Upload trainer_state.json
Browse files- trainer_state.json +340 -0
trainer_state.json
ADDED
|
@@ -0,0 +1,340 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 3.0,
|
| 5 |
+
"global_step": 24450,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.06,
|
| 12 |
+
"learning_rate": 4.897750511247444e-05,
|
| 13 |
+
"loss": 1.7982,
|
| 14 |
+
"step": 500
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.12,
|
| 18 |
+
"learning_rate": 4.7955010224948876e-05,
|
| 19 |
+
"loss": 1.6538,
|
| 20 |
+
"step": 1000
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.18,
|
| 24 |
+
"learning_rate": 4.693251533742332e-05,
|
| 25 |
+
"loss": 1.5864,
|
| 26 |
+
"step": 1500
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.25,
|
| 30 |
+
"learning_rate": 4.591002044989775e-05,
|
| 31 |
+
"loss": 1.5518,
|
| 32 |
+
"step": 2000
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.31,
|
| 36 |
+
"learning_rate": 4.488752556237219e-05,
|
| 37 |
+
"loss": 1.5189,
|
| 38 |
+
"step": 2500
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.37,
|
| 42 |
+
"learning_rate": 4.386503067484663e-05,
|
| 43 |
+
"loss": 1.5078,
|
| 44 |
+
"step": 3000
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.43,
|
| 48 |
+
"learning_rate": 4.2842535787321066e-05,
|
| 49 |
+
"loss": 1.4769,
|
| 50 |
+
"step": 3500
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.49,
|
| 54 |
+
"learning_rate": 4.18200408997955e-05,
|
| 55 |
+
"loss": 1.4499,
|
| 56 |
+
"step": 4000
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.55,
|
| 60 |
+
"learning_rate": 4.079754601226994e-05,
|
| 61 |
+
"loss": 1.4412,
|
| 62 |
+
"step": 4500
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.61,
|
| 66 |
+
"learning_rate": 3.9775051124744376e-05,
|
| 67 |
+
"loss": 1.4154,
|
| 68 |
+
"step": 5000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.67,
|
| 72 |
+
"learning_rate": 3.875255623721882e-05,
|
| 73 |
+
"loss": 1.4142,
|
| 74 |
+
"step": 5500
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.74,
|
| 78 |
+
"learning_rate": 3.773006134969325e-05,
|
| 79 |
+
"loss": 1.4084,
|
| 80 |
+
"step": 6000
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.8,
|
| 84 |
+
"learning_rate": 3.670756646216769e-05,
|
| 85 |
+
"loss": 1.3804,
|
| 86 |
+
"step": 6500
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.86,
|
| 90 |
+
"learning_rate": 3.568507157464213e-05,
|
| 91 |
+
"loss": 1.3813,
|
| 92 |
+
"step": 7000
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.92,
|
| 96 |
+
"learning_rate": 3.4662576687116566e-05,
|
| 97 |
+
"loss": 1.3566,
|
| 98 |
+
"step": 7500
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.98,
|
| 102 |
+
"learning_rate": 3.3640081799591e-05,
|
| 103 |
+
"loss": 1.3624,
|
| 104 |
+
"step": 8000
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 1.0,
|
| 108 |
+
"eval_accuracy": 0.7132048832822165,
|
| 109 |
+
"eval_loss": 1.3430438041687012,
|
| 110 |
+
"eval_runtime": 167.0718,
|
| 111 |
+
"eval_samples_per_second": 91.2,
|
| 112 |
+
"eval_steps_per_second": 11.402,
|
| 113 |
+
"step": 8150
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 1.04,
|
| 117 |
+
"learning_rate": 3.261758691206544e-05,
|
| 118 |
+
"loss": 1.3519,
|
| 119 |
+
"step": 8500
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 1.1,
|
| 123 |
+
"learning_rate": 3.159509202453988e-05,
|
| 124 |
+
"loss": 1.3203,
|
| 125 |
+
"step": 9000
|
| 126 |
+
},
|
| 127 |
+
{
|
| 128 |
+
"epoch": 1.17,
|
| 129 |
+
"learning_rate": 3.057259713701431e-05,
|
| 130 |
+
"loss": 1.3242,
|
| 131 |
+
"step": 9500
|
| 132 |
+
},
|
| 133 |
+
{
|
| 134 |
+
"epoch": 1.23,
|
| 135 |
+
"learning_rate": 2.9550102249488753e-05,
|
| 136 |
+
"loss": 1.3069,
|
| 137 |
+
"step": 10000
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 1.29,
|
| 141 |
+
"learning_rate": 2.8527607361963193e-05,
|
| 142 |
+
"loss": 1.2915,
|
| 143 |
+
"step": 10500
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 1.35,
|
| 147 |
+
"learning_rate": 2.7505112474437626e-05,
|
| 148 |
+
"loss": 1.3011,
|
| 149 |
+
"step": 11000
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 1.41,
|
| 153 |
+
"learning_rate": 2.6482617586912066e-05,
|
| 154 |
+
"loss": 1.299,
|
| 155 |
+
"step": 11500
|
| 156 |
+
},
|
| 157 |
+
{
|
| 158 |
+
"epoch": 1.47,
|
| 159 |
+
"learning_rate": 2.5460122699386503e-05,
|
| 160 |
+
"loss": 1.2907,
|
| 161 |
+
"step": 12000
|
| 162 |
+
},
|
| 163 |
+
{
|
| 164 |
+
"epoch": 1.53,
|
| 165 |
+
"learning_rate": 2.4437627811860943e-05,
|
| 166 |
+
"loss": 1.2647,
|
| 167 |
+
"step": 12500
|
| 168 |
+
},
|
| 169 |
+
{
|
| 170 |
+
"epoch": 1.6,
|
| 171 |
+
"learning_rate": 2.341513292433538e-05,
|
| 172 |
+
"loss": 1.2744,
|
| 173 |
+
"step": 13000
|
| 174 |
+
},
|
| 175 |
+
{
|
| 176 |
+
"epoch": 1.66,
|
| 177 |
+
"learning_rate": 2.239263803680982e-05,
|
| 178 |
+
"loss": 1.2614,
|
| 179 |
+
"step": 13500
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 1.72,
|
| 183 |
+
"learning_rate": 2.1370143149284256e-05,
|
| 184 |
+
"loss": 1.2614,
|
| 185 |
+
"step": 14000
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 1.78,
|
| 189 |
+
"learning_rate": 2.0347648261758693e-05,
|
| 190 |
+
"loss": 1.2551,
|
| 191 |
+
"step": 14500
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 1.84,
|
| 195 |
+
"learning_rate": 1.932515337423313e-05,
|
| 196 |
+
"loss": 1.2401,
|
| 197 |
+
"step": 15000
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 1.9,
|
| 201 |
+
"learning_rate": 1.8302658486707566e-05,
|
| 202 |
+
"loss": 1.2415,
|
| 203 |
+
"step": 15500
|
| 204 |
+
},
|
| 205 |
+
{
|
| 206 |
+
"epoch": 1.96,
|
| 207 |
+
"learning_rate": 1.7280163599182006e-05,
|
| 208 |
+
"loss": 1.2328,
|
| 209 |
+
"step": 16000
|
| 210 |
+
},
|
| 211 |
+
{
|
| 212 |
+
"epoch": 2.0,
|
| 213 |
+
"eval_accuracy": 0.7264976637836763,
|
| 214 |
+
"eval_loss": 1.2673941850662231,
|
| 215 |
+
"eval_runtime": 167.9128,
|
| 216 |
+
"eval_samples_per_second": 90.744,
|
| 217 |
+
"eval_steps_per_second": 11.345,
|
| 218 |
+
"step": 16300
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 2.02,
|
| 222 |
+
"learning_rate": 1.6257668711656443e-05,
|
| 223 |
+
"loss": 1.2239,
|
| 224 |
+
"step": 16500
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 2.09,
|
| 228 |
+
"learning_rate": 1.523517382413088e-05,
|
| 229 |
+
"loss": 1.2146,
|
| 230 |
+
"step": 17000
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 2.15,
|
| 234 |
+
"learning_rate": 1.4212678936605318e-05,
|
| 235 |
+
"loss": 1.2142,
|
| 236 |
+
"step": 17500
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 2.21,
|
| 240 |
+
"learning_rate": 1.3190184049079754e-05,
|
| 241 |
+
"loss": 1.2113,
|
| 242 |
+
"step": 18000
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 2.27,
|
| 246 |
+
"learning_rate": 1.2167689161554193e-05,
|
| 247 |
+
"loss": 1.2011,
|
| 248 |
+
"step": 18500
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 2.33,
|
| 252 |
+
"learning_rate": 1.1145194274028631e-05,
|
| 253 |
+
"loss": 1.2008,
|
| 254 |
+
"step": 19000
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 2.39,
|
| 258 |
+
"learning_rate": 1.0122699386503068e-05,
|
| 259 |
+
"loss": 1.1976,
|
| 260 |
+
"step": 19500
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 2.45,
|
| 264 |
+
"learning_rate": 9.100204498977506e-06,
|
| 265 |
+
"loss": 1.1867,
|
| 266 |
+
"step": 20000
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 2.52,
|
| 270 |
+
"learning_rate": 8.077709611451943e-06,
|
| 271 |
+
"loss": 1.1797,
|
| 272 |
+
"step": 20500
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 2.58,
|
| 276 |
+
"learning_rate": 7.05521472392638e-06,
|
| 277 |
+
"loss": 1.1877,
|
| 278 |
+
"step": 21000
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 2.64,
|
| 282 |
+
"learning_rate": 6.032719836400819e-06,
|
| 283 |
+
"loss": 1.1828,
|
| 284 |
+
"step": 21500
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 2.7,
|
| 288 |
+
"learning_rate": 5.0102249488752554e-06,
|
| 289 |
+
"loss": 1.175,
|
| 290 |
+
"step": 22000
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 2.76,
|
| 294 |
+
"learning_rate": 3.987730061349693e-06,
|
| 295 |
+
"loss": 1.1786,
|
| 296 |
+
"step": 22500
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 2.82,
|
| 300 |
+
"learning_rate": 2.965235173824131e-06,
|
| 301 |
+
"loss": 1.1747,
|
| 302 |
+
"step": 23000
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 2.88,
|
| 306 |
+
"learning_rate": 1.942740286298569e-06,
|
| 307 |
+
"loss": 1.1736,
|
| 308 |
+
"step": 23500
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 2.94,
|
| 312 |
+
"learning_rate": 9.202453987730062e-07,
|
| 313 |
+
"loss": 1.1753,
|
| 314 |
+
"step": 24000
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 3.0,
|
| 318 |
+
"eval_accuracy": 0.7363739433117852,
|
| 319 |
+
"eval_loss": 1.2128783464431763,
|
| 320 |
+
"eval_runtime": 167.207,
|
| 321 |
+
"eval_samples_per_second": 91.127,
|
| 322 |
+
"eval_steps_per_second": 11.393,
|
| 323 |
+
"step": 24450
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"epoch": 3.0,
|
| 327 |
+
"step": 24450,
|
| 328 |
+
"total_flos": 1.0143417323453616e+17,
|
| 329 |
+
"train_loss": 1.316010577430023,
|
| 330 |
+
"train_runtime": 32228.2163,
|
| 331 |
+
"train_samples_per_second": 24.276,
|
| 332 |
+
"train_steps_per_second": 0.759
|
| 333 |
+
}
|
| 334 |
+
],
|
| 335 |
+
"max_steps": 24450,
|
| 336 |
+
"num_train_epochs": 3,
|
| 337 |
+
"total_flos": 1.0143417323453616e+17,
|
| 338 |
+
"trial_name": null,
|
| 339 |
+
"trial_params": null
|
| 340 |
+
}
|