246 |
+
"learning_rate": 1.4781444134150047e-05,
|
3247 |
+
"loss": 1.1956,
|
3248 |
+
"step": 457
|
3249 |
+
},
|
3250 |
+
{
|
3251 |
+
"epoch": 0.05822573596599261,
|
3252 |
+
"grad_norm": 0.20643867552280426,
|
3253 |
+
"learning_rate": 1.447959214067155e-05,
|
3254 |
+
"loss": 1.2708,
|
3255 |
+
"step": 458
|
3256 |
+
},
|
3257 |
+
{
|
3258 |
+
"epoch": 0.05835286639386596,
|
3259 |
+
"grad_norm": 0.203439399600029,
|
3260 |
+
"learning_rate": 1.4180613372743679e-05,
|
3261 |
+
"loss": 1.1935,
|
3262 |
+
"step": 459
|
3263 |
+
},
|
3264 |
+
{
|
3265 |
+
"epoch": 0.0584799968217393,
|
3266 |
+
"grad_norm": 0.2026398777961731,
|
3267 |
+
"learning_rate": 1.3884517875143544e-05,
|
3268 |
+
"loss": 1.2331,
|
3269 |
+
"step": 460
|
3270 |
+
},
|
3271 |
+
{
|
3272 |
+
"epoch": 0.05860712724961265,
|
3273 |
+
"grad_norm": 0.20543427765369415,
|
3274 |
+
"learning_rate": 1.3591315595779108e-05,
|
3275 |
+
"loss": 1.2039,
|
3276 |
+
"step": 461
|
3277 |
+
},
|
3278 |
+
{
|
3279 |
+
"epoch": 0.058734257677485996,
|
3280 |
+
"grad_norm": 0.2049439400434494,
|
3281 |
+
"learning_rate": 1.3301016385355092e-05,
|
3282 |
+
"loss": 1.1847,
|
3283 |
+
"step": 462
|
3284 |
+
},
|
3285 |
+
{
|
3286 |
+
"epoch": 0.05886138810535934,
|
3287 |
+
"grad_norm": 0.20000450313091278,
|
3288 |
+
"learning_rate": 1.3013629997041853e-05,
|
3289 |
+
"loss": 1.2432,
|
3290 |
+
"step": 463
|
3291 |
+
},
|
3292 |
+
{
|
3293 |
+
"epoch": 0.05898851853323269,
|
3294 |
+
"grad_norm": 0.20123903453350067,
|
3295 |
+
"learning_rate": 1.2729166086147803e-05,
|
3296 |
+
"loss": 1.2698,
|
3297 |
+
"step": 464
|
3298 |
+
},
|
3299 |
+
{
|
3300 |
+
"epoch": 0.059115648961106035,
|
3301 |
+
"grad_norm": 0.20424886047840118,
|
3302 |
+
"learning_rate": 1.2447634209795e-05,
|
3303 |
+
"loss": 1.2635,
|
3304 |
+
"step": 465
|
3305 |
+
},
|
3306 |
+
{
|
3307 |
+
"epoch": 0.05924277938897938,
|
3308 |
+
"grad_norm": 0.19165809452533722,
|
3309 |
+
"learning_rate": 1.2169043826598058e-05,
|
3310 |
+
"loss": 1.1772,
|
3311 |
+
"step": 466
|
3312 |
+
},
|
3313 |
+
{
|
3314 |
+
"epoch": 0.05936990981685273,
|
3315 |
+
"grad_norm": 0.19973435997962952,
|
3316 |
+
"learning_rate": 1.1893404296346423e-05,
|
3317 |
+
"loss": 1.2326,
|
3318 |
+
"step": 467
|
3319 |
+
},
|
3320 |
+
{
|
3321 |
+
"epoch": 0.059497040244726074,
|
3322 |
+
"grad_norm": 0.20723304152488708,
|
3323 |
+
"learning_rate": 1.1620724879689792e-05,
|
3324 |
+
"loss": 1.2451,
|
3325 |
+
"step": 468
|
3326 |
+
},
|
3327 |
+
{
|
3328 |
+
"epoch": 0.05962417067259942,
|
3329 |
+
"grad_norm": 0.1965390294790268,
|
3330 |
+
"learning_rate": 1.135101473782706e-05,
|
3331 |
+
"loss": 1.2154,
|
3332 |
+
"step": 469
|
3333 |
+
},
|
3334 |
+
{
|
3335 |
+
"epoch": 0.05975130110047277,
|
3336 |
+
"grad_norm": 0.19724096357822418,
|
3337 |
+
"learning_rate": 1.1084282932198541e-05,
|
3338 |
+
"loss": 1.2092,
|
3339 |
+
"step": 470
|
3340 |
+
},
|
3341 |
+
{
|
3342 |
+
"epoch": 0.05987843152834611,
|
3343 |
+
"grad_norm": 0.19244475662708282,
|
3344 |
+
"learning_rate": 1.0820538424181515e-05,
|
3345 |
+
"loss": 1.12,
|
3346 |
+
"step": 471
|
3347 |
+
},
|
3348 |
+
{
|
3349 |
+
"epoch": 0.06000556195621946,
|
3350 |
+
"grad_norm": 0.20514468848705292,
|
3351 |
+
"learning_rate": 1.0559790074789133e-05,
|
3352 |
+
"loss": 1.1499,
|
3353 |
+
"step": 472
|
3354 |
+
},
|
3355 |
+
{
|
3356 |
+
"epoch": 0.06013269238409281,
|
3357 |
+
"grad_norm": 0.19560429453849792,
|
3358 |
+
"learning_rate": 1.030204664437271e-05,
|
3359 |
+
"loss": 1.2266,
|
3360 |
+
"step": 473
|
3361 |
+
},
|
3362 |
+
{
|
3363 |
+
"epoch": 0.06025982281196615,
|
3364 |
+
"grad_norm": 0.20317231118679047,
|
3365 |
+
"learning_rate": 1.0047316792327499e-05,
|
3366 |
+
"loss": 1.1913,
|
3367 |
+
"step": 474
|
3368 |
+
},
|
3369 |
+
{
|
3370 |
+
"epoch": 0.060386953239839496,
|
3371 |
+
"grad_norm": 0.19489817321300507,
|
3372 |
+
"learning_rate": 9.795609076801625e-06,
|
3373 |
+
"loss": 1.2461,
|
3374 |
+
"step": 475
|
3375 |
+
},
|
3376 |
+
{
|
3377 |
+
"epoch": 0.06051408366771285,
|
3378 |
+
"grad_norm": 0.1974310576915741,
|
3379 |
+
"learning_rate": 9.546931954408622e-06,
|
3380 |
+
"loss": 1.1443,
|
3381 |
+
"step": 476
|
3382 |
+
},
|
3383 |
+
{
|
3384 |
+
"epoch": 0.06064121409558619,
|
3385 |
+
"grad_norm": 0.19223681092262268,
|
3386 |
+
"learning_rate": 9.301293779943321e-06,
|
3387 |
+
"loss": 1.1864,
|
3388 |
+
"step": 477
|
3389 |
+
},
|
3390 |
+
{
|
3391 |
+
"epoch": 0.060768344523459535,
|
3392 |
+
"grad_norm": 0.2029593586921692,
|
3393 |
+
"learning_rate": 9.058702806101172e-06,
|
3394 |
+
"loss": 1.2081,
|
3395 |
+
"step": 478
|
3396 |
+
},
|
3397 |
+
{
|
3398 |
+
"epoch": 0.060895474951332886,
|
3399 |
+
"grad_norm": 0.20288948714733124,
|
3400 |
+
"learning_rate": 8.819167183200905e-06,
|
3401 |
+
"loss": 1.2794,
|
3402 |
+
"step": 479
|
3403 |
+
},
|
3404 |
+
{
|
3405 |
+
"epoch": 0.06102260537920623,
|
3406 |
+
"grad_norm": 0.20905464887619019,
|
3407 |
+
"learning_rate": 8.58269495891081e-06,
|
3408 |
+
"loss": 1.2125,
|
3409 |
+
"step": 480
|
3410 |
+
},
|
3411 |
+
{
|
3412 |
+
"epoch": 0.061149735807079574,
|
3413 |
+
"grad_norm": 0.2005111575126648,
|
3414 |
+
"learning_rate": 8.349294077978265e-06,
|
3415 |
+
"loss": 1.2183,
|
3416 |
+
"step": 481
|
3417 |
+
},
|
3418 |
+
{
|
3419 |
+
"epoch": 0.061276866234952924,
|
3420 |
+
"grad_norm": 0.20153559744358063,
|
3421 |
+
"learning_rate": 8.118972381962853e-06,
|
3422 |
+
"loss": 1.2213,
|
3423 |
+
"step": 482
|
3424 |
+
},
|
3425 |
+
{
|
3426 |
+
"epoch": 0.06140399666282627,
|
3427 |
+
"grad_norm": 0.20067910850048065,
|
3428 |
+
"learning_rate": 7.891737608972927e-06,
|
3429 |
+
"loss": 1.1919,
|
3430 |
+
"step": 483
|
3431 |
+
},
|
3432 |
+
{
|
3433 |
+
"epoch": 0.06153112709069961,
|
3434 |
+
"grad_norm": 0.2007380872964859,
|
3435 |
+
"learning_rate": 7.6675973934056e-06,
|
3436 |
+
"loss": 1.2182,
|
3437 |
+
"step": 484
|
3438 |
+
},
|
3439 |
+
{
|
3440 |
+
"epoch": 0.06165825751857296,
|
3441 |
+
"grad_norm": 0.19200630486011505,
|
3442 |
+
"learning_rate": 7.4465592656903114e-06,
|
3443 |
+
"loss": 1.2089,
|
3444 |
+
"step": 485
|
3445 |
+
},
|
3446 |
+
{
|
3447 |
+
"epoch": 0.06178538794644631,
|
3448 |
+
"grad_norm": 0.19756072759628296,
|
3449 |
+
"learning_rate": 7.228630652035717e-06,
|
3450 |
+
"loss": 1.2236,
|
3451 |
+
"step": 486
|
3452 |
+
},
|
3453 |
+
{
|
3454 |
+
"epoch": 0.06191251837431965,
|
3455 |
+
"grad_norm": 0.20330612361431122,
|
3456 |
+
"learning_rate": 7.0138188741803225e-06,
|
3457 |
+
"loss": 1.1603,
|
3458 |
+
"step": 487
|
3459 |
+
},
|
3460 |
+
{
|
3461 |
+
"epoch": 0.062039648802193,
|
3462 |
+
"grad_norm": 0.2018778920173645,
|
3463 |
+
"learning_rate": 6.802131149146373e-06,
|
3464 |
+
"loss": 1.2256,
|
3465 |
+
"step": 488
|
3466 |
+
},
|
3467 |
+
{
|
3468 |
+
"epoch": 0.062166779230066346,
|
3469 |
+
"grad_norm": 0.19810867309570312,
|
3470 |
+
"learning_rate": 6.59357458899752e-06,
|
3471 |
+
"loss": 1.2123,
|
3472 |
+
"step": 489
|
3473 |
+
},
|
3474 |
+
{
|
3475 |
+
"epoch": 0.06229390965793969,
|
3476 |
+
"grad_norm": 0.19678053259849548,
|
3477 |
+
"learning_rate": 6.388156200599726e-06,
|
3478 |
+
"loss": 1.2266,
|
3479 |
+
"step": 490
|
3480 |
+
},
|
3481 |
+
{
|
3482 |
+
"epoch": 0.06242104008581304,
|
3483 |
+
"grad_norm": 0.20404191315174103,
|
3484 |
+
"learning_rate": 6.185882885385952e-06,
|
3485 |
+
"loss": 1.2376,
|
3486 |
+
"step": 491
|
3487 |
+
},
|
3488 |
+
{
|
3489 |
+
"epoch": 0.06254817051368639,
|
3490 |
+
"grad_norm": 0.20588378608226776,
|
3491 |
+
"learning_rate": 5.986761439124289e-06,
|
3492 |
+
"loss": 1.2604,
|
3493 |
+
"step": 492
|
3494 |
+
},
|
3495 |
+
{
|
3496 |
+
"epoch": 0.06267530094155974,
|
3497 |
+
"grad_norm": 0.19841307401657104,
|
3498 |
+
"learning_rate": 5.790798551689592e-06,
|
3499 |
+
"loss": 1.2849,
|
3500 |
+
"step": 493
|
3501 |
+
},
|
3502 |
+
{
|
3503 |
+
"epoch": 0.06280243136943307,
|
3504 |
+
"grad_norm": 0.20967555046081543,
|
3505 |
+
"learning_rate": 5.598000806838766e-06,
|
3506 |
+
"loss": 1.2321,
|
3507 |
+
"step": 494
|
3508 |
+
},
|
3509 |
+
{
|
3510 |
+
"epoch": 0.06292956179730642,
|
3511 |
+
"grad_norm": 0.19931279122829437,
|
3512 |
+
"learning_rate": 5.408374681989548e-06,
|
3513 |
+
"loss": 1.278,
|
3514 |
+
"step": 495
|
3515 |
+
},
|
3516 |
+
{
|
3517 |
+
"epoch": 0.06305669222517978,
|
3518 |
+
"grad_norm": 0.19611623883247375,
|
3519 |
+
"learning_rate": 5.221926548002876e-06,
|
3520 |
+
"loss": 1.1999,
|
3521 |
+
"step": 496
|
3522 |
+
},
|
3523 |
+
{
|
3524 |
+
"epoch": 0.06318382265305311,
|
3525 |
+
"grad_norm": 0.19863781332969666,
|
3526 |
+
"learning_rate": 5.038662668968886e-06,
|
3527 |
+
"loss": 1.2295,
|
3528 |
+
"step": 497
|
3529 |
+
},
|
3530 |
+
{
|
3531 |
+
"epoch": 0.06331095308092646,
|
3532 |
+
"grad_norm": 0.19629532098770142,
|
3533 |
+
"learning_rate": 4.858589201996433e-06,
|
3534 |
+
"loss": 1.2194,
|
3535 |
+
"step": 498
|
3536 |
+
},
|
3537 |
+
{
|
3538 |
+
"epoch": 0.06343808350879981,
|
3539 |
+
"grad_norm": 0.21103760600090027,
|
3540 |
+
"learning_rate": 4.681712197006205e-06,
|
3541 |
+
"loss": 1.2861,
|
3542 |
+
"step": 499
|
3543 |
+
},
|
3544 |
+
{
|
3545 |
+
"epoch": 0.06356521393667315,
|
3546 |
+
"grad_norm": 0.20111264288425446,
|
3547 |
+
"learning_rate": 4.508037596527526e-06,
|
3548 |
+
"loss": 1.2123,
|
3549 |
+
"step": 500
|
3550 |
+
},
|
3551 |
+
{
|
3552 |
+
"epoch": 0.06356521393667315,
|
3553 |
+
"eval_loss": 1.22141695022583,
|
3554 |
+
"eval_runtime": 1257.3017,
|
3555 |
+
"eval_samples_per_second": 3.977,
|
3556 |
+
"eval_steps_per_second": 0.994,
|
3557 |
+
"step": 500
|
3558 |
}
|
3559 |
],
|
3560 |
"logging_steps": 1,
|
|
|
3583 |
"attributes": {}
|
3584 |
}
|
3585 |
},
|
3586 |
+
"total_flos": 5.182000113647616e+18,
|
3587 |
"train_batch_size": 4,
|
3588 |
"trial_name": null,
|
3589 |
"trial_params": null
|