End of training
Browse files- all_results.json +11 -11
- eval_results.json +6 -6
- train_results.json +6 -6
- trainer_state.json +497 -209
all_results.json
CHANGED
|
@@ -1,13 +1,13 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"eval_accuracy": 0.
|
| 4 |
-
"eval_loss": 0.
|
| 5 |
-
"eval_runtime":
|
| 6 |
-
"eval_samples_per_second":
|
| 7 |
-
"eval_steps_per_second":
|
| 8 |
-
"total_flos":
|
| 9 |
-
"train_loss": 0.
|
| 10 |
-
"train_runtime":
|
| 11 |
-
"train_samples_per_second":
|
| 12 |
-
"train_steps_per_second":
|
| 13 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 2.99,
|
| 3 |
+
"eval_accuracy": 0.973,
|
| 4 |
+
"eval_loss": 0.079922154545784,
|
| 5 |
+
"eval_runtime": 20.1587,
|
| 6 |
+
"eval_samples_per_second": 248.032,
|
| 7 |
+
"eval_steps_per_second": 7.788,
|
| 8 |
+
"total_flos": 3.3497451642252165e+18,
|
| 9 |
+
"train_loss": 0.5413832122217669,
|
| 10 |
+
"train_runtime": 1364.9822,
|
| 11 |
+
"train_samples_per_second": 98.902,
|
| 12 |
+
"train_steps_per_second": 0.771
|
| 13 |
}
|
eval_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"eval_accuracy": 0.
|
| 4 |
-
"eval_loss": 0.
|
| 5 |
-
"eval_runtime":
|
| 6 |
-
"eval_samples_per_second":
|
| 7 |
-
"eval_steps_per_second":
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 2.99,
|
| 3 |
+
"eval_accuracy": 0.973,
|
| 4 |
+
"eval_loss": 0.079922154545784,
|
| 5 |
+
"eval_runtime": 20.1587,
|
| 6 |
+
"eval_samples_per_second": 248.032,
|
| 7 |
+
"eval_steps_per_second": 7.788
|
| 8 |
}
|
train_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"total_flos":
|
| 4 |
-
"train_loss": 0.
|
| 5 |
-
"train_runtime":
|
| 6 |
-
"train_samples_per_second":
|
| 7 |
-
"train_steps_per_second":
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 2.99,
|
| 3 |
+
"total_flos": 3.3497451642252165e+18,
|
| 4 |
+
"train_loss": 0.5413832122217669,
|
| 5 |
+
"train_runtime": 1364.9822,
|
| 6 |
+
"train_samples_per_second": 98.902,
|
| 7 |
+
"train_steps_per_second": 0.771
|
| 8 |
}
|
trainer_state.json
CHANGED
|
@@ -1,397 +1,685 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 0.
|
| 3 |
-
"best_model_checkpoint": "swin-tiny-patch4-window7-224-finetuned-eurosat/checkpoint-
|
| 4 |
-
"epoch":
|
| 5 |
"eval_steps": 500,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
-
"epoch": 0.
|
| 13 |
-
"learning_rate":
|
| 14 |
-
"loss": 2.
|
| 15 |
"step": 10
|
| 16 |
},
|
| 17 |
{
|
| 18 |
-
"epoch": 0.
|
| 19 |
-
"learning_rate":
|
| 20 |
-
"loss": 2.
|
| 21 |
"step": 20
|
| 22 |
},
|
| 23 |
{
|
| 24 |
-
"epoch": 0.
|
| 25 |
-
"learning_rate":
|
| 26 |
-
"loss":
|
| 27 |
"step": 30
|
| 28 |
},
|
| 29 |
{
|
| 30 |
-
"epoch": 0.
|
| 31 |
-
"learning_rate":
|
| 32 |
-
"loss":
|
| 33 |
"step": 40
|
| 34 |
},
|
| 35 |
{
|
| 36 |
-
"epoch": 0.
|
| 37 |
-
"learning_rate":
|
| 38 |
-
"loss":
|
| 39 |
"step": 50
|
| 40 |
},
|
| 41 |
{
|
| 42 |
-
"epoch": 0.
|
| 43 |
-
"learning_rate":
|
| 44 |
-
"loss":
|
| 45 |
"step": 60
|
| 46 |
},
|
| 47 |
{
|
| 48 |
-
"epoch": 0.
|
| 49 |
-
"learning_rate":
|
| 50 |
-
"loss":
|
| 51 |
"step": 70
|
| 52 |
},
|
| 53 |
{
|
| 54 |
-
"epoch": 0.
|
| 55 |
-
"learning_rate":
|
| 56 |
-
"loss": 0.
|
| 57 |
"step": 80
|
| 58 |
},
|
| 59 |
{
|
| 60 |
-
"epoch": 0.
|
| 61 |
-
"learning_rate": 4.
|
| 62 |
-
"loss": 0.
|
| 63 |
"step": 90
|
| 64 |
},
|
| 65 |
{
|
| 66 |
-
"epoch": 0.
|
| 67 |
-
"learning_rate": 4.
|
| 68 |
-
"loss": 0.
|
| 69 |
"step": 100
|
| 70 |
},
|
| 71 |
{
|
| 72 |
-
"epoch": 0.
|
| 73 |
-
"learning_rate": 4.
|
| 74 |
-
"loss": 0.
|
| 75 |
"step": 110
|
| 76 |
},
|
| 77 |
{
|
| 78 |
-
"epoch": 0.
|
| 79 |
-
"learning_rate": 4.
|
| 80 |
-
"loss": 0.
|
| 81 |
"step": 120
|
| 82 |
},
|
| 83 |
{
|
| 84 |
-
"epoch": 0.
|
| 85 |
-
"learning_rate": 4.
|
| 86 |
-
"loss": 0.
|
| 87 |
"step": 130
|
| 88 |
},
|
| 89 |
{
|
| 90 |
-
"epoch": 0.
|
| 91 |
-
"learning_rate": 4.
|
| 92 |
-
"loss": 0.
|
| 93 |
"step": 140
|
| 94 |
},
|
| 95 |
{
|
| 96 |
-
"epoch": 0.
|
| 97 |
-
"learning_rate": 4.
|
| 98 |
-
"loss": 0.
|
| 99 |
"step": 150
|
| 100 |
},
|
| 101 |
{
|
| 102 |
-
"epoch": 0.
|
| 103 |
-
"learning_rate":
|
| 104 |
-
"loss": 0.
|
| 105 |
"step": 160
|
| 106 |
},
|
| 107 |
{
|
| 108 |
-
"epoch": 0.
|
| 109 |
-
"learning_rate":
|
| 110 |
-
"loss": 0.
|
| 111 |
"step": 170
|
| 112 |
},
|
| 113 |
{
|
| 114 |
-
"epoch": 0.
|
| 115 |
-
"learning_rate":
|
| 116 |
-
"loss": 0.
|
| 117 |
"step": 180
|
| 118 |
},
|
| 119 |
{
|
| 120 |
-
"epoch":
|
| 121 |
-
"learning_rate":
|
| 122 |
-
"loss": 0.
|
| 123 |
"step": 190
|
| 124 |
},
|
| 125 |
{
|
| 126 |
-
"epoch":
|
| 127 |
-
"
|
| 128 |
-
"
|
| 129 |
-
"eval_runtime": 5.1297,
|
| 130 |
-
"eval_samples_per_second": 526.346,
|
| 131 |
-
"eval_steps_per_second": 16.57,
|
| 132 |
-
"step": 190
|
| 133 |
-
},
|
| 134 |
-
{
|
| 135 |
-
"epoch": 1.05,
|
| 136 |
-
"learning_rate": 3.606237816764133e-05,
|
| 137 |
-
"loss": 0.2245,
|
| 138 |
"step": 200
|
| 139 |
},
|
| 140 |
{
|
| 141 |
-
"epoch":
|
| 142 |
-
"learning_rate":
|
| 143 |
-
"loss": 0.
|
| 144 |
"step": 210
|
| 145 |
},
|
| 146 |
{
|
| 147 |
-
"epoch":
|
| 148 |
-
"learning_rate":
|
| 149 |
-
"loss": 0.
|
| 150 |
"step": 220
|
| 151 |
},
|
| 152 |
{
|
| 153 |
-
"epoch":
|
| 154 |
-
"learning_rate":
|
| 155 |
-
"loss": 0.
|
| 156 |
"step": 230
|
| 157 |
},
|
| 158 |
{
|
| 159 |
-
"epoch":
|
| 160 |
-
"learning_rate":
|
| 161 |
-
"loss": 0.
|
| 162 |
"step": 240
|
| 163 |
},
|
| 164 |
{
|
| 165 |
-
"epoch":
|
| 166 |
-
"learning_rate":
|
| 167 |
-
"loss": 0.
|
| 168 |
"step": 250
|
| 169 |
},
|
| 170 |
{
|
| 171 |
-
"epoch":
|
| 172 |
-
"learning_rate":
|
| 173 |
-
"loss": 0.
|
| 174 |
"step": 260
|
| 175 |
},
|
| 176 |
{
|
| 177 |
-
"epoch":
|
| 178 |
-
"learning_rate":
|
| 179 |
-
"loss": 0.
|
| 180 |
"step": 270
|
| 181 |
},
|
| 182 |
{
|
| 183 |
-
"epoch":
|
| 184 |
-
"learning_rate":
|
| 185 |
-
"loss": 0.
|
| 186 |
"step": 280
|
| 187 |
},
|
| 188 |
{
|
| 189 |
-
"epoch":
|
| 190 |
-
"learning_rate":
|
| 191 |
-
"loss": 0.
|
| 192 |
"step": 290
|
| 193 |
},
|
| 194 |
{
|
| 195 |
-
"epoch":
|
| 196 |
-
"learning_rate":
|
| 197 |
-
"loss": 0.
|
| 198 |
"step": 300
|
| 199 |
},
|
| 200 |
{
|
| 201 |
-
"epoch":
|
| 202 |
-
"learning_rate":
|
| 203 |
-
"loss": 0.
|
| 204 |
"step": 310
|
| 205 |
},
|
| 206 |
{
|
| 207 |
-
"epoch":
|
| 208 |
-
"learning_rate":
|
| 209 |
-
"loss": 0.
|
| 210 |
"step": 320
|
| 211 |
},
|
| 212 |
{
|
| 213 |
-
"epoch":
|
| 214 |
-
"learning_rate":
|
| 215 |
-
"loss": 0.
|
| 216 |
"step": 330
|
| 217 |
},
|
| 218 |
{
|
| 219 |
-
"epoch":
|
| 220 |
-
"learning_rate":
|
| 221 |
-
"loss": 0.
|
| 222 |
"step": 340
|
| 223 |
},
|
| 224 |
{
|
| 225 |
-
"epoch": 1.
|
| 226 |
-
"learning_rate":
|
| 227 |
-
"loss": 0.
|
| 228 |
"step": 350
|
| 229 |
},
|
| 230 |
{
|
| 231 |
-
"epoch": 1.
|
| 232 |
-
"
|
| 233 |
-
"
|
| 234 |
-
"
|
|
|
|
|
|
|
|
|
|
| 235 |
},
|
| 236 |
{
|
| 237 |
-
"epoch": 1.
|
| 238 |
-
"learning_rate":
|
| 239 |
-
"loss": 0.
|
| 240 |
-
"step":
|
| 241 |
},
|
| 242 |
{
|
| 243 |
-
"epoch":
|
| 244 |
-
"learning_rate":
|
| 245 |
-
"loss": 0.
|
| 246 |
-
"step":
|
| 247 |
},
|
| 248 |
{
|
| 249 |
-
"epoch":
|
| 250 |
-
"
|
| 251 |
-
"
|
| 252 |
-
"eval_runtime": 4.4943,
|
| 253 |
-
"eval_samples_per_second": 600.762,
|
| 254 |
-
"eval_steps_per_second": 18.913,
|
| 255 |
"step": 380
|
| 256 |
},
|
| 257 |
{
|
| 258 |
-
"epoch":
|
| 259 |
-
"learning_rate":
|
| 260 |
-
"loss": 0.
|
| 261 |
"step": 390
|
| 262 |
},
|
| 263 |
{
|
| 264 |
-
"epoch":
|
| 265 |
-
"learning_rate":
|
| 266 |
-
"loss": 0.
|
| 267 |
"step": 400
|
| 268 |
},
|
| 269 |
{
|
| 270 |
-
"epoch":
|
| 271 |
-
"learning_rate":
|
| 272 |
-
"loss": 0.
|
| 273 |
"step": 410
|
| 274 |
},
|
| 275 |
{
|
| 276 |
-
"epoch":
|
| 277 |
-
"learning_rate":
|
| 278 |
-
"loss": 0.
|
| 279 |
"step": 420
|
| 280 |
},
|
| 281 |
{
|
| 282 |
-
"epoch":
|
| 283 |
-
"learning_rate":
|
| 284 |
-
"loss": 0.
|
| 285 |
"step": 430
|
| 286 |
},
|
| 287 |
{
|
| 288 |
-
"epoch":
|
| 289 |
-
"learning_rate":
|
| 290 |
-
"loss": 0.
|
| 291 |
"step": 440
|
| 292 |
},
|
| 293 |
{
|
| 294 |
-
"epoch":
|
| 295 |
-
"learning_rate":
|
| 296 |
-
"loss": 0.
|
| 297 |
"step": 450
|
| 298 |
},
|
| 299 |
{
|
| 300 |
-
"epoch":
|
| 301 |
-
"learning_rate":
|
| 302 |
-
"loss": 0.
|
| 303 |
"step": 460
|
| 304 |
},
|
| 305 |
{
|
| 306 |
-
"epoch":
|
| 307 |
-
"learning_rate":
|
| 308 |
-
"loss": 0.
|
| 309 |
"step": 470
|
| 310 |
},
|
| 311 |
{
|
| 312 |
-
"epoch":
|
| 313 |
-
"learning_rate":
|
| 314 |
-
"loss": 0.
|
| 315 |
"step": 480
|
| 316 |
},
|
| 317 |
{
|
| 318 |
-
"epoch":
|
| 319 |
-
"learning_rate":
|
| 320 |
-
"loss": 0.
|
| 321 |
"step": 490
|
| 322 |
},
|
| 323 |
{
|
| 324 |
-
"epoch":
|
| 325 |
-
"learning_rate":
|
| 326 |
-
"loss": 0.
|
| 327 |
"step": 500
|
| 328 |
},
|
| 329 |
{
|
| 330 |
-
"epoch":
|
| 331 |
-
"learning_rate":
|
| 332 |
-
"loss": 0.
|
| 333 |
"step": 510
|
| 334 |
},
|
| 335 |
{
|
| 336 |
-
"epoch":
|
| 337 |
-
"learning_rate":
|
| 338 |
-
"loss": 0.
|
| 339 |
"step": 520
|
| 340 |
},
|
| 341 |
{
|
| 342 |
-
"epoch":
|
| 343 |
-
"learning_rate":
|
| 344 |
-
"loss": 0.
|
| 345 |
"step": 530
|
| 346 |
},
|
| 347 |
{
|
| 348 |
-
"epoch":
|
| 349 |
-
"learning_rate": 2.
|
| 350 |
-
"loss": 0.
|
| 351 |
"step": 540
|
| 352 |
},
|
| 353 |
{
|
| 354 |
-
"epoch":
|
| 355 |
-
"learning_rate":
|
| 356 |
-
"loss": 0.
|
| 357 |
"step": 550
|
| 358 |
},
|
| 359 |
{
|
| 360 |
-
"epoch":
|
| 361 |
-
"learning_rate":
|
| 362 |
-
"loss": 0.
|
| 363 |
"step": 560
|
| 364 |
},
|
| 365 |
{
|
| 366 |
-
"epoch":
|
| 367 |
-
"learning_rate":
|
| 368 |
-
"loss": 0.
|
| 369 |
"step": 570
|
| 370 |
},
|
| 371 |
{
|
| 372 |
-
"epoch":
|
| 373 |
-
"
|
| 374 |
-
"
|
| 375 |
-
"
|
| 376 |
-
|
| 377 |
-
|
| 378 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 379 |
},
|
| 380 |
{
|
| 381 |
-
"epoch":
|
| 382 |
-
"step":
|
| 383 |
-
"total_flos":
|
| 384 |
-
"train_loss": 0.
|
| 385 |
-
"train_runtime":
|
| 386 |
-
"train_samples_per_second":
|
| 387 |
-
"train_steps_per_second":
|
| 388 |
}
|
| 389 |
],
|
| 390 |
"logging_steps": 10,
|
| 391 |
-
"max_steps":
|
| 392 |
"num_train_epochs": 3,
|
| 393 |
"save_steps": 500,
|
| 394 |
-
"total_flos":
|
| 395 |
"trial_name": null,
|
| 396 |
"trial_params": null
|
| 397 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 0.973,
|
| 3 |
+
"best_model_checkpoint": "swin-tiny-patch4-window7-224-finetuned-eurosat/checkpoint-1053",
|
| 4 |
+
"epoch": 2.9936034115138592,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 1053,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
+
"epoch": 0.03,
|
| 13 |
+
"learning_rate": 4.716981132075472e-06,
|
| 14 |
+
"loss": 2.3037,
|
| 15 |
"step": 10
|
| 16 |
},
|
| 17 |
{
|
| 18 |
+
"epoch": 0.06,
|
| 19 |
+
"learning_rate": 9.433962264150944e-06,
|
| 20 |
+
"loss": 2.2484,
|
| 21 |
"step": 20
|
| 22 |
},
|
| 23 |
{
|
| 24 |
+
"epoch": 0.09,
|
| 25 |
+
"learning_rate": 1.4150943396226415e-05,
|
| 26 |
+
"loss": 2.109,
|
| 27 |
"step": 30
|
| 28 |
},
|
| 29 |
{
|
| 30 |
+
"epoch": 0.11,
|
| 31 |
+
"learning_rate": 1.8867924528301888e-05,
|
| 32 |
+
"loss": 1.8715,
|
| 33 |
"step": 40
|
| 34 |
},
|
| 35 |
{
|
| 36 |
+
"epoch": 0.14,
|
| 37 |
+
"learning_rate": 2.358490566037736e-05,
|
| 38 |
+
"loss": 1.5984,
|
| 39 |
"step": 50
|
| 40 |
},
|
| 41 |
{
|
| 42 |
+
"epoch": 0.17,
|
| 43 |
+
"learning_rate": 2.830188679245283e-05,
|
| 44 |
+
"loss": 1.323,
|
| 45 |
"step": 60
|
| 46 |
},
|
| 47 |
{
|
| 48 |
+
"epoch": 0.2,
|
| 49 |
+
"learning_rate": 3.30188679245283e-05,
|
| 50 |
+
"loss": 1.1205,
|
| 51 |
"step": 70
|
| 52 |
},
|
| 53 |
{
|
| 54 |
+
"epoch": 0.23,
|
| 55 |
+
"learning_rate": 3.7735849056603776e-05,
|
| 56 |
+
"loss": 0.9739,
|
| 57 |
"step": 80
|
| 58 |
},
|
| 59 |
{
|
| 60 |
+
"epoch": 0.26,
|
| 61 |
+
"learning_rate": 4.245283018867925e-05,
|
| 62 |
+
"loss": 0.9502,
|
| 63 |
"step": 90
|
| 64 |
},
|
| 65 |
{
|
| 66 |
+
"epoch": 0.28,
|
| 67 |
+
"learning_rate": 4.716981132075472e-05,
|
| 68 |
+
"loss": 0.8486,
|
| 69 |
"step": 100
|
| 70 |
},
|
| 71 |
{
|
| 72 |
+
"epoch": 0.31,
|
| 73 |
+
"learning_rate": 4.978880675818374e-05,
|
| 74 |
+
"loss": 0.7495,
|
| 75 |
"step": 110
|
| 76 |
},
|
| 77 |
{
|
| 78 |
+
"epoch": 0.34,
|
| 79 |
+
"learning_rate": 4.9260823653643085e-05,
|
| 80 |
+
"loss": 0.7036,
|
| 81 |
"step": 120
|
| 82 |
},
|
| 83 |
{
|
| 84 |
+
"epoch": 0.37,
|
| 85 |
+
"learning_rate": 4.8732840549102435e-05,
|
| 86 |
+
"loss": 0.6417,
|
| 87 |
"step": 130
|
| 88 |
},
|
| 89 |
{
|
| 90 |
+
"epoch": 0.4,
|
| 91 |
+
"learning_rate": 4.820485744456177e-05,
|
| 92 |
+
"loss": 0.6558,
|
| 93 |
"step": 140
|
| 94 |
},
|
| 95 |
{
|
| 96 |
+
"epoch": 0.43,
|
| 97 |
+
"learning_rate": 4.767687434002112e-05,
|
| 98 |
+
"loss": 0.5976,
|
| 99 |
"step": 150
|
| 100 |
},
|
| 101 |
{
|
| 102 |
+
"epoch": 0.45,
|
| 103 |
+
"learning_rate": 4.7148891235480466e-05,
|
| 104 |
+
"loss": 0.6296,
|
| 105 |
"step": 160
|
| 106 |
},
|
| 107 |
{
|
| 108 |
+
"epoch": 0.48,
|
| 109 |
+
"learning_rate": 4.662090813093981e-05,
|
| 110 |
+
"loss": 0.6067,
|
| 111 |
"step": 170
|
| 112 |
},
|
| 113 |
{
|
| 114 |
+
"epoch": 0.51,
|
| 115 |
+
"learning_rate": 4.609292502639916e-05,
|
| 116 |
+
"loss": 0.5513,
|
| 117 |
"step": 180
|
| 118 |
},
|
| 119 |
{
|
| 120 |
+
"epoch": 0.54,
|
| 121 |
+
"learning_rate": 4.55649419218585e-05,
|
| 122 |
+
"loss": 0.5577,
|
| 123 |
"step": 190
|
| 124 |
},
|
| 125 |
{
|
| 126 |
+
"epoch": 0.57,
|
| 127 |
+
"learning_rate": 4.503695881731785e-05,
|
| 128 |
+
"loss": 0.5611,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 129 |
"step": 200
|
| 130 |
},
|
| 131 |
{
|
| 132 |
+
"epoch": 0.6,
|
| 133 |
+
"learning_rate": 4.45089757127772e-05,
|
| 134 |
+
"loss": 0.5693,
|
| 135 |
"step": 210
|
| 136 |
},
|
| 137 |
{
|
| 138 |
+
"epoch": 0.63,
|
| 139 |
+
"learning_rate": 4.398099260823654e-05,
|
| 140 |
+
"loss": 0.5354,
|
| 141 |
"step": 220
|
| 142 |
},
|
| 143 |
{
|
| 144 |
+
"epoch": 0.65,
|
| 145 |
+
"learning_rate": 4.3453009503695884e-05,
|
| 146 |
+
"loss": 0.5329,
|
| 147 |
"step": 230
|
| 148 |
},
|
| 149 |
{
|
| 150 |
+
"epoch": 0.68,
|
| 151 |
+
"learning_rate": 4.292502639915523e-05,
|
| 152 |
+
"loss": 0.544,
|
| 153 |
"step": 240
|
| 154 |
},
|
| 155 |
{
|
| 156 |
+
"epoch": 0.71,
|
| 157 |
+
"learning_rate": 4.239704329461457e-05,
|
| 158 |
+
"loss": 0.5865,
|
| 159 |
"step": 250
|
| 160 |
},
|
| 161 |
{
|
| 162 |
+
"epoch": 0.74,
|
| 163 |
+
"learning_rate": 4.186906019007392e-05,
|
| 164 |
+
"loss": 0.4998,
|
| 165 |
"step": 260
|
| 166 |
},
|
| 167 |
{
|
| 168 |
+
"epoch": 0.77,
|
| 169 |
+
"learning_rate": 4.1341077085533265e-05,
|
| 170 |
+
"loss": 0.5228,
|
| 171 |
"step": 270
|
| 172 |
},
|
| 173 |
{
|
| 174 |
+
"epoch": 0.8,
|
| 175 |
+
"learning_rate": 4.081309398099261e-05,
|
| 176 |
+
"loss": 0.4614,
|
| 177 |
"step": 280
|
| 178 |
},
|
| 179 |
{
|
| 180 |
+
"epoch": 0.82,
|
| 181 |
+
"learning_rate": 4.028511087645195e-05,
|
| 182 |
+
"loss": 0.5211,
|
| 183 |
"step": 290
|
| 184 |
},
|
| 185 |
{
|
| 186 |
+
"epoch": 0.85,
|
| 187 |
+
"learning_rate": 3.97571277719113e-05,
|
| 188 |
+
"loss": 0.5343,
|
| 189 |
"step": 300
|
| 190 |
},
|
| 191 |
{
|
| 192 |
+
"epoch": 0.88,
|
| 193 |
+
"learning_rate": 3.9229144667370646e-05,
|
| 194 |
+
"loss": 0.5135,
|
| 195 |
"step": 310
|
| 196 |
},
|
| 197 |
{
|
| 198 |
+
"epoch": 0.91,
|
| 199 |
+
"learning_rate": 3.870116156282999e-05,
|
| 200 |
+
"loss": 0.4749,
|
| 201 |
"step": 320
|
| 202 |
},
|
| 203 |
{
|
| 204 |
+
"epoch": 0.94,
|
| 205 |
+
"learning_rate": 3.817317845828934e-05,
|
| 206 |
+
"loss": 0.4425,
|
| 207 |
"step": 330
|
| 208 |
},
|
| 209 |
{
|
| 210 |
+
"epoch": 0.97,
|
| 211 |
+
"learning_rate": 3.764519535374868e-05,
|
| 212 |
+
"loss": 0.512,
|
| 213 |
"step": 340
|
| 214 |
},
|
| 215 |
{
|
| 216 |
+
"epoch": 1.0,
|
| 217 |
+
"learning_rate": 3.711721224920803e-05,
|
| 218 |
+
"loss": 0.5238,
|
| 219 |
"step": 350
|
| 220 |
},
|
| 221 |
{
|
| 222 |
+
"epoch": 1.0,
|
| 223 |
+
"eval_accuracy": 0.9512,
|
| 224 |
+
"eval_loss": 0.14167343080043793,
|
| 225 |
+
"eval_runtime": 20.0773,
|
| 226 |
+
"eval_samples_per_second": 249.038,
|
| 227 |
+
"eval_steps_per_second": 7.82,
|
| 228 |
+
"step": 351
|
| 229 |
},
|
| 230 |
{
|
| 231 |
+
"epoch": 1.02,
|
| 232 |
+
"learning_rate": 3.658922914466738e-05,
|
| 233 |
+
"loss": 0.4826,
|
| 234 |
+
"step": 360
|
| 235 |
},
|
| 236 |
{
|
| 237 |
+
"epoch": 1.05,
|
| 238 |
+
"learning_rate": 3.6061246040126714e-05,
|
| 239 |
+
"loss": 0.4862,
|
| 240 |
+
"step": 370
|
| 241 |
},
|
| 242 |
{
|
| 243 |
+
"epoch": 1.08,
|
| 244 |
+
"learning_rate": 3.5533262935586064e-05,
|
| 245 |
+
"loss": 0.4651,
|
|
|
|
|
|
|
|
|
|
| 246 |
"step": 380
|
| 247 |
},
|
| 248 |
{
|
| 249 |
+
"epoch": 1.11,
|
| 250 |
+
"learning_rate": 3.500527983104541e-05,
|
| 251 |
+
"loss": 0.4217,
|
| 252 |
"step": 390
|
| 253 |
},
|
| 254 |
{
|
| 255 |
+
"epoch": 1.14,
|
| 256 |
+
"learning_rate": 3.447729672650475e-05,
|
| 257 |
+
"loss": 0.4617,
|
| 258 |
"step": 400
|
| 259 |
},
|
| 260 |
{
|
| 261 |
+
"epoch": 1.17,
|
| 262 |
+
"learning_rate": 3.3949313621964095e-05,
|
| 263 |
+
"loss": 0.4409,
|
| 264 |
"step": 410
|
| 265 |
},
|
| 266 |
{
|
| 267 |
+
"epoch": 1.19,
|
| 268 |
+
"learning_rate": 3.3421330517423445e-05,
|
| 269 |
+
"loss": 0.4056,
|
| 270 |
"step": 420
|
| 271 |
},
|
| 272 |
{
|
| 273 |
+
"epoch": 1.22,
|
| 274 |
+
"learning_rate": 3.289334741288279e-05,
|
| 275 |
+
"loss": 0.4872,
|
| 276 |
"step": 430
|
| 277 |
},
|
| 278 |
{
|
| 279 |
+
"epoch": 1.25,
|
| 280 |
+
"learning_rate": 3.236536430834213e-05,
|
| 281 |
+
"loss": 0.4484,
|
| 282 |
"step": 440
|
| 283 |
},
|
| 284 |
{
|
| 285 |
+
"epoch": 1.28,
|
| 286 |
+
"learning_rate": 3.183738120380148e-05,
|
| 287 |
+
"loss": 0.4289,
|
| 288 |
"step": 450
|
| 289 |
},
|
| 290 |
{
|
| 291 |
+
"epoch": 1.31,
|
| 292 |
+
"learning_rate": 3.130939809926082e-05,
|
| 293 |
+
"loss": 0.4384,
|
| 294 |
"step": 460
|
| 295 |
},
|
| 296 |
{
|
| 297 |
+
"epoch": 1.34,
|
| 298 |
+
"learning_rate": 3.078141499472017e-05,
|
| 299 |
+
"loss": 0.4248,
|
| 300 |
"step": 470
|
| 301 |
},
|
| 302 |
{
|
| 303 |
+
"epoch": 1.36,
|
| 304 |
+
"learning_rate": 3.0253431890179517e-05,
|
| 305 |
+
"loss": 0.4518,
|
| 306 |
"step": 480
|
| 307 |
},
|
| 308 |
{
|
| 309 |
+
"epoch": 1.39,
|
| 310 |
+
"learning_rate": 2.972544878563886e-05,
|
| 311 |
+
"loss": 0.4412,
|
| 312 |
"step": 490
|
| 313 |
},
|
| 314 |
{
|
| 315 |
+
"epoch": 1.42,
|
| 316 |
+
"learning_rate": 2.9197465681098207e-05,
|
| 317 |
+
"loss": 0.4597,
|
| 318 |
"step": 500
|
| 319 |
},
|
| 320 |
{
|
| 321 |
+
"epoch": 1.45,
|
| 322 |
+
"learning_rate": 2.8669482576557548e-05,
|
| 323 |
+
"loss": 0.458,
|
| 324 |
"step": 510
|
| 325 |
},
|
| 326 |
{
|
| 327 |
+
"epoch": 1.48,
|
| 328 |
+
"learning_rate": 2.8141499472016898e-05,
|
| 329 |
+
"loss": 0.4359,
|
| 330 |
"step": 520
|
| 331 |
},
|
| 332 |
{
|
| 333 |
+
"epoch": 1.51,
|
| 334 |
+
"learning_rate": 2.7613516367476245e-05,
|
| 335 |
+
"loss": 0.4051,
|
| 336 |
"step": 530
|
| 337 |
},
|
| 338 |
{
|
| 339 |
+
"epoch": 1.54,
|
| 340 |
+
"learning_rate": 2.7085533262935585e-05,
|
| 341 |
+
"loss": 0.4527,
|
| 342 |
"step": 540
|
| 343 |
},
|
| 344 |
{
|
| 345 |
+
"epoch": 1.56,
|
| 346 |
+
"learning_rate": 2.6557550158394935e-05,
|
| 347 |
+
"loss": 0.4189,
|
| 348 |
"step": 550
|
| 349 |
},
|
| 350 |
{
|
| 351 |
+
"epoch": 1.59,
|
| 352 |
+
"learning_rate": 2.6029567053854276e-05,
|
| 353 |
+
"loss": 0.4014,
|
| 354 |
"step": 560
|
| 355 |
},
|
| 356 |
{
|
| 357 |
+
"epoch": 1.62,
|
| 358 |
+
"learning_rate": 2.5501583949313622e-05,
|
| 359 |
+
"loss": 0.4195,
|
| 360 |
"step": 570
|
| 361 |
},
|
| 362 |
{
|
| 363 |
+
"epoch": 1.65,
|
| 364 |
+
"learning_rate": 2.497360084477297e-05,
|
| 365 |
+
"loss": 0.4497,
|
| 366 |
+
"step": 580
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 1.68,
|
| 370 |
+
"learning_rate": 2.4445617740232313e-05,
|
| 371 |
+
"loss": 0.4156,
|
| 372 |
+
"step": 590
|
| 373 |
+
},
|
| 374 |
+
{
|
| 375 |
+
"epoch": 1.71,
|
| 376 |
+
"learning_rate": 2.391763463569166e-05,
|
| 377 |
+
"loss": 0.4268,
|
| 378 |
+
"step": 600
|
| 379 |
+
},
|
| 380 |
+
{
|
| 381 |
+
"epoch": 1.73,
|
| 382 |
+
"learning_rate": 2.3389651531151003e-05,
|
| 383 |
+
"loss": 0.436,
|
| 384 |
+
"step": 610
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 1.76,
|
| 388 |
+
"learning_rate": 2.286166842661035e-05,
|
| 389 |
+
"loss": 0.392,
|
| 390 |
+
"step": 620
|
| 391 |
+
},
|
| 392 |
+
{
|
| 393 |
+
"epoch": 1.79,
|
| 394 |
+
"learning_rate": 2.2333685322069694e-05,
|
| 395 |
+
"loss": 0.4059,
|
| 396 |
+
"step": 630
|
| 397 |
+
},
|
| 398 |
+
{
|
| 399 |
+
"epoch": 1.82,
|
| 400 |
+
"learning_rate": 2.180570221752904e-05,
|
| 401 |
+
"loss": 0.4014,
|
| 402 |
+
"step": 640
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 1.85,
|
| 406 |
+
"learning_rate": 2.1277719112988384e-05,
|
| 407 |
+
"loss": 0.4304,
|
| 408 |
+
"step": 650
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 1.88,
|
| 412 |
+
"learning_rate": 2.074973600844773e-05,
|
| 413 |
+
"loss": 0.3908,
|
| 414 |
+
"step": 660
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"epoch": 1.9,
|
| 418 |
+
"learning_rate": 2.0221752903907075e-05,
|
| 419 |
+
"loss": 0.3963,
|
| 420 |
+
"step": 670
|
| 421 |
+
},
|
| 422 |
+
{
|
| 423 |
+
"epoch": 1.93,
|
| 424 |
+
"learning_rate": 1.9693769799366422e-05,
|
| 425 |
+
"loss": 0.3438,
|
| 426 |
+
"step": 680
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 1.96,
|
| 430 |
+
"learning_rate": 1.9165786694825765e-05,
|
| 431 |
+
"loss": 0.3941,
|
| 432 |
+
"step": 690
|
| 433 |
+
},
|
| 434 |
+
{
|
| 435 |
+
"epoch": 1.99,
|
| 436 |
+
"learning_rate": 1.863780359028511e-05,
|
| 437 |
+
"loss": 0.3708,
|
| 438 |
+
"step": 700
|
| 439 |
+
},
|
| 440 |
+
{
|
| 441 |
+
"epoch": 2.0,
|
| 442 |
+
"eval_accuracy": 0.9688,
|
| 443 |
+
"eval_loss": 0.09748782962560654,
|
| 444 |
+
"eval_runtime": 20.0916,
|
| 445 |
+
"eval_samples_per_second": 248.86,
|
| 446 |
+
"eval_steps_per_second": 7.814,
|
| 447 |
+
"step": 703
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 2.02,
|
| 451 |
+
"learning_rate": 1.810982048574446e-05,
|
| 452 |
+
"loss": 0.3771,
|
| 453 |
+
"step": 710
|
| 454 |
+
},
|
| 455 |
+
{
|
| 456 |
+
"epoch": 2.05,
|
| 457 |
+
"learning_rate": 1.7581837381203803e-05,
|
| 458 |
+
"loss": 0.3559,
|
| 459 |
+
"step": 720
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 2.08,
|
| 463 |
+
"learning_rate": 1.7053854276663146e-05,
|
| 464 |
+
"loss": 0.379,
|
| 465 |
+
"step": 730
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 2.1,
|
| 469 |
+
"learning_rate": 1.6525871172122493e-05,
|
| 470 |
+
"loss": 0.3301,
|
| 471 |
+
"step": 740
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 2.13,
|
| 475 |
+
"learning_rate": 1.5997888067581837e-05,
|
| 476 |
+
"loss": 0.3691,
|
| 477 |
+
"step": 750
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"epoch": 2.16,
|
| 481 |
+
"learning_rate": 1.5469904963041184e-05,
|
| 482 |
+
"loss": 0.3257,
|
| 483 |
+
"step": 760
|
| 484 |
+
},
|
| 485 |
+
{
|
| 486 |
+
"epoch": 2.19,
|
| 487 |
+
"learning_rate": 1.4941921858500529e-05,
|
| 488 |
+
"loss": 0.3673,
|
| 489 |
+
"step": 770
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 2.22,
|
| 493 |
+
"learning_rate": 1.4413938753959874e-05,
|
| 494 |
+
"loss": 0.3701,
|
| 495 |
+
"step": 780
|
| 496 |
+
},
|
| 497 |
+
{
|
| 498 |
+
"epoch": 2.25,
|
| 499 |
+
"learning_rate": 1.388595564941922e-05,
|
| 500 |
+
"loss": 0.3434,
|
| 501 |
+
"step": 790
|
| 502 |
+
},
|
| 503 |
+
{
|
| 504 |
+
"epoch": 2.27,
|
| 505 |
+
"learning_rate": 1.3357972544878563e-05,
|
| 506 |
+
"loss": 0.3811,
|
| 507 |
+
"step": 800
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 2.3,
|
| 511 |
+
"learning_rate": 1.2829989440337912e-05,
|
| 512 |
+
"loss": 0.3483,
|
| 513 |
+
"step": 810
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 2.33,
|
| 517 |
+
"learning_rate": 1.2302006335797255e-05,
|
| 518 |
+
"loss": 0.369,
|
| 519 |
+
"step": 820
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 2.36,
|
| 523 |
+
"learning_rate": 1.17740232312566e-05,
|
| 524 |
+
"loss": 0.4224,
|
| 525 |
+
"step": 830
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 2.39,
|
| 529 |
+
"learning_rate": 1.1246040126715946e-05,
|
| 530 |
+
"loss": 0.3386,
|
| 531 |
+
"step": 840
|
| 532 |
+
},
|
| 533 |
+
{
|
| 534 |
+
"epoch": 2.42,
|
| 535 |
+
"learning_rate": 1.0718057022175291e-05,
|
| 536 |
+
"loss": 0.3552,
|
| 537 |
+
"step": 850
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 2.44,
|
| 541 |
+
"learning_rate": 1.0190073917634636e-05,
|
| 542 |
+
"loss": 0.3818,
|
| 543 |
+
"step": 860
|
| 544 |
+
},
|
| 545 |
+
{
|
| 546 |
+
"epoch": 2.47,
|
| 547 |
+
"learning_rate": 9.662090813093982e-06,
|
| 548 |
+
"loss": 0.3485,
|
| 549 |
+
"step": 870
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 2.5,
|
| 553 |
+
"learning_rate": 9.134107708553327e-06,
|
| 554 |
+
"loss": 0.3713,
|
| 555 |
+
"step": 880
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 2.53,
|
| 559 |
+
"learning_rate": 8.606124604012672e-06,
|
| 560 |
+
"loss": 0.3233,
|
| 561 |
+
"step": 890
|
| 562 |
+
},
|
| 563 |
+
{
|
| 564 |
+
"epoch": 2.56,
|
| 565 |
+
"learning_rate": 8.078141499472017e-06,
|
| 566 |
+
"loss": 0.3931,
|
| 567 |
+
"step": 900
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 2.59,
|
| 571 |
+
"learning_rate": 7.5501583949313625e-06,
|
| 572 |
+
"loss": 0.3595,
|
| 573 |
+
"step": 910
|
| 574 |
+
},
|
| 575 |
+
{
|
| 576 |
+
"epoch": 2.62,
|
| 577 |
+
"learning_rate": 7.022175290390708e-06,
|
| 578 |
+
"loss": 0.368,
|
| 579 |
+
"step": 920
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"epoch": 2.64,
|
| 583 |
+
"learning_rate": 6.494192185850054e-06,
|
| 584 |
+
"loss": 0.3494,
|
| 585 |
+
"step": 930
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 2.67,
|
| 589 |
+
"learning_rate": 5.966209081309398e-06,
|
| 590 |
+
"loss": 0.3388,
|
| 591 |
+
"step": 940
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 2.7,
|
| 595 |
+
"learning_rate": 5.438225976768744e-06,
|
| 596 |
+
"loss": 0.3311,
|
| 597 |
+
"step": 950
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 2.73,
|
| 601 |
+
"learning_rate": 4.910242872228089e-06,
|
| 602 |
+
"loss": 0.3335,
|
| 603 |
+
"step": 960
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"epoch": 2.76,
|
| 607 |
+
"learning_rate": 4.382259767687434e-06,
|
| 608 |
+
"loss": 0.3763,
|
| 609 |
+
"step": 970
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"epoch": 2.79,
|
| 613 |
+
"learning_rate": 3.854276663146779e-06,
|
| 614 |
+
"loss": 0.3199,
|
| 615 |
+
"step": 980
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 2.81,
|
| 619 |
+
"learning_rate": 3.326293558606125e-06,
|
| 620 |
+
"loss": 0.3904,
|
| 621 |
+
"step": 990
|
| 622 |
+
},
|
| 623 |
+
{
|
| 624 |
+
"epoch": 2.84,
|
| 625 |
+
"learning_rate": 2.79831045406547e-06,
|
| 626 |
+
"loss": 0.3893,
|
| 627 |
+
"step": 1000
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 2.87,
|
| 631 |
+
"learning_rate": 2.2703273495248154e-06,
|
| 632 |
+
"loss": 0.36,
|
| 633 |
+
"step": 1010
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 2.9,
|
| 637 |
+
"learning_rate": 1.7423442449841606e-06,
|
| 638 |
+
"loss": 0.3613,
|
| 639 |
+
"step": 1020
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 2.93,
|
| 643 |
+
"learning_rate": 1.2143611404435059e-06,
|
| 644 |
+
"loss": 0.3556,
|
| 645 |
+
"step": 1030
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 2.96,
|
| 649 |
+
"learning_rate": 6.863780359028511e-07,
|
| 650 |
+
"loss": 0.3358,
|
| 651 |
+
"step": 1040
|
| 652 |
+
},
|
| 653 |
+
{
|
| 654 |
+
"epoch": 2.99,
|
| 655 |
+
"learning_rate": 1.5839493136219642e-07,
|
| 656 |
+
"loss": 0.3201,
|
| 657 |
+
"step": 1050
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 2.99,
|
| 661 |
+
"eval_accuracy": 0.973,
|
| 662 |
+
"eval_loss": 0.079922154545784,
|
| 663 |
+
"eval_runtime": 20.1022,
|
| 664 |
+
"eval_samples_per_second": 248.73,
|
| 665 |
+
"eval_steps_per_second": 7.81,
|
| 666 |
+
"step": 1053
|
| 667 |
},
|
| 668 |
{
|
| 669 |
+
"epoch": 2.99,
|
| 670 |
+
"step": 1053,
|
| 671 |
+
"total_flos": 3.3497451642252165e+18,
|
| 672 |
+
"train_loss": 0.5413832122217669,
|
| 673 |
+
"train_runtime": 1364.9822,
|
| 674 |
+
"train_samples_per_second": 98.902,
|
| 675 |
+
"train_steps_per_second": 0.771
|
| 676 |
}
|
| 677 |
],
|
| 678 |
"logging_steps": 10,
|
| 679 |
+
"max_steps": 1053,
|
| 680 |
"num_train_epochs": 3,
|
| 681 |
"save_steps": 500,
|
| 682 |
+
"total_flos": 3.3497451642252165e+18,
|
| 683 |
"trial_name": null,
|
| 684 |
"trial_params": null
|
| 685 |
}
|