Training in progress, step 4000, checkpoint
Browse files
last-checkpoint/model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 737580392
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:92c7c2466e9547634a505ccaf2590f9e4d9d15d2f31d94aa4c0cfe5f155dc10b
|
| 3 |
size 737580392
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1475248442
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:05b07ba25a847922c59fe9c0ee222039fd2b55eb27e7164ec80572760094d906
|
| 3 |
size 1475248442
|
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 sha256:3b8a0379bdd10765d4926325b17779ba084884beedfbdf271680e1d1bd136b43
|
| 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 1000
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ea8c56402fe28ab6610db127ee707a0d7bbb7e8371ebb7f77b59566a41c7f5ef
|
| 3 |
size 1000
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch": 1.
|
| 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,
|
|
@@ -3017,6 +3017,436 @@
|
|
| 3017 |
"eval_spearman_manhattan": 0.8222685344671697,
|
| 3018 |
"eval_steps_per_second": 15.727,
|
| 3019 |
"step": 3500
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3020 |
}
|
| 3021 |
],
|
| 3022 |
"logging_steps": 10,
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.8744142455482662,
|
| 5 |
"eval_steps": 100,
|
| 6 |
+
"global_step": 4000,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 3017 |
"eval_spearman_manhattan": 0.8222685344671697,
|
| 3018 |
"eval_steps_per_second": 15.727,
|
| 3019 |
"step": 3500
|
| 3020 |
+
},
|
| 3021 |
+
{
|
| 3022 |
+
"epoch": 1.6447985004686037,
|
| 3023 |
+
"grad_norm": 0.9714635610580444,
|
| 3024 |
+
"learning_rate": 4.486000468603562e-05,
|
| 3025 |
+
"loss": 0.1841,
|
| 3026 |
+
"step": 3510
|
| 3027 |
+
},
|
| 3028 |
+
{
|
| 3029 |
+
"epoch": 1.6494845360824741,
|
| 3030 |
+
"grad_norm": 1.4874918460845947,
|
| 3031 |
+
"learning_rate": 4.484536082474227e-05,
|
| 3032 |
+
"loss": 0.1833,
|
| 3033 |
+
"step": 3520
|
| 3034 |
+
},
|
| 3035 |
+
{
|
| 3036 |
+
"epoch": 1.6541705716963448,
|
| 3037 |
+
"grad_norm": 1.3226478099822998,
|
| 3038 |
+
"learning_rate": 4.4830716963448926e-05,
|
| 3039 |
+
"loss": 0.1946,
|
| 3040 |
+
"step": 3530
|
| 3041 |
+
},
|
| 3042 |
+
{
|
| 3043 |
+
"epoch": 1.6588566073102156,
|
| 3044 |
+
"grad_norm": 1.6590946912765503,
|
| 3045 |
+
"learning_rate": 4.481607310215558e-05,
|
| 3046 |
+
"loss": 0.192,
|
| 3047 |
+
"step": 3540
|
| 3048 |
+
},
|
| 3049 |
+
{
|
| 3050 |
+
"epoch": 1.6635426429240863,
|
| 3051 |
+
"grad_norm": 1.8779162168502808,
|
| 3052 |
+
"learning_rate": 4.480142924086223e-05,
|
| 3053 |
+
"loss": 0.1841,
|
| 3054 |
+
"step": 3550
|
| 3055 |
+
},
|
| 3056 |
+
{
|
| 3057 |
+
"epoch": 1.668228678537957,
|
| 3058 |
+
"grad_norm": 1.6211146116256714,
|
| 3059 |
+
"learning_rate": 4.478678537956888e-05,
|
| 3060 |
+
"loss": 0.1943,
|
| 3061 |
+
"step": 3560
|
| 3062 |
+
},
|
| 3063 |
+
{
|
| 3064 |
+
"epoch": 1.6729147141518275,
|
| 3065 |
+
"grad_norm": 1.4703700542449951,
|
| 3066 |
+
"learning_rate": 4.477214151827554e-05,
|
| 3067 |
+
"loss": 0.1728,
|
| 3068 |
+
"step": 3570
|
| 3069 |
+
},
|
| 3070 |
+
{
|
| 3071 |
+
"epoch": 1.6776007497656982,
|
| 3072 |
+
"grad_norm": 1.689462661743164,
|
| 3073 |
+
"learning_rate": 4.47574976569822e-05,
|
| 3074 |
+
"loss": 0.1857,
|
| 3075 |
+
"step": 3580
|
| 3076 |
+
},
|
| 3077 |
+
{
|
| 3078 |
+
"epoch": 1.6822867853795689,
|
| 3079 |
+
"grad_norm": 1.7197692394256592,
|
| 3080 |
+
"learning_rate": 4.474285379568885e-05,
|
| 3081 |
+
"loss": 0.2084,
|
| 3082 |
+
"step": 3590
|
| 3083 |
+
},
|
| 3084 |
+
{
|
| 3085 |
+
"epoch": 1.6869728209934396,
|
| 3086 |
+
"grad_norm": 1.2906155586242676,
|
| 3087 |
+
"learning_rate": 4.4728209934395506e-05,
|
| 3088 |
+
"loss": 0.1982,
|
| 3089 |
+
"step": 3600
|
| 3090 |
+
},
|
| 3091 |
+
{
|
| 3092 |
+
"epoch": 1.6869728209934396,
|
| 3093 |
+
"eval_loss": 0.03799282759428024,
|
| 3094 |
+
"eval_pearson_cosine": 0.8220264449208883,
|
| 3095 |
+
"eval_pearson_dot": 0.7551673016916052,
|
| 3096 |
+
"eval_pearson_euclidean": 0.818234030509938,
|
| 3097 |
+
"eval_pearson_manhattan": 0.8196075002869208,
|
| 3098 |
+
"eval_runtime": 5.9912,
|
| 3099 |
+
"eval_samples_per_second": 250.368,
|
| 3100 |
+
"eval_spearman_cosine": 0.8233208131275441,
|
| 3101 |
+
"eval_spearman_dot": 0.7535137341757772,
|
| 3102 |
+
"eval_spearman_euclidean": 0.8248668260426036,
|
| 3103 |
+
"eval_spearman_manhattan": 0.825693429330301,
|
| 3104 |
+
"eval_steps_per_second": 15.69,
|
| 3105 |
+
"step": 3600
|
| 3106 |
+
},
|
| 3107 |
+
{
|
| 3108 |
+
"epoch": 1.6916588566073103,
|
| 3109 |
+
"grad_norm": 1.595618724822998,
|
| 3110 |
+
"learning_rate": 4.471356607310216e-05,
|
| 3111 |
+
"loss": 0.1828,
|
| 3112 |
+
"step": 3610
|
| 3113 |
+
},
|
| 3114 |
+
{
|
| 3115 |
+
"epoch": 1.6963448922211808,
|
| 3116 |
+
"grad_norm": 1.280032753944397,
|
| 3117 |
+
"learning_rate": 4.4698922211808815e-05,
|
| 3118 |
+
"loss": 0.1748,
|
| 3119 |
+
"step": 3620
|
| 3120 |
+
},
|
| 3121 |
+
{
|
| 3122 |
+
"epoch": 1.7010309278350515,
|
| 3123 |
+
"grad_norm": 1.2117244005203247,
|
| 3124 |
+
"learning_rate": 4.468427835051547e-05,
|
| 3125 |
+
"loss": 0.1758,
|
| 3126 |
+
"step": 3630
|
| 3127 |
+
},
|
| 3128 |
+
{
|
| 3129 |
+
"epoch": 1.7057169634489222,
|
| 3130 |
+
"grad_norm": 1.7422757148742676,
|
| 3131 |
+
"learning_rate": 4.466963448922212e-05,
|
| 3132 |
+
"loss": 0.17,
|
| 3133 |
+
"step": 3640
|
| 3134 |
+
},
|
| 3135 |
+
{
|
| 3136 |
+
"epoch": 1.710402999062793,
|
| 3137 |
+
"grad_norm": 1.6089109182357788,
|
| 3138 |
+
"learning_rate": 4.465499062792877e-05,
|
| 3139 |
+
"loss": 0.1745,
|
| 3140 |
+
"step": 3650
|
| 3141 |
+
},
|
| 3142 |
+
{
|
| 3143 |
+
"epoch": 1.7150890346766636,
|
| 3144 |
+
"grad_norm": 1.664401650428772,
|
| 3145 |
+
"learning_rate": 4.4640346766635425e-05,
|
| 3146 |
+
"loss": 0.153,
|
| 3147 |
+
"step": 3660
|
| 3148 |
+
},
|
| 3149 |
+
{
|
| 3150 |
+
"epoch": 1.7197750702905341,
|
| 3151 |
+
"grad_norm": 1.5395989418029785,
|
| 3152 |
+
"learning_rate": 4.462570290534208e-05,
|
| 3153 |
+
"loss": 0.2049,
|
| 3154 |
+
"step": 3670
|
| 3155 |
+
},
|
| 3156 |
+
{
|
| 3157 |
+
"epoch": 1.7244611059044048,
|
| 3158 |
+
"grad_norm": 1.8465495109558105,
|
| 3159 |
+
"learning_rate": 4.4611059044048734e-05,
|
| 3160 |
+
"loss": 0.208,
|
| 3161 |
+
"step": 3680
|
| 3162 |
+
},
|
| 3163 |
+
{
|
| 3164 |
+
"epoch": 1.7291471415182755,
|
| 3165 |
+
"grad_norm": 1.6127640008926392,
|
| 3166 |
+
"learning_rate": 4.459641518275539e-05,
|
| 3167 |
+
"loss": 0.1732,
|
| 3168 |
+
"step": 3690
|
| 3169 |
+
},
|
| 3170 |
+
{
|
| 3171 |
+
"epoch": 1.7338331771321462,
|
| 3172 |
+
"grad_norm": 1.2604060173034668,
|
| 3173 |
+
"learning_rate": 4.458177132146205e-05,
|
| 3174 |
+
"loss": 0.1824,
|
| 3175 |
+
"step": 3700
|
| 3176 |
+
},
|
| 3177 |
+
{
|
| 3178 |
+
"epoch": 1.7338331771321462,
|
| 3179 |
+
"eval_loss": 0.03523562103509903,
|
| 3180 |
+
"eval_pearson_cosine": 0.8246188621544412,
|
| 3181 |
+
"eval_pearson_dot": 0.7566728482844667,
|
| 3182 |
+
"eval_pearson_euclidean": 0.816607137727027,
|
| 3183 |
+
"eval_pearson_manhattan": 0.8181353811719987,
|
| 3184 |
+
"eval_runtime": 6.4257,
|
| 3185 |
+
"eval_samples_per_second": 233.438,
|
| 3186 |
+
"eval_spearman_cosine": 0.8251725549021353,
|
| 3187 |
+
"eval_spearman_dot": 0.7553705266499796,
|
| 3188 |
+
"eval_spearman_euclidean": 0.8232824851511874,
|
| 3189 |
+
"eval_spearman_manhattan": 0.8242087634144919,
|
| 3190 |
+
"eval_steps_per_second": 14.629,
|
| 3191 |
+
"step": 3700
|
| 3192 |
+
},
|
| 3193 |
+
{
|
| 3194 |
+
"epoch": 1.738519212746017,
|
| 3195 |
+
"grad_norm": 1.4244945049285889,
|
| 3196 |
+
"learning_rate": 4.45671274601687e-05,
|
| 3197 |
+
"loss": 0.1704,
|
| 3198 |
+
"step": 3710
|
| 3199 |
+
},
|
| 3200 |
+
{
|
| 3201 |
+
"epoch": 1.7432052483598874,
|
| 3202 |
+
"grad_norm": 1.5206220149993896,
|
| 3203 |
+
"learning_rate": 4.455248359887536e-05,
|
| 3204 |
+
"loss": 0.1921,
|
| 3205 |
+
"step": 3720
|
| 3206 |
+
},
|
| 3207 |
+
{
|
| 3208 |
+
"epoch": 1.7478912839737581,
|
| 3209 |
+
"grad_norm": 1.596449375152588,
|
| 3210 |
+
"learning_rate": 4.453783973758201e-05,
|
| 3211 |
+
"loss": 0.1714,
|
| 3212 |
+
"step": 3730
|
| 3213 |
+
},
|
| 3214 |
+
{
|
| 3215 |
+
"epoch": 1.7525773195876289,
|
| 3216 |
+
"grad_norm": 1.6774275302886963,
|
| 3217 |
+
"learning_rate": 4.452319587628866e-05,
|
| 3218 |
+
"loss": 0.1889,
|
| 3219 |
+
"step": 3740
|
| 3220 |
+
},
|
| 3221 |
+
{
|
| 3222 |
+
"epoch": 1.7572633552014996,
|
| 3223 |
+
"grad_norm": 1.4467116594314575,
|
| 3224 |
+
"learning_rate": 4.4508552014995314e-05,
|
| 3225 |
+
"loss": 0.2031,
|
| 3226 |
+
"step": 3750
|
| 3227 |
+
},
|
| 3228 |
+
{
|
| 3229 |
+
"epoch": 1.7619493908153703,
|
| 3230 |
+
"grad_norm": 1.722493290901184,
|
| 3231 |
+
"learning_rate": 4.449390815370197e-05,
|
| 3232 |
+
"loss": 0.1931,
|
| 3233 |
+
"step": 3760
|
| 3234 |
+
},
|
| 3235 |
+
{
|
| 3236 |
+
"epoch": 1.7666354264292408,
|
| 3237 |
+
"grad_norm": 1.8555185794830322,
|
| 3238 |
+
"learning_rate": 4.447926429240862e-05,
|
| 3239 |
+
"loss": 0.211,
|
| 3240 |
+
"step": 3770
|
| 3241 |
+
},
|
| 3242 |
+
{
|
| 3243 |
+
"epoch": 1.7713214620431117,
|
| 3244 |
+
"grad_norm": 1.5005462169647217,
|
| 3245 |
+
"learning_rate": 4.4464620431115277e-05,
|
| 3246 |
+
"loss": 0.1939,
|
| 3247 |
+
"step": 3780
|
| 3248 |
+
},
|
| 3249 |
+
{
|
| 3250 |
+
"epoch": 1.7760074976569822,
|
| 3251 |
+
"grad_norm": 1.88517165184021,
|
| 3252 |
+
"learning_rate": 4.444997656982193e-05,
|
| 3253 |
+
"loss": 0.1868,
|
| 3254 |
+
"step": 3790
|
| 3255 |
+
},
|
| 3256 |
+
{
|
| 3257 |
+
"epoch": 1.780693533270853,
|
| 3258 |
+
"grad_norm": 1.7735782861709595,
|
| 3259 |
+
"learning_rate": 4.4435332708528585e-05,
|
| 3260 |
+
"loss": 0.2009,
|
| 3261 |
+
"step": 3800
|
| 3262 |
+
},
|
| 3263 |
+
{
|
| 3264 |
+
"epoch": 1.780693533270853,
|
| 3265 |
+
"eval_loss": 0.03577423095703125,
|
| 3266 |
+
"eval_pearson_cosine": 0.8269622117688868,
|
| 3267 |
+
"eval_pearson_dot": 0.7668511969090961,
|
| 3268 |
+
"eval_pearson_euclidean": 0.8089875954267027,
|
| 3269 |
+
"eval_pearson_manhattan": 0.8104624503351374,
|
| 3270 |
+
"eval_runtime": 6.3923,
|
| 3271 |
+
"eval_samples_per_second": 234.659,
|
| 3272 |
+
"eval_spearman_cosine": 0.8277880382919593,
|
| 3273 |
+
"eval_spearman_dot": 0.7654651690867051,
|
| 3274 |
+
"eval_spearman_euclidean": 0.8164441074213089,
|
| 3275 |
+
"eval_spearman_manhattan": 0.8181458418663312,
|
| 3276 |
+
"eval_steps_per_second": 14.705,
|
| 3277 |
+
"step": 3800
|
| 3278 |
+
},
|
| 3279 |
+
{
|
| 3280 |
+
"epoch": 1.7853795688847236,
|
| 3281 |
+
"grad_norm": 1.697729229927063,
|
| 3282 |
+
"learning_rate": 4.442068884723524e-05,
|
| 3283 |
+
"loss": 0.1902,
|
| 3284 |
+
"step": 3810
|
| 3285 |
+
},
|
| 3286 |
+
{
|
| 3287 |
+
"epoch": 1.790065604498594,
|
| 3288 |
+
"grad_norm": 1.3950953483581543,
|
| 3289 |
+
"learning_rate": 4.44060449859419e-05,
|
| 3290 |
+
"loss": 0.1937,
|
| 3291 |
+
"step": 3820
|
| 3292 |
+
},
|
| 3293 |
+
{
|
| 3294 |
+
"epoch": 1.794751640112465,
|
| 3295 |
+
"grad_norm": 2.385718584060669,
|
| 3296 |
+
"learning_rate": 4.4391401124648555e-05,
|
| 3297 |
+
"loss": 0.2259,
|
| 3298 |
+
"step": 3830
|
| 3299 |
+
},
|
| 3300 |
+
{
|
| 3301 |
+
"epoch": 1.7994376757263355,
|
| 3302 |
+
"grad_norm": 1.0628514289855957,
|
| 3303 |
+
"learning_rate": 4.43767572633552e-05,
|
| 3304 |
+
"loss": 0.1611,
|
| 3305 |
+
"step": 3840
|
| 3306 |
+
},
|
| 3307 |
+
{
|
| 3308 |
+
"epoch": 1.8041237113402062,
|
| 3309 |
+
"grad_norm": 1.6918362379074097,
|
| 3310 |
+
"learning_rate": 4.4362113402061856e-05,
|
| 3311 |
+
"loss": 0.1767,
|
| 3312 |
+
"step": 3850
|
| 3313 |
+
},
|
| 3314 |
+
{
|
| 3315 |
+
"epoch": 1.808809746954077,
|
| 3316 |
+
"grad_norm": 1.2459661960601807,
|
| 3317 |
+
"learning_rate": 4.434746954076851e-05,
|
| 3318 |
+
"loss": 0.1477,
|
| 3319 |
+
"step": 3860
|
| 3320 |
+
},
|
| 3321 |
+
{
|
| 3322 |
+
"epoch": 1.8134957825679474,
|
| 3323 |
+
"grad_norm": 1.2925353050231934,
|
| 3324 |
+
"learning_rate": 4.4332825679475165e-05,
|
| 3325 |
+
"loss": 0.1848,
|
| 3326 |
+
"step": 3870
|
| 3327 |
+
},
|
| 3328 |
+
{
|
| 3329 |
+
"epoch": 1.8181818181818183,
|
| 3330 |
+
"grad_norm": 1.3787219524383545,
|
| 3331 |
+
"learning_rate": 4.431818181818182e-05,
|
| 3332 |
+
"loss": 0.2216,
|
| 3333 |
+
"step": 3880
|
| 3334 |
+
},
|
| 3335 |
+
{
|
| 3336 |
+
"epoch": 1.8228678537956888,
|
| 3337 |
+
"grad_norm": 1.531141996383667,
|
| 3338 |
+
"learning_rate": 4.4303537956888473e-05,
|
| 3339 |
+
"loss": 0.2087,
|
| 3340 |
+
"step": 3890
|
| 3341 |
+
},
|
| 3342 |
+
{
|
| 3343 |
+
"epoch": 1.8275538894095595,
|
| 3344 |
+
"grad_norm": 1.4053128957748413,
|
| 3345 |
+
"learning_rate": 4.428889409559513e-05,
|
| 3346 |
+
"loss": 0.1899,
|
| 3347 |
+
"step": 3900
|
| 3348 |
+
},
|
| 3349 |
+
{
|
| 3350 |
+
"epoch": 1.8275538894095595,
|
| 3351 |
+
"eval_loss": 0.03851017728447914,
|
| 3352 |
+
"eval_pearson_cosine": 0.8239765523882259,
|
| 3353 |
+
"eval_pearson_dot": 0.7418402003946696,
|
| 3354 |
+
"eval_pearson_euclidean": 0.8110650670007509,
|
| 3355 |
+
"eval_pearson_manhattan": 0.8133011255650899,
|
| 3356 |
+
"eval_runtime": 6.2641,
|
| 3357 |
+
"eval_samples_per_second": 239.46,
|
| 3358 |
+
"eval_spearman_cosine": 0.8252211874177356,
|
| 3359 |
+
"eval_spearman_dot": 0.7383247864885149,
|
| 3360 |
+
"eval_spearman_euclidean": 0.8180097764566381,
|
| 3361 |
+
"eval_spearman_manhattan": 0.8202301926079525,
|
| 3362 |
+
"eval_steps_per_second": 15.006,
|
| 3363 |
+
"step": 3900
|
| 3364 |
+
},
|
| 3365 |
+
{
|
| 3366 |
+
"epoch": 1.8322399250234302,
|
| 3367 |
+
"grad_norm": 1.564635157585144,
|
| 3368 |
+
"learning_rate": 4.427425023430178e-05,
|
| 3369 |
+
"loss": 0.1912,
|
| 3370 |
+
"step": 3910
|
| 3371 |
+
},
|
| 3372 |
+
{
|
| 3373 |
+
"epoch": 1.8369259606373007,
|
| 3374 |
+
"grad_norm": 1.525820255279541,
|
| 3375 |
+
"learning_rate": 4.4259606373008436e-05,
|
| 3376 |
+
"loss": 0.1725,
|
| 3377 |
+
"step": 3920
|
| 3378 |
+
},
|
| 3379 |
+
{
|
| 3380 |
+
"epoch": 1.8416119962511717,
|
| 3381 |
+
"grad_norm": 1.6183199882507324,
|
| 3382 |
+
"learning_rate": 4.424496251171509e-05,
|
| 3383 |
+
"loss": 0.1771,
|
| 3384 |
+
"step": 3930
|
| 3385 |
+
},
|
| 3386 |
+
{
|
| 3387 |
+
"epoch": 1.8462980318650422,
|
| 3388 |
+
"grad_norm": 1.614140510559082,
|
| 3389 |
+
"learning_rate": 4.4230318650421745e-05,
|
| 3390 |
+
"loss": 0.2085,
|
| 3391 |
+
"step": 3940
|
| 3392 |
+
},
|
| 3393 |
+
{
|
| 3394 |
+
"epoch": 1.8509840674789129,
|
| 3395 |
+
"grad_norm": 1.6187098026275635,
|
| 3396 |
+
"learning_rate": 4.42156747891284e-05,
|
| 3397 |
+
"loss": 0.1841,
|
| 3398 |
+
"step": 3950
|
| 3399 |
+
},
|
| 3400 |
+
{
|
| 3401 |
+
"epoch": 1.8556701030927836,
|
| 3402 |
+
"grad_norm": 1.9337131977081299,
|
| 3403 |
+
"learning_rate": 4.4201030927835053e-05,
|
| 3404 |
+
"loss": 0.1814,
|
| 3405 |
+
"step": 3960
|
| 3406 |
+
},
|
| 3407 |
+
{
|
| 3408 |
+
"epoch": 1.860356138706654,
|
| 3409 |
+
"grad_norm": 1.4200247526168823,
|
| 3410 |
+
"learning_rate": 4.418638706654171e-05,
|
| 3411 |
+
"loss": 0.2138,
|
| 3412 |
+
"step": 3970
|
| 3413 |
+
},
|
| 3414 |
+
{
|
| 3415 |
+
"epoch": 1.865042174320525,
|
| 3416 |
+
"grad_norm": 1.2319873571395874,
|
| 3417 |
+
"learning_rate": 4.417174320524836e-05,
|
| 3418 |
+
"loss": 0.1811,
|
| 3419 |
+
"step": 3980
|
| 3420 |
+
},
|
| 3421 |
+
{
|
| 3422 |
+
"epoch": 1.8697282099343955,
|
| 3423 |
+
"grad_norm": 1.5090093612670898,
|
| 3424 |
+
"learning_rate": 4.4157099343955016e-05,
|
| 3425 |
+
"loss": 0.2013,
|
| 3426 |
+
"step": 3990
|
| 3427 |
+
},
|
| 3428 |
+
{
|
| 3429 |
+
"epoch": 1.8744142455482662,
|
| 3430 |
+
"grad_norm": 1.4747837781906128,
|
| 3431 |
+
"learning_rate": 4.414245548266167e-05,
|
| 3432 |
+
"loss": 0.1858,
|
| 3433 |
+
"step": 4000
|
| 3434 |
+
},
|
| 3435 |
+
{
|
| 3436 |
+
"epoch": 1.8744142455482662,
|
| 3437 |
+
"eval_loss": 0.03368546813726425,
|
| 3438 |
+
"eval_pearson_cosine": 0.8281146927252365,
|
| 3439 |
+
"eval_pearson_dot": 0.761976509126896,
|
| 3440 |
+
"eval_pearson_euclidean": 0.810243447652832,
|
| 3441 |
+
"eval_pearson_manhattan": 0.8121517436135477,
|
| 3442 |
+
"eval_runtime": 7.0014,
|
| 3443 |
+
"eval_samples_per_second": 214.242,
|
| 3444 |
+
"eval_spearman_cosine": 0.8273786901092443,
|
| 3445 |
+
"eval_spearman_dot": 0.7590219133940205,
|
| 3446 |
+
"eval_spearman_euclidean": 0.8180153411946929,
|
| 3447 |
+
"eval_spearman_manhattan": 0.8198041702608989,
|
| 3448 |
+
"eval_steps_per_second": 13.426,
|
| 3449 |
+
"step": 4000
|
| 3450 |
}
|
| 3451 |
],
|
| 3452 |
"logging_steps": 10,
|