Training in progress, step 4500, 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 2718107304
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5fbd5aa58a02999ada4c0f3cd43ca41dde21582eb2d5dc255f13cbc8023650b6
|
| 3 |
size 2718107304
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 145486330
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7d7ccc1eca1a7ceff3e0dde210062328c431e027749f3f43c26d6230e844198d
|
| 3 |
size 145486330
|
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:b5b09ae3a06fa6068073a53264cd1c287209ec69eb82f509ea260660e6955ead
|
| 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 sha256:48993fe83df0cc4ae4d85d5c3e846862f5ffd116f5d7e38ed8352e362fdf4bbb
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch": 0.
|
| 5 |
"eval_steps": 1000,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -2839,6 +2839,356 @@
|
|
| 2839 |
"eval_samples_per_second": 9.649,
|
| 2840 |
"eval_steps_per_second": 1.206,
|
| 2841 |
"step": 4000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2842 |
}
|
| 2843 |
],
|
| 2844 |
"logging_steps": 10,
|
|
@@ -2858,7 +3208,7 @@
|
|
| 2858 |
"attributes": {}
|
| 2859 |
}
|
| 2860 |
},
|
| 2861 |
-
"total_flos":
|
| 2862 |
"train_batch_size": 4,
|
| 2863 |
"trial_name": null,
|
| 2864 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.8192244675040962,
|
| 5 |
"eval_steps": 1000,
|
| 6 |
+
"global_step": 4500,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 2839 |
"eval_samples_per_second": 9.649,
|
| 2840 |
"eval_steps_per_second": 1.206,
|
| 2841 |
"step": 4000
|
| 2842 |
+
},
|
| 2843 |
+
{
|
| 2844 |
+
"epoch": 0.7300200254869834,
|
| 2845 |
+
"grad_norm": 5.740320205688477,
|
| 2846 |
+
"learning_rate": 2.0232804874534313e-05,
|
| 2847 |
+
"loss": 2.4582,
|
| 2848 |
+
"step": 4010
|
| 2849 |
+
},
|
| 2850 |
+
{
|
| 2851 |
+
"epoch": 0.7318405243036592,
|
| 2852 |
+
"grad_norm": 5.987521171569824,
|
| 2853 |
+
"learning_rate": 1.998062403243704e-05,
|
| 2854 |
+
"loss": 2.4499,
|
| 2855 |
+
"step": 4020
|
| 2856 |
+
},
|
| 2857 |
+
{
|
| 2858 |
+
"epoch": 0.733661023120335,
|
| 2859 |
+
"grad_norm": 5.337474346160889,
|
| 2860 |
+
"learning_rate": 1.9729631629190042e-05,
|
| 2861 |
+
"loss": 2.4598,
|
| 2862 |
+
"step": 4030
|
| 2863 |
+
},
|
| 2864 |
+
{
|
| 2865 |
+
"epoch": 0.7354815219370108,
|
| 2866 |
+
"grad_norm": 5.655992031097412,
|
| 2867 |
+
"learning_rate": 1.9479837601346457e-05,
|
| 2868 |
+
"loss": 2.4601,
|
| 2869 |
+
"step": 4040
|
| 2870 |
+
},
|
| 2871 |
+
{
|
| 2872 |
+
"epoch": 0.7373020207536866,
|
| 2873 |
+
"grad_norm": 5.7331414222717285,
|
| 2874 |
+
"learning_rate": 1.923125183801678e-05,
|
| 2875 |
+
"loss": 2.4445,
|
| 2876 |
+
"step": 4050
|
| 2877 |
+
},
|
| 2878 |
+
{
|
| 2879 |
+
"epoch": 0.7391225195703622,
|
| 2880 |
+
"grad_norm": 5.471503257751465,
|
| 2881 |
+
"learning_rate": 1.898388418047753e-05,
|
| 2882 |
+
"loss": 2.4683,
|
| 2883 |
+
"step": 4060
|
| 2884 |
+
},
|
| 2885 |
+
{
|
| 2886 |
+
"epoch": 0.740943018387038,
|
| 2887 |
+
"grad_norm": 5.409184455871582,
|
| 2888 |
+
"learning_rate": 1.87377444217815e-05,
|
| 2889 |
+
"loss": 2.4358,
|
| 2890 |
+
"step": 4070
|
| 2891 |
+
},
|
| 2892 |
+
{
|
| 2893 |
+
"epoch": 0.7427635172037138,
|
| 2894 |
+
"grad_norm": 5.579779148101807,
|
| 2895 |
+
"learning_rate": 1.8492842306370182e-05,
|
| 2896 |
+
"loss": 2.4989,
|
| 2897 |
+
"step": 4080
|
| 2898 |
+
},
|
| 2899 |
+
{
|
| 2900 |
+
"epoch": 0.7445840160203896,
|
| 2901 |
+
"grad_norm": 5.366626262664795,
|
| 2902 |
+
"learning_rate": 1.8249187529687895e-05,
|
| 2903 |
+
"loss": 2.4102,
|
| 2904 |
+
"step": 4090
|
| 2905 |
+
},
|
| 2906 |
+
{
|
| 2907 |
+
"epoch": 0.7464045148370654,
|
| 2908 |
+
"grad_norm": 5.263418674468994,
|
| 2909 |
+
"learning_rate": 1.8006789737797984e-05,
|
| 2910 |
+
"loss": 2.4573,
|
| 2911 |
+
"step": 4100
|
| 2912 |
+
},
|
| 2913 |
+
{
|
| 2914 |
+
"epoch": 0.7482250136537412,
|
| 2915 |
+
"grad_norm": 5.129177570343018,
|
| 2916 |
+
"learning_rate": 1.7765658527000966e-05,
|
| 2917 |
+
"loss": 2.4792,
|
| 2918 |
+
"step": 4110
|
| 2919 |
+
},
|
| 2920 |
+
{
|
| 2921 |
+
"epoch": 0.7500455124704168,
|
| 2922 |
+
"grad_norm": 6.237401962280273,
|
| 2923 |
+
"learning_rate": 1.7525803443454615e-05,
|
| 2924 |
+
"loss": 2.479,
|
| 2925 |
+
"step": 4120
|
| 2926 |
+
},
|
| 2927 |
+
{
|
| 2928 |
+
"epoch": 0.7518660112870926,
|
| 2929 |
+
"grad_norm": 6.163425445556641,
|
| 2930 |
+
"learning_rate": 1.728723398279603e-05,
|
| 2931 |
+
"loss": 2.4222,
|
| 2932 |
+
"step": 4130
|
| 2933 |
+
},
|
| 2934 |
+
{
|
| 2935 |
+
"epoch": 0.7536865101037684,
|
| 2936 |
+
"grad_norm": 5.254932403564453,
|
| 2937 |
+
"learning_rate": 1.7049959589765686e-05,
|
| 2938 |
+
"loss": 2.4307,
|
| 2939 |
+
"step": 4140
|
| 2940 |
+
},
|
| 2941 |
+
{
|
| 2942 |
+
"epoch": 0.7555070089204442,
|
| 2943 |
+
"grad_norm": 6.144068717956543,
|
| 2944 |
+
"learning_rate": 1.6813989657833534e-05,
|
| 2945 |
+
"loss": 2.4923,
|
| 2946 |
+
"step": 4150
|
| 2947 |
+
},
|
| 2948 |
+
{
|
| 2949 |
+
"epoch": 0.75732750773712,
|
| 2950 |
+
"grad_norm": 5.038397789001465,
|
| 2951 |
+
"learning_rate": 1.6579333528827205e-05,
|
| 2952 |
+
"loss": 2.4732,
|
| 2953 |
+
"step": 4160
|
| 2954 |
+
},
|
| 2955 |
+
{
|
| 2956 |
+
"epoch": 0.7591480065537958,
|
| 2957 |
+
"grad_norm": 5.2848076820373535,
|
| 2958 |
+
"learning_rate": 1.634600049256204e-05,
|
| 2959 |
+
"loss": 2.4651,
|
| 2960 |
+
"step": 4170
|
| 2961 |
+
},
|
| 2962 |
+
{
|
| 2963 |
+
"epoch": 0.7609685053704716,
|
| 2964 |
+
"grad_norm": 5.274468898773193,
|
| 2965 |
+
"learning_rate": 1.611399978647342e-05,
|
| 2966 |
+
"loss": 2.4407,
|
| 2967 |
+
"step": 4180
|
| 2968 |
+
},
|
| 2969 |
+
{
|
| 2970 |
+
"epoch": 0.7627890041871472,
|
| 2971 |
+
"grad_norm": 5.039272308349609,
|
| 2972 |
+
"learning_rate": 1.588334059525099e-05,
|
| 2973 |
+
"loss": 2.4892,
|
| 2974 |
+
"step": 4190
|
| 2975 |
+
},
|
| 2976 |
+
{
|
| 2977 |
+
"epoch": 0.764609503003823,
|
| 2978 |
+
"grad_norm": 5.7419867515563965,
|
| 2979 |
+
"learning_rate": 1.5654032050475138e-05,
|
| 2980 |
+
"loss": 2.4456,
|
| 2981 |
+
"step": 4200
|
| 2982 |
+
},
|
| 2983 |
+
{
|
| 2984 |
+
"epoch": 0.7664300018204988,
|
| 2985 |
+
"grad_norm": 5.30146598815918,
|
| 2986 |
+
"learning_rate": 1.5426083230255405e-05,
|
| 2987 |
+
"loss": 2.4515,
|
| 2988 |
+
"step": 4210
|
| 2989 |
+
},
|
| 2990 |
+
{
|
| 2991 |
+
"epoch": 0.7682505006371746,
|
| 2992 |
+
"grad_norm": 4.977199077606201,
|
| 2993 |
+
"learning_rate": 1.5199503158871115e-05,
|
| 2994 |
+
"loss": 2.4317,
|
| 2995 |
+
"step": 4220
|
| 2996 |
+
},
|
| 2997 |
+
{
|
| 2998 |
+
"epoch": 0.7700709994538504,
|
| 2999 |
+
"grad_norm": 5.318095684051514,
|
| 3000 |
+
"learning_rate": 1.4974300806414082e-05,
|
| 3001 |
+
"loss": 2.403,
|
| 3002 |
+
"step": 4230
|
| 3003 |
+
},
|
| 3004 |
+
{
|
| 3005 |
+
"epoch": 0.7718914982705262,
|
| 3006 |
+
"grad_norm": 5.638497352600098,
|
| 3007 |
+
"learning_rate": 1.4750485088433592e-05,
|
| 3008 |
+
"loss": 2.4327,
|
| 3009 |
+
"step": 4240
|
| 3010 |
+
},
|
| 3011 |
+
{
|
| 3012 |
+
"epoch": 0.7737119970872018,
|
| 3013 |
+
"grad_norm": 5.739340305328369,
|
| 3014 |
+
"learning_rate": 1.4528064865583301e-05,
|
| 3015 |
+
"loss": 2.4266,
|
| 3016 |
+
"step": 4250
|
| 3017 |
+
},
|
| 3018 |
+
{
|
| 3019 |
+
"epoch": 0.7755324959038776,
|
| 3020 |
+
"grad_norm": 4.749205112457275,
|
| 3021 |
+
"learning_rate": 1.4307048943270606e-05,
|
| 3022 |
+
"loss": 2.4136,
|
| 3023 |
+
"step": 4260
|
| 3024 |
+
},
|
| 3025 |
+
{
|
| 3026 |
+
"epoch": 0.7773529947205534,
|
| 3027 |
+
"grad_norm": 5.616302490234375,
|
| 3028 |
+
"learning_rate": 1.4087446071307903e-05,
|
| 3029 |
+
"loss": 2.4197,
|
| 3030 |
+
"step": 4270
|
| 3031 |
+
},
|
| 3032 |
+
{
|
| 3033 |
+
"epoch": 0.7791734935372292,
|
| 3034 |
+
"grad_norm": 5.402510643005371,
|
| 3035 |
+
"learning_rate": 1.3869264943566263e-05,
|
| 3036 |
+
"loss": 2.4194,
|
| 3037 |
+
"step": 4280
|
| 3038 |
+
},
|
| 3039 |
+
{
|
| 3040 |
+
"epoch": 0.780993992353905,
|
| 3041 |
+
"grad_norm": 5.278769493103027,
|
| 3042 |
+
"learning_rate": 1.3652514197631277e-05,
|
| 3043 |
+
"loss": 2.4351,
|
| 3044 |
+
"step": 4290
|
| 3045 |
+
},
|
| 3046 |
+
{
|
| 3047 |
+
"epoch": 0.7828144911705808,
|
| 3048 |
+
"grad_norm": 6.828596115112305,
|
| 3049 |
+
"learning_rate": 1.343720241446103e-05,
|
| 3050 |
+
"loss": 2.3813,
|
| 3051 |
+
"step": 4300
|
| 3052 |
+
},
|
| 3053 |
+
{
|
| 3054 |
+
"epoch": 0.7846349899872566,
|
| 3055 |
+
"grad_norm": 5.306332588195801,
|
| 3056 |
+
"learning_rate": 1.322333811804643e-05,
|
| 3057 |
+
"loss": 2.4133,
|
| 3058 |
+
"step": 4310
|
| 3059 |
+
},
|
| 3060 |
+
{
|
| 3061 |
+
"epoch": 0.7864554888039322,
|
| 3062 |
+
"grad_norm": 5.437227249145508,
|
| 3063 |
+
"learning_rate": 1.3010929775073765e-05,
|
| 3064 |
+
"loss": 2.4166,
|
| 3065 |
+
"step": 4320
|
| 3066 |
+
},
|
| 3067 |
+
{
|
| 3068 |
+
"epoch": 0.788275987620608,
|
| 3069 |
+
"grad_norm": 5.493254661560059,
|
| 3070 |
+
"learning_rate": 1.2799985794589497e-05,
|
| 3071 |
+
"loss": 2.3842,
|
| 3072 |
+
"step": 4330
|
| 3073 |
+
},
|
| 3074 |
+
{
|
| 3075 |
+
"epoch": 0.7900964864372838,
|
| 3076 |
+
"grad_norm": 5.259057521820068,
|
| 3077 |
+
"learning_rate": 1.2590514527667336e-05,
|
| 3078 |
+
"loss": 2.3783,
|
| 3079 |
+
"step": 4340
|
| 3080 |
+
},
|
| 3081 |
+
{
|
| 3082 |
+
"epoch": 0.7919169852539596,
|
| 3083 |
+
"grad_norm": 5.750987529754639,
|
| 3084 |
+
"learning_rate": 1.2382524267077645e-05,
|
| 3085 |
+
"loss": 2.4202,
|
| 3086 |
+
"step": 4350
|
| 3087 |
+
},
|
| 3088 |
+
{
|
| 3089 |
+
"epoch": 0.7937374840706354,
|
| 3090 |
+
"grad_norm": 4.952456951141357,
|
| 3091 |
+
"learning_rate": 1.2176023246959133e-05,
|
| 3092 |
+
"loss": 2.4393,
|
| 3093 |
+
"step": 4360
|
| 3094 |
+
},
|
| 3095 |
+
{
|
| 3096 |
+
"epoch": 0.7955579828873112,
|
| 3097 |
+
"grad_norm": 5.3008713722229,
|
| 3098 |
+
"learning_rate": 1.1971019642492942e-05,
|
| 3099 |
+
"loss": 2.375,
|
| 3100 |
+
"step": 4370
|
| 3101 |
+
},
|
| 3102 |
+
{
|
| 3103 |
+
"epoch": 0.7973784817039868,
|
| 3104 |
+
"grad_norm": 4.872366428375244,
|
| 3105 |
+
"learning_rate": 1.176752156957886e-05,
|
| 3106 |
+
"loss": 2.4257,
|
| 3107 |
+
"step": 4380
|
| 3108 |
+
},
|
| 3109 |
+
{
|
| 3110 |
+
"epoch": 0.7991989805206626,
|
| 3111 |
+
"grad_norm": 5.488797664642334,
|
| 3112 |
+
"learning_rate": 1.1565537084514123e-05,
|
| 3113 |
+
"loss": 2.4424,
|
| 3114 |
+
"step": 4390
|
| 3115 |
+
},
|
| 3116 |
+
{
|
| 3117 |
+
"epoch": 0.8010194793373384,
|
| 3118 |
+
"grad_norm": 5.145867824554443,
|
| 3119 |
+
"learning_rate": 1.1365074183674468e-05,
|
| 3120 |
+
"loss": 2.4806,
|
| 3121 |
+
"step": 4400
|
| 3122 |
+
},
|
| 3123 |
+
{
|
| 3124 |
+
"epoch": 0.8028399781540142,
|
| 3125 |
+
"grad_norm": 5.343238353729248,
|
| 3126 |
+
"learning_rate": 1.116614080319754e-05,
|
| 3127 |
+
"loss": 2.4321,
|
| 3128 |
+
"step": 4410
|
| 3129 |
+
},
|
| 3130 |
+
{
|
| 3131 |
+
"epoch": 0.80466047697069,
|
| 3132 |
+
"grad_norm": 5.240965366363525,
|
| 3133 |
+
"learning_rate": 1.0968744818668691e-05,
|
| 3134 |
+
"loss": 2.4358,
|
| 3135 |
+
"step": 4420
|
| 3136 |
+
},
|
| 3137 |
+
{
|
| 3138 |
+
"epoch": 0.8064809757873658,
|
| 3139 |
+
"grad_norm": 5.5220513343811035,
|
| 3140 |
+
"learning_rate": 1.0772894044809229e-05,
|
| 3141 |
+
"loss": 2.442,
|
| 3142 |
+
"step": 4430
|
| 3143 |
+
},
|
| 3144 |
+
{
|
| 3145 |
+
"epoch": 0.8083014746040416,
|
| 3146 |
+
"grad_norm": 4.8629045486450195,
|
| 3147 |
+
"learning_rate": 1.0578596235166998e-05,
|
| 3148 |
+
"loss": 2.4567,
|
| 3149 |
+
"step": 4440
|
| 3150 |
+
},
|
| 3151 |
+
{
|
| 3152 |
+
"epoch": 0.8101219734207172,
|
| 3153 |
+
"grad_norm": 5.297680854797363,
|
| 3154 |
+
"learning_rate": 1.0385859081809508e-05,
|
| 3155 |
+
"loss": 2.4544,
|
| 3156 |
+
"step": 4450
|
| 3157 |
+
},
|
| 3158 |
+
{
|
| 3159 |
+
"epoch": 0.811942472237393,
|
| 3160 |
+
"grad_norm": 5.134615898132324,
|
| 3161 |
+
"learning_rate": 1.0194690215019292e-05,
|
| 3162 |
+
"loss": 2.4656,
|
| 3163 |
+
"step": 4460
|
| 3164 |
+
},
|
| 3165 |
+
{
|
| 3166 |
+
"epoch": 0.8137629710540688,
|
| 3167 |
+
"grad_norm": 5.012113571166992,
|
| 3168 |
+
"learning_rate": 1.0005097202991948e-05,
|
| 3169 |
+
"loss": 2.382,
|
| 3170 |
+
"step": 4470
|
| 3171 |
+
},
|
| 3172 |
+
{
|
| 3173 |
+
"epoch": 0.8155834698707446,
|
| 3174 |
+
"grad_norm": 5.369142532348633,
|
| 3175 |
+
"learning_rate": 9.817087551536414e-06,
|
| 3176 |
+
"loss": 2.4584,
|
| 3177 |
+
"step": 4480
|
| 3178 |
+
},
|
| 3179 |
+
{
|
| 3180 |
+
"epoch": 0.8174039686874204,
|
| 3181 |
+
"grad_norm": 5.545107841491699,
|
| 3182 |
+
"learning_rate": 9.630668703777922e-06,
|
| 3183 |
+
"loss": 2.4013,
|
| 3184 |
+
"step": 4490
|
| 3185 |
+
},
|
| 3186 |
+
{
|
| 3187 |
+
"epoch": 0.8192244675040962,
|
| 3188 |
+
"grad_norm": 4.933434963226318,
|
| 3189 |
+
"learning_rate": 9.445848039863252e-06,
|
| 3190 |
+
"loss": 2.4516,
|
| 3191 |
+
"step": 4500
|
| 3192 |
}
|
| 3193 |
],
|
| 3194 |
"logging_steps": 10,
|
|
|
|
| 3208 |
"attributes": {}
|
| 3209 |
}
|
| 3210 |
},
|
| 3211 |
+
"total_flos": 1.052165096865792e+18,
|
| 3212 |
"train_batch_size": 4,
|
| 3213 |
"trial_name": null,
|
| 3214 |
"trial_params": null
|