Simin Fan
commited on
Commit
•
e4d1af0
1
Parent(s):
87daf46
Upload trainer_state.json
Browse files- trainer_state.json +256 -0
trainer_state.json
ADDED
|
@@ -0,0 +1,256 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 11.580775911986104,
|
| 5 |
+
"global_step": 40000,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.29,
|
| 12 |
+
"learning_rate": 6e-06,
|
| 13 |
+
"loss": 4.3587,
|
| 14 |
+
"step": 1000
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.58,
|
| 18 |
+
"learning_rate": 1.2e-05,
|
| 19 |
+
"loss": 3.473,
|
| 20 |
+
"step": 2000
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.87,
|
| 24 |
+
"learning_rate": 1.4807435555099107e-05,
|
| 25 |
+
"loss": 3.257,
|
| 26 |
+
"step": 3000
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 1.16,
|
| 30 |
+
"learning_rate": 1.442230666529732e-05,
|
| 31 |
+
"loss": 3.0935,
|
| 32 |
+
"step": 4000
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 1.45,
|
| 36 |
+
"learning_rate": 1.4037177775495533e-05,
|
| 37 |
+
"loss": 3.0049,
|
| 38 |
+
"step": 5000
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 1.74,
|
| 42 |
+
"learning_rate": 1.3652048885693745e-05,
|
| 43 |
+
"loss": 2.932,
|
| 44 |
+
"step": 6000
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 2.03,
|
| 48 |
+
"learning_rate": 1.326691999589196e-05,
|
| 49 |
+
"loss": 2.8843,
|
| 50 |
+
"step": 7000
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 2.32,
|
| 54 |
+
"learning_rate": 1.2881791106090172e-05,
|
| 55 |
+
"loss": 2.7991,
|
| 56 |
+
"step": 8000
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 2.61,
|
| 60 |
+
"learning_rate": 1.2496662216288385e-05,
|
| 61 |
+
"loss": 2.7699,
|
| 62 |
+
"step": 9000
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 2.9,
|
| 66 |
+
"learning_rate": 1.2111533326486598e-05,
|
| 67 |
+
"loss": 2.7408,
|
| 68 |
+
"step": 10000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 3.18,
|
| 72 |
+
"learning_rate": 1.172640443668481e-05,
|
| 73 |
+
"loss": 2.6829,
|
| 74 |
+
"step": 11000
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 3.47,
|
| 78 |
+
"learning_rate": 1.1341275546883024e-05,
|
| 79 |
+
"loss": 2.6527,
|
| 80 |
+
"step": 12000
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 3.76,
|
| 84 |
+
"learning_rate": 1.0956146657081237e-05,
|
| 85 |
+
"loss": 2.6339,
|
| 86 |
+
"step": 13000
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 4.05,
|
| 90 |
+
"learning_rate": 1.057101776727945e-05,
|
| 91 |
+
"loss": 2.6152,
|
| 92 |
+
"step": 14000
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 4.34,
|
| 96 |
+
"learning_rate": 1.0185888877477662e-05,
|
| 97 |
+
"loss": 2.5695,
|
| 98 |
+
"step": 15000
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 4.63,
|
| 102 |
+
"learning_rate": 9.800759987675876e-06,
|
| 103 |
+
"loss": 2.5439,
|
| 104 |
+
"step": 16000
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 4.92,
|
| 108 |
+
"learning_rate": 9.41563109787409e-06,
|
| 109 |
+
"loss": 2.5504,
|
| 110 |
+
"step": 17000
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 5.21,
|
| 114 |
+
"learning_rate": 9.030502208072302e-06,
|
| 115 |
+
"loss": 2.5106,
|
| 116 |
+
"step": 18000
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 5.5,
|
| 120 |
+
"learning_rate": 8.645373318270514e-06,
|
| 121 |
+
"loss": 2.4734,
|
| 122 |
+
"step": 19000
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 5.79,
|
| 126 |
+
"learning_rate": 8.260244428468728e-06,
|
| 127 |
+
"loss": 2.4931,
|
| 128 |
+
"step": 20000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 6.08,
|
| 132 |
+
"learning_rate": 7.875115538666941e-06,
|
| 133 |
+
"loss": 2.4626,
|
| 134 |
+
"step": 21000
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 6.37,
|
| 138 |
+
"learning_rate": 7.489986648865154e-06,
|
| 139 |
+
"loss": 2.4341,
|
| 140 |
+
"step": 22000
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 6.66,
|
| 144 |
+
"learning_rate": 7.104857759063367e-06,
|
| 145 |
+
"loss": 2.4316,
|
| 146 |
+
"step": 23000
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 6.95,
|
| 150 |
+
"learning_rate": 6.71972886926158e-06,
|
| 151 |
+
"loss": 2.4158,
|
| 152 |
+
"step": 24000
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 7.24,
|
| 156 |
+
"learning_rate": 6.334599979459793e-06,
|
| 157 |
+
"loss": 2.4055,
|
| 158 |
+
"step": 25000
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 7.53,
|
| 162 |
+
"learning_rate": 5.949471089658006e-06,
|
| 163 |
+
"loss": 2.3773,
|
| 164 |
+
"step": 26000
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 7.82,
|
| 168 |
+
"learning_rate": 5.564342199856218e-06,
|
| 169 |
+
"loss": 2.3877,
|
| 170 |
+
"step": 27000
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 8.11,
|
| 174 |
+
"learning_rate": 5.179213310054432e-06,
|
| 175 |
+
"loss": 2.3826,
|
| 176 |
+
"step": 28000
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 8.4,
|
| 180 |
+
"learning_rate": 4.794084420252645e-06,
|
| 181 |
+
"loss": 2.3419,
|
| 182 |
+
"step": 29000
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 8.69,
|
| 186 |
+
"learning_rate": 4.408955530450858e-06,
|
| 187 |
+
"loss": 2.3447,
|
| 188 |
+
"step": 30000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 8.98,
|
| 192 |
+
"learning_rate": 4.023826640649071e-06,
|
| 193 |
+
"loss": 2.3602,
|
| 194 |
+
"step": 31000
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 9.26,
|
| 198 |
+
"learning_rate": 3.6386977508472838e-06,
|
| 199 |
+
"loss": 2.3315,
|
| 200 |
+
"step": 32000
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 9.55,
|
| 204 |
+
"learning_rate": 3.2535688610454968e-06,
|
| 205 |
+
"loss": 2.3116,
|
| 206 |
+
"step": 33000
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 9.84,
|
| 210 |
+
"learning_rate": 2.8684399712437098e-06,
|
| 211 |
+
"loss": 2.3285,
|
| 212 |
+
"step": 34000
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 10.13,
|
| 216 |
+
"learning_rate": 2.4833110814419224e-06,
|
| 217 |
+
"loss": 2.3151,
|
| 218 |
+
"step": 35000
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 10.42,
|
| 222 |
+
"learning_rate": 2.0981821916401354e-06,
|
| 223 |
+
"loss": 2.2991,
|
| 224 |
+
"step": 36000
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 10.71,
|
| 228 |
+
"learning_rate": 1.7130533018383486e-06,
|
| 229 |
+
"loss": 2.2984,
|
| 230 |
+
"step": 37000
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 11.0,
|
| 234 |
+
"learning_rate": 1.3279244120365616e-06,
|
| 235 |
+
"loss": 2.31,
|
| 236 |
+
"step": 38000
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 11.29,
|
| 240 |
+
"learning_rate": 9.427955222347745e-07,
|
| 241 |
+
"loss": 2.2897,
|
| 242 |
+
"step": 39000
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 11.58,
|
| 246 |
+
"learning_rate": 5.576666324329876e-07,
|
| 247 |
+
"loss": 2.2826,
|
| 248 |
+
"step": 40000
|
| 249 |
+
}
|
| 250 |
+
],
|
| 251 |
+
"max_steps": 41448,
|
| 252 |
+
"num_train_epochs": 12,
|
| 253 |
+
"total_flos": 4.877556310867968e+16,
|
| 254 |
+
"trial_name": null,
|
| 255 |
+
"trial_params": null
|
| 256 |
+
}
|