Commit
•
4f5f129
1
Parent(s):
289f5a7
End of training
Browse files- all_results.json +8 -8
- eval_nbest_predictions.json +2 -2
- eval_predictions.json +0 -0
- eval_results.json +3 -3
- runs/Oct03_14-05-24_g3072/events.out.tfevents.1664833033.g3072.47072.2 +3 -0
- train_results.json +5 -5
- trainer_state.json +71 -107
all_results.json
CHANGED
|
@@ -1,11 +1,11 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 2.0,
|
| 3 |
-
"eval_exact_match":
|
| 4 |
-
"eval_f1":
|
| 5 |
-
"eval_samples":
|
| 6 |
-
"train_loss":
|
| 7 |
-
"train_runtime":
|
| 8 |
-
"train_samples":
|
| 9 |
-
"train_samples_per_second":
|
| 10 |
-
"train_steps_per_second":
|
| 11 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 2.0,
|
| 3 |
+
"eval_exact_match": 68.40113528855251,
|
| 4 |
+
"eval_f1": 78.86662409665142,
|
| 5 |
+
"eval_samples": 10793,
|
| 6 |
+
"train_loss": 1.629506731429398,
|
| 7 |
+
"train_runtime": 1781.7081,
|
| 8 |
+
"train_samples": 88599,
|
| 9 |
+
"train_samples_per_second": 99.454,
|
| 10 |
+
"train_steps_per_second": 3.108
|
| 11 |
}
|
eval_nbest_predictions.json
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7bd6006a5f99175b8bdf67683d7a256f558361eadafaabe15f9825b72f3c6939
|
| 3 |
+
size 49559748
|
eval_predictions.json
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|
eval_results.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 2.0,
|
| 3 |
-
"eval_exact_match":
|
| 4 |
-
"eval_f1":
|
| 5 |
-
"eval_samples":
|
| 6 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 2.0,
|
| 3 |
+
"eval_exact_match": 68.40113528855251,
|
| 4 |
+
"eval_f1": 78.86662409665142,
|
| 5 |
+
"eval_samples": 10793
|
| 6 |
}
|
runs/Oct03_14-05-24_g3072/events.out.tfevents.1664833033.g3072.47072.2
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7d127a2ee602459608a7b31461ce2759bb12bec0b3a87ec210e5793fff5c7170
|
| 3 |
+
size 191
|
train_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 2.0,
|
| 3 |
-
"train_loss":
|
| 4 |
-
"train_runtime":
|
| 5 |
-
"train_samples":
|
| 6 |
-
"train_samples_per_second":
|
| 7 |
-
"train_steps_per_second":
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 2.0,
|
| 3 |
+
"train_loss": 1.629506731429398,
|
| 4 |
+
"train_runtime": 1781.7081,
|
| 5 |
+
"train_samples": 88599,
|
| 6 |
+
"train_samples_per_second": 99.454,
|
| 7 |
+
"train_steps_per_second": 3.108
|
| 8 |
}
|
trainer_state.json
CHANGED
|
@@ -10,412 +10,376 @@
|
|
| 10 |
{
|
| 11 |
"epoch": 0.04,
|
| 12 |
"learning_rate": 2.945828819068256e-05,
|
| 13 |
-
"loss":
|
| 14 |
"step": 100
|
| 15 |
},
|
| 16 |
{
|
| 17 |
"epoch": 0.07,
|
| 18 |
"learning_rate": 2.8916576381365114e-05,
|
| 19 |
-
"loss":
|
| 20 |
"step": 200
|
| 21 |
},
|
| 22 |
{
|
| 23 |
"epoch": 0.11,
|
| 24 |
"learning_rate": 2.837486457204767e-05,
|
| 25 |
-
"loss":
|
| 26 |
"step": 300
|
| 27 |
},
|
| 28 |
{
|
| 29 |
"epoch": 0.14,
|
| 30 |
"learning_rate": 2.783315276273023e-05,
|
| 31 |
-
"loss":
|
| 32 |
"step": 400
|
| 33 |
},
|
| 34 |
{
|
| 35 |
"epoch": 0.18,
|
| 36 |
"learning_rate": 2.7291440953412788e-05,
|
| 37 |
-
"loss":
|
| 38 |
-
"step": 500
|
| 39 |
-
},
|
| 40 |
-
{
|
| 41 |
-
"epoch": 0.18,
|
| 42 |
-
"eval_exact_match": 3.91674550614948,
|
| 43 |
-
"eval_f1": 7.935587309728481,
|
| 44 |
"step": 500
|
| 45 |
},
|
| 46 |
{
|
| 47 |
"epoch": 0.22,
|
| 48 |
"learning_rate": 2.674972914409534e-05,
|
| 49 |
-
"loss":
|
| 50 |
"step": 600
|
| 51 |
},
|
| 52 |
{
|
| 53 |
"epoch": 0.25,
|
| 54 |
"learning_rate": 2.6208017334777897e-05,
|
| 55 |
-
"loss":
|
| 56 |
"step": 700
|
| 57 |
},
|
| 58 |
{
|
| 59 |
"epoch": 0.29,
|
| 60 |
"learning_rate": 2.5666305525460455e-05,
|
| 61 |
-
"loss":
|
| 62 |
"step": 800
|
| 63 |
},
|
| 64 |
{
|
| 65 |
"epoch": 0.33,
|
| 66 |
"learning_rate": 2.5124593716143014e-05,
|
| 67 |
-
"loss":
|
| 68 |
"step": 900
|
| 69 |
},
|
| 70 |
{
|
| 71 |
"epoch": 0.36,
|
| 72 |
"learning_rate": 2.4582881906825568e-05,
|
| 73 |
-
"loss":
|
| 74 |
"step": 1000
|
| 75 |
},
|
| 76 |
{
|
| 77 |
"epoch": 0.36,
|
| 78 |
-
"eval_exact_match":
|
| 79 |
-
"eval_f1":
|
| 80 |
"step": 1000
|
| 81 |
},
|
| 82 |
{
|
| 83 |
"epoch": 0.4,
|
| 84 |
"learning_rate": 2.4041170097508126e-05,
|
| 85 |
-
"loss":
|
| 86 |
"step": 1100
|
| 87 |
},
|
| 88 |
{
|
| 89 |
"epoch": 0.43,
|
| 90 |
"learning_rate": 2.3499458288190684e-05,
|
| 91 |
-
"loss":
|
| 92 |
"step": 1200
|
| 93 |
},
|
| 94 |
{
|
| 95 |
"epoch": 0.47,
|
| 96 |
"learning_rate": 2.2957746478873243e-05,
|
| 97 |
-
"loss":
|
| 98 |
"step": 1300
|
| 99 |
},
|
| 100 |
{
|
| 101 |
"epoch": 0.51,
|
| 102 |
"learning_rate": 2.2416034669555797e-05,
|
| 103 |
-
"loss":
|
| 104 |
"step": 1400
|
| 105 |
},
|
| 106 |
{
|
| 107 |
"epoch": 0.54,
|
| 108 |
"learning_rate": 2.1874322860238355e-05,
|
| 109 |
-
"loss":
|
| 110 |
-
"step": 1500
|
| 111 |
-
},
|
| 112 |
-
{
|
| 113 |
-
"epoch": 0.54,
|
| 114 |
-
"eval_exact_match": 5.922421948912016,
|
| 115 |
-
"eval_f1": 11.530882619495905,
|
| 116 |
"step": 1500
|
| 117 |
},
|
| 118 |
{
|
| 119 |
"epoch": 0.58,
|
| 120 |
"learning_rate": 2.133261105092091e-05,
|
| 121 |
-
"loss":
|
| 122 |
"step": 1600
|
| 123 |
},
|
| 124 |
{
|
| 125 |
"epoch": 0.61,
|
| 126 |
"learning_rate": 2.0790899241603465e-05,
|
| 127 |
-
"loss":
|
| 128 |
"step": 1700
|
| 129 |
},
|
| 130 |
{
|
| 131 |
"epoch": 0.65,
|
| 132 |
"learning_rate": 2.0249187432286023e-05,
|
| 133 |
-
"loss":
|
| 134 |
"step": 1800
|
| 135 |
},
|
| 136 |
{
|
| 137 |
"epoch": 0.69,
|
| 138 |
"learning_rate": 1.970747562296858e-05,
|
| 139 |
-
"loss":
|
| 140 |
"step": 1900
|
| 141 |
},
|
| 142 |
{
|
| 143 |
"epoch": 0.72,
|
| 144 |
"learning_rate": 1.916576381365114e-05,
|
| 145 |
-
"loss":
|
| 146 |
"step": 2000
|
| 147 |
},
|
| 148 |
{
|
| 149 |
"epoch": 0.72,
|
| 150 |
-
"eval_exact_match":
|
| 151 |
-
"eval_f1":
|
| 152 |
"step": 2000
|
| 153 |
},
|
| 154 |
{
|
| 155 |
"epoch": 0.76,
|
| 156 |
"learning_rate": 1.8624052004333694e-05,
|
| 157 |
-
"loss":
|
| 158 |
"step": 2100
|
| 159 |
},
|
| 160 |
{
|
| 161 |
"epoch": 0.79,
|
| 162 |
"learning_rate": 1.8082340195016252e-05,
|
| 163 |
-
"loss":
|
| 164 |
"step": 2200
|
| 165 |
},
|
| 166 |
{
|
| 167 |
"epoch": 0.83,
|
| 168 |
"learning_rate": 1.754062838569881e-05,
|
| 169 |
-
"loss":
|
| 170 |
"step": 2300
|
| 171 |
},
|
| 172 |
{
|
| 173 |
"epoch": 0.87,
|
| 174 |
"learning_rate": 1.6998916576381368e-05,
|
| 175 |
-
"loss":
|
| 176 |
"step": 2400
|
| 177 |
},
|
| 178 |
{
|
| 179 |
"epoch": 0.9,
|
| 180 |
"learning_rate": 1.6457204767063923e-05,
|
| 181 |
-
"loss":
|
| 182 |
-
"step": 2500
|
| 183 |
-
},
|
| 184 |
-
{
|
| 185 |
-
"epoch": 0.9,
|
| 186 |
-
"eval_exact_match": 9.810785241248817,
|
| 187 |
-
"eval_f1": 17.14007217511169,
|
| 188 |
"step": 2500
|
| 189 |
},
|
| 190 |
{
|
| 191 |
"epoch": 0.94,
|
| 192 |
"learning_rate": 1.5915492957746478e-05,
|
| 193 |
-
"loss":
|
| 194 |
"step": 2600
|
| 195 |
},
|
| 196 |
{
|
| 197 |
"epoch": 0.98,
|
| 198 |
"learning_rate": 1.5373781148429036e-05,
|
| 199 |
-
"loss":
|
| 200 |
"step": 2700
|
| 201 |
},
|
| 202 |
{
|
| 203 |
"epoch": 1.01,
|
| 204 |
"learning_rate": 1.4832069339111594e-05,
|
| 205 |
-
"loss":
|
| 206 |
"step": 2800
|
| 207 |
},
|
| 208 |
{
|
| 209 |
"epoch": 1.05,
|
| 210 |
"learning_rate": 1.4290357529794148e-05,
|
| 211 |
-
"loss":
|
| 212 |
"step": 2900
|
| 213 |
},
|
| 214 |
{
|
| 215 |
"epoch": 1.08,
|
| 216 |
"learning_rate": 1.3748645720476707e-05,
|
| 217 |
-
"loss":
|
| 218 |
"step": 3000
|
| 219 |
},
|
| 220 |
{
|
| 221 |
"epoch": 1.08,
|
| 222 |
-
"eval_exact_match":
|
| 223 |
-
"eval_f1":
|
| 224 |
"step": 3000
|
| 225 |
},
|
| 226 |
{
|
| 227 |
"epoch": 1.12,
|
| 228 |
"learning_rate": 1.3206933911159263e-05,
|
| 229 |
-
"loss":
|
| 230 |
"step": 3100
|
| 231 |
},
|
| 232 |
{
|
| 233 |
"epoch": 1.16,
|
| 234 |
"learning_rate": 1.2665222101841821e-05,
|
| 235 |
-
"loss":
|
| 236 |
"step": 3200
|
| 237 |
},
|
| 238 |
{
|
| 239 |
"epoch": 1.19,
|
| 240 |
"learning_rate": 1.2123510292524377e-05,
|
| 241 |
-
"loss":
|
| 242 |
"step": 3300
|
| 243 |
},
|
| 244 |
{
|
| 245 |
"epoch": 1.23,
|
| 246 |
"learning_rate": 1.1581798483206934e-05,
|
| 247 |
-
"loss":
|
| 248 |
"step": 3400
|
| 249 |
},
|
| 250 |
{
|
| 251 |
"epoch": 1.26,
|
| 252 |
"learning_rate": 1.104008667388949e-05,
|
| 253 |
-
"loss":
|
| 254 |
-
"step": 3500
|
| 255 |
-
},
|
| 256 |
-
{
|
| 257 |
-
"epoch": 1.26,
|
| 258 |
-
"eval_exact_match": 18.68495742667928,
|
| 259 |
-
"eval_f1": 28.805039874191937,
|
| 260 |
"step": 3500
|
| 261 |
},
|
| 262 |
{
|
| 263 |
"epoch": 1.3,
|
| 264 |
"learning_rate": 1.0498374864572048e-05,
|
| 265 |
-
"loss":
|
| 266 |
"step": 3600
|
| 267 |
},
|
| 268 |
{
|
| 269 |
"epoch": 1.34,
|
| 270 |
"learning_rate": 9.956663055254605e-06,
|
| 271 |
-
"loss":
|
| 272 |
"step": 3700
|
| 273 |
},
|
| 274 |
{
|
| 275 |
"epoch": 1.37,
|
| 276 |
"learning_rate": 9.414951245937163e-06,
|
| 277 |
-
"loss":
|
| 278 |
"step": 3800
|
| 279 |
},
|
| 280 |
{
|
| 281 |
"epoch": 1.41,
|
| 282 |
"learning_rate": 8.873239436619718e-06,
|
| 283 |
-
"loss":
|
| 284 |
"step": 3900
|
| 285 |
},
|
| 286 |
{
|
| 287 |
"epoch": 1.44,
|
| 288 |
"learning_rate": 8.331527627302276e-06,
|
| 289 |
-
"loss":
|
| 290 |
"step": 4000
|
| 291 |
},
|
| 292 |
{
|
| 293 |
"epoch": 1.44,
|
| 294 |
-
"eval_exact_match":
|
| 295 |
-
"eval_f1":
|
| 296 |
"step": 4000
|
| 297 |
},
|
| 298 |
{
|
| 299 |
"epoch": 1.48,
|
| 300 |
"learning_rate": 7.789815817984832e-06,
|
| 301 |
-
"loss":
|
| 302 |
"step": 4100
|
| 303 |
},
|
| 304 |
{
|
| 305 |
"epoch": 1.52,
|
| 306 |
"learning_rate": 7.248104008667389e-06,
|
| 307 |
-
"loss":
|
| 308 |
"step": 4200
|
| 309 |
},
|
| 310 |
{
|
| 311 |
"epoch": 1.55,
|
| 312 |
"learning_rate": 6.706392199349946e-06,
|
| 313 |
-
"loss":
|
| 314 |
"step": 4300
|
| 315 |
},
|
| 316 |
{
|
| 317 |
"epoch": 1.59,
|
| 318 |
"learning_rate": 6.164680390032503e-06,
|
| 319 |
-
"loss":
|
| 320 |
"step": 4400
|
| 321 |
},
|
| 322 |
{
|
| 323 |
"epoch": 1.63,
|
| 324 |
"learning_rate": 5.62296858071506e-06,
|
| 325 |
-
"loss":
|
| 326 |
-
"step": 4500
|
| 327 |
-
},
|
| 328 |
-
{
|
| 329 |
-
"epoch": 1.63,
|
| 330 |
-
"eval_exact_match": 20.085146641438033,
|
| 331 |
-
"eval_f1": 30.456921284874575,
|
| 332 |
"step": 4500
|
| 333 |
},
|
| 334 |
{
|
| 335 |
"epoch": 1.66,
|
| 336 |
"learning_rate": 5.081256771397617e-06,
|
| 337 |
-
"loss":
|
| 338 |
"step": 4600
|
| 339 |
},
|
| 340 |
{
|
| 341 |
"epoch": 1.7,
|
| 342 |
"learning_rate": 4.539544962080174e-06,
|
| 343 |
-
"loss":
|
| 344 |
"step": 4700
|
| 345 |
},
|
| 346 |
{
|
| 347 |
"epoch": 1.73,
|
| 348 |
"learning_rate": 3.99783315276273e-06,
|
| 349 |
-
"loss":
|
| 350 |
"step": 4800
|
| 351 |
},
|
| 352 |
{
|
| 353 |
"epoch": 1.77,
|
| 354 |
"learning_rate": 3.456121343445287e-06,
|
| 355 |
-
"loss":
|
| 356 |
"step": 4900
|
| 357 |
},
|
| 358 |
{
|
| 359 |
"epoch": 1.81,
|
| 360 |
"learning_rate": 2.914409534127844e-06,
|
| 361 |
-
"loss":
|
| 362 |
"step": 5000
|
| 363 |
},
|
| 364 |
{
|
| 365 |
"epoch": 1.81,
|
| 366 |
-
"eval_exact_match":
|
| 367 |
-
"eval_f1":
|
| 368 |
"step": 5000
|
| 369 |
},
|
| 370 |
{
|
| 371 |
"epoch": 1.84,
|
| 372 |
"learning_rate": 2.3726977248104013e-06,
|
| 373 |
-
"loss":
|
| 374 |
"step": 5100
|
| 375 |
},
|
| 376 |
{
|
| 377 |
"epoch": 1.88,
|
| 378 |
"learning_rate": 1.8309859154929577e-06,
|
| 379 |
-
"loss":
|
| 380 |
"step": 5200
|
| 381 |
},
|
| 382 |
{
|
| 383 |
"epoch": 1.91,
|
| 384 |
"learning_rate": 1.2892741061755147e-06,
|
| 385 |
-
"loss":
|
| 386 |
"step": 5300
|
| 387 |
},
|
| 388 |
{
|
| 389 |
"epoch": 1.95,
|
| 390 |
"learning_rate": 7.475622968580716e-07,
|
| 391 |
-
"loss":
|
| 392 |
"step": 5400
|
| 393 |
},
|
| 394 |
{
|
| 395 |
"epoch": 1.99,
|
| 396 |
"learning_rate": 2.0585048754062841e-07,
|
| 397 |
-
"loss":
|
| 398 |
-
"step": 5500
|
| 399 |
-
},
|
| 400 |
-
{
|
| 401 |
-
"epoch": 1.99,
|
| 402 |
-
"eval_exact_match": 21.788079470198674,
|
| 403 |
-
"eval_f1": 32.61873772930997,
|
| 404 |
"step": 5500
|
| 405 |
},
|
| 406 |
{
|
| 407 |
"epoch": 2.0,
|
| 408 |
"step": 5538,
|
| 409 |
-
"total_flos":
|
| 410 |
-
"train_loss":
|
| 411 |
-
"train_runtime":
|
| 412 |
-
"train_samples_per_second":
|
| 413 |
-
"train_steps_per_second":
|
| 414 |
}
|
| 415 |
],
|
| 416 |
"max_steps": 5538,
|
| 417 |
"num_train_epochs": 2,
|
| 418 |
-
"total_flos":
|
| 419 |
"trial_name": null,
|
| 420 |
"trial_params": null
|
| 421 |
}
|
|
|
|
| 10 |
{
|
| 11 |
"epoch": 0.04,
|
| 12 |
"learning_rate": 2.945828819068256e-05,
|
| 13 |
+
"loss": 4.3199,
|
| 14 |
"step": 100
|
| 15 |
},
|
| 16 |
{
|
| 17 |
"epoch": 0.07,
|
| 18 |
"learning_rate": 2.8916576381365114e-05,
|
| 19 |
+
"loss": 2.7819,
|
| 20 |
"step": 200
|
| 21 |
},
|
| 22 |
{
|
| 23 |
"epoch": 0.11,
|
| 24 |
"learning_rate": 2.837486457204767e-05,
|
| 25 |
+
"loss": 2.4102,
|
| 26 |
"step": 300
|
| 27 |
},
|
| 28 |
{
|
| 29 |
"epoch": 0.14,
|
| 30 |
"learning_rate": 2.783315276273023e-05,
|
| 31 |
+
"loss": 2.2734,
|
| 32 |
"step": 400
|
| 33 |
},
|
| 34 |
{
|
| 35 |
"epoch": 0.18,
|
| 36 |
"learning_rate": 2.7291440953412788e-05,
|
| 37 |
+
"loss": 2.1137,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 38 |
"step": 500
|
| 39 |
},
|
| 40 |
{
|
| 41 |
"epoch": 0.22,
|
| 42 |
"learning_rate": 2.674972914409534e-05,
|
| 43 |
+
"loss": 2.0087,
|
| 44 |
"step": 600
|
| 45 |
},
|
| 46 |
{
|
| 47 |
"epoch": 0.25,
|
| 48 |
"learning_rate": 2.6208017334777897e-05,
|
| 49 |
+
"loss": 1.9438,
|
| 50 |
"step": 700
|
| 51 |
},
|
| 52 |
{
|
| 53 |
"epoch": 0.29,
|
| 54 |
"learning_rate": 2.5666305525460455e-05,
|
| 55 |
+
"loss": 1.9278,
|
| 56 |
"step": 800
|
| 57 |
},
|
| 58 |
{
|
| 59 |
"epoch": 0.33,
|
| 60 |
"learning_rate": 2.5124593716143014e-05,
|
| 61 |
+
"loss": 1.8629,
|
| 62 |
"step": 900
|
| 63 |
},
|
| 64 |
{
|
| 65 |
"epoch": 0.36,
|
| 66 |
"learning_rate": 2.4582881906825568e-05,
|
| 67 |
+
"loss": 1.8102,
|
| 68 |
"step": 1000
|
| 69 |
},
|
| 70 |
{
|
| 71 |
"epoch": 0.36,
|
| 72 |
+
"eval_exact_match": 60.84200567644276,
|
| 73 |
+
"eval_f1": 72.2063882008199,
|
| 74 |
"step": 1000
|
| 75 |
},
|
| 76 |
{
|
| 77 |
"epoch": 0.4,
|
| 78 |
"learning_rate": 2.4041170097508126e-05,
|
| 79 |
+
"loss": 1.8028,
|
| 80 |
"step": 1100
|
| 81 |
},
|
| 82 |
{
|
| 83 |
"epoch": 0.43,
|
| 84 |
"learning_rate": 2.3499458288190684e-05,
|
| 85 |
+
"loss": 1.7879,
|
| 86 |
"step": 1200
|
| 87 |
},
|
| 88 |
{
|
| 89 |
"epoch": 0.47,
|
| 90 |
"learning_rate": 2.2957746478873243e-05,
|
| 91 |
+
"loss": 1.7333,
|
| 92 |
"step": 1300
|
| 93 |
},
|
| 94 |
{
|
| 95 |
"epoch": 0.51,
|
| 96 |
"learning_rate": 2.2416034669555797e-05,
|
| 97 |
+
"loss": 1.6751,
|
| 98 |
"step": 1400
|
| 99 |
},
|
| 100 |
{
|
| 101 |
"epoch": 0.54,
|
| 102 |
"learning_rate": 2.1874322860238355e-05,
|
| 103 |
+
"loss": 1.6925,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 104 |
"step": 1500
|
| 105 |
},
|
| 106 |
{
|
| 107 |
"epoch": 0.58,
|
| 108 |
"learning_rate": 2.133261105092091e-05,
|
| 109 |
+
"loss": 1.6655,
|
| 110 |
"step": 1600
|
| 111 |
},
|
| 112 |
{
|
| 113 |
"epoch": 0.61,
|
| 114 |
"learning_rate": 2.0790899241603465e-05,
|
| 115 |
+
"loss": 1.6675,
|
| 116 |
"step": 1700
|
| 117 |
},
|
| 118 |
{
|
| 119 |
"epoch": 0.65,
|
| 120 |
"learning_rate": 2.0249187432286023e-05,
|
| 121 |
+
"loss": 1.6169,
|
| 122 |
"step": 1800
|
| 123 |
},
|
| 124 |
{
|
| 125 |
"epoch": 0.69,
|
| 126 |
"learning_rate": 1.970747562296858e-05,
|
| 127 |
+
"loss": 1.6765,
|
| 128 |
"step": 1900
|
| 129 |
},
|
| 130 |
{
|
| 131 |
"epoch": 0.72,
|
| 132 |
"learning_rate": 1.916576381365114e-05,
|
| 133 |
+
"loss": 1.6591,
|
| 134 |
"step": 2000
|
| 135 |
},
|
| 136 |
{
|
| 137 |
"epoch": 0.72,
|
| 138 |
+
"eval_exact_match": 64.56953642384106,
|
| 139 |
+
"eval_f1": 75.74673565914026,
|
| 140 |
"step": 2000
|
| 141 |
},
|
| 142 |
{
|
| 143 |
"epoch": 0.76,
|
| 144 |
"learning_rate": 1.8624052004333694e-05,
|
| 145 |
+
"loss": 1.6145,
|
| 146 |
"step": 2100
|
| 147 |
},
|
| 148 |
{
|
| 149 |
"epoch": 0.79,
|
| 150 |
"learning_rate": 1.8082340195016252e-05,
|
| 151 |
+
"loss": 1.569,
|
| 152 |
"step": 2200
|
| 153 |
},
|
| 154 |
{
|
| 155 |
"epoch": 0.83,
|
| 156 |
"learning_rate": 1.754062838569881e-05,
|
| 157 |
+
"loss": 1.5951,
|
| 158 |
"step": 2300
|
| 159 |
},
|
| 160 |
{
|
| 161 |
"epoch": 0.87,
|
| 162 |
"learning_rate": 1.6998916576381368e-05,
|
| 163 |
+
"loss": 1.5693,
|
| 164 |
"step": 2400
|
| 165 |
},
|
| 166 |
{
|
| 167 |
"epoch": 0.9,
|
| 168 |
"learning_rate": 1.6457204767063923e-05,
|
| 169 |
+
"loss": 1.5276,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 170 |
"step": 2500
|
| 171 |
},
|
| 172 |
{
|
| 173 |
"epoch": 0.94,
|
| 174 |
"learning_rate": 1.5915492957746478e-05,
|
| 175 |
+
"loss": 1.5738,
|
| 176 |
"step": 2600
|
| 177 |
},
|
| 178 |
{
|
| 179 |
"epoch": 0.98,
|
| 180 |
"learning_rate": 1.5373781148429036e-05,
|
| 181 |
+
"loss": 1.5839,
|
| 182 |
"step": 2700
|
| 183 |
},
|
| 184 |
{
|
| 185 |
"epoch": 1.01,
|
| 186 |
"learning_rate": 1.4832069339111594e-05,
|
| 187 |
+
"loss": 1.5627,
|
| 188 |
"step": 2800
|
| 189 |
},
|
| 190 |
{
|
| 191 |
"epoch": 1.05,
|
| 192 |
"learning_rate": 1.4290357529794148e-05,
|
| 193 |
+
"loss": 1.3985,
|
| 194 |
"step": 2900
|
| 195 |
},
|
| 196 |
{
|
| 197 |
"epoch": 1.08,
|
| 198 |
"learning_rate": 1.3748645720476707e-05,
|
| 199 |
+
"loss": 1.401,
|
| 200 |
"step": 3000
|
| 201 |
},
|
| 202 |
{
|
| 203 |
"epoch": 1.08,
|
| 204 |
+
"eval_exact_match": 66.69820245979186,
|
| 205 |
+
"eval_f1": 77.4518814156339,
|
| 206 |
"step": 3000
|
| 207 |
},
|
| 208 |
{
|
| 209 |
"epoch": 1.12,
|
| 210 |
"learning_rate": 1.3206933911159263e-05,
|
| 211 |
+
"loss": 1.374,
|
| 212 |
"step": 3100
|
| 213 |
},
|
| 214 |
{
|
| 215 |
"epoch": 1.16,
|
| 216 |
"learning_rate": 1.2665222101841821e-05,
|
| 217 |
+
"loss": 1.3751,
|
| 218 |
"step": 3200
|
| 219 |
},
|
| 220 |
{
|
| 221 |
"epoch": 1.19,
|
| 222 |
"learning_rate": 1.2123510292524377e-05,
|
| 223 |
+
"loss": 1.3844,
|
| 224 |
"step": 3300
|
| 225 |
},
|
| 226 |
{
|
| 227 |
"epoch": 1.23,
|
| 228 |
"learning_rate": 1.1581798483206934e-05,
|
| 229 |
+
"loss": 1.3643,
|
| 230 |
"step": 3400
|
| 231 |
},
|
| 232 |
{
|
| 233 |
"epoch": 1.26,
|
| 234 |
"learning_rate": 1.104008667388949e-05,
|
| 235 |
+
"loss": 1.4097,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 236 |
"step": 3500
|
| 237 |
},
|
| 238 |
{
|
| 239 |
"epoch": 1.3,
|
| 240 |
"learning_rate": 1.0498374864572048e-05,
|
| 241 |
+
"loss": 1.3747,
|
| 242 |
"step": 3600
|
| 243 |
},
|
| 244 |
{
|
| 245 |
"epoch": 1.34,
|
| 246 |
"learning_rate": 9.956663055254605e-06,
|
| 247 |
+
"loss": 1.3667,
|
| 248 |
"step": 3700
|
| 249 |
},
|
| 250 |
{
|
| 251 |
"epoch": 1.37,
|
| 252 |
"learning_rate": 9.414951245937163e-06,
|
| 253 |
+
"loss": 1.333,
|
| 254 |
"step": 3800
|
| 255 |
},
|
| 256 |
{
|
| 257 |
"epoch": 1.41,
|
| 258 |
"learning_rate": 8.873239436619718e-06,
|
| 259 |
+
"loss": 1.384,
|
| 260 |
"step": 3900
|
| 261 |
},
|
| 262 |
{
|
| 263 |
"epoch": 1.44,
|
| 264 |
"learning_rate": 8.331527627302276e-06,
|
| 265 |
+
"loss": 1.3544,
|
| 266 |
"step": 4000
|
| 267 |
},
|
| 268 |
{
|
| 269 |
"epoch": 1.44,
|
| 270 |
+
"eval_exact_match": 67.53074739829707,
|
| 271 |
+
"eval_f1": 78.10232802082328,
|
| 272 |
"step": 4000
|
| 273 |
},
|
| 274 |
{
|
| 275 |
"epoch": 1.48,
|
| 276 |
"learning_rate": 7.789815817984832e-06,
|
| 277 |
+
"loss": 1.3762,
|
| 278 |
"step": 4100
|
| 279 |
},
|
| 280 |
{
|
| 281 |
"epoch": 1.52,
|
| 282 |
"learning_rate": 7.248104008667389e-06,
|
| 283 |
+
"loss": 1.3628,
|
| 284 |
"step": 4200
|
| 285 |
},
|
| 286 |
{
|
| 287 |
"epoch": 1.55,
|
| 288 |
"learning_rate": 6.706392199349946e-06,
|
| 289 |
+
"loss": 1.3639,
|
| 290 |
"step": 4300
|
| 291 |
},
|
| 292 |
{
|
| 293 |
"epoch": 1.59,
|
| 294 |
"learning_rate": 6.164680390032503e-06,
|
| 295 |
+
"loss": 1.397,
|
| 296 |
"step": 4400
|
| 297 |
},
|
| 298 |
{
|
| 299 |
"epoch": 1.63,
|
| 300 |
"learning_rate": 5.62296858071506e-06,
|
| 301 |
+
"loss": 1.3298,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 302 |
"step": 4500
|
| 303 |
},
|
| 304 |
{
|
| 305 |
"epoch": 1.66,
|
| 306 |
"learning_rate": 5.081256771397617e-06,
|
| 307 |
+
"loss": 1.3641,
|
| 308 |
"step": 4600
|
| 309 |
},
|
| 310 |
{
|
| 311 |
"epoch": 1.7,
|
| 312 |
"learning_rate": 4.539544962080174e-06,
|
| 313 |
+
"loss": 1.3164,
|
| 314 |
"step": 4700
|
| 315 |
},
|
| 316 |
{
|
| 317 |
"epoch": 1.73,
|
| 318 |
"learning_rate": 3.99783315276273e-06,
|
| 319 |
+
"loss": 1.3582,
|
| 320 |
"step": 4800
|
| 321 |
},
|
| 322 |
{
|
| 323 |
"epoch": 1.77,
|
| 324 |
"learning_rate": 3.456121343445287e-06,
|
| 325 |
+
"loss": 1.3273,
|
| 326 |
"step": 4900
|
| 327 |
},
|
| 328 |
{
|
| 329 |
"epoch": 1.81,
|
| 330 |
"learning_rate": 2.914409534127844e-06,
|
| 331 |
+
"loss": 1.3176,
|
| 332 |
"step": 5000
|
| 333 |
},
|
| 334 |
{
|
| 335 |
"epoch": 1.81,
|
| 336 |
+
"eval_exact_match": 68.1456953642384,
|
| 337 |
+
"eval_f1": 78.69175735985112,
|
| 338 |
"step": 5000
|
| 339 |
},
|
| 340 |
{
|
| 341 |
"epoch": 1.84,
|
| 342 |
"learning_rate": 2.3726977248104013e-06,
|
| 343 |
+
"loss": 1.2995,
|
| 344 |
"step": 5100
|
| 345 |
},
|
| 346 |
{
|
| 347 |
"epoch": 1.88,
|
| 348 |
"learning_rate": 1.8309859154929577e-06,
|
| 349 |
+
"loss": 1.3327,
|
| 350 |
"step": 5200
|
| 351 |
},
|
| 352 |
{
|
| 353 |
"epoch": 1.91,
|
| 354 |
"learning_rate": 1.2892741061755147e-06,
|
| 355 |
+
"loss": 1.3254,
|
| 356 |
"step": 5300
|
| 357 |
},
|
| 358 |
{
|
| 359 |
"epoch": 1.95,
|
| 360 |
"learning_rate": 7.475622968580716e-07,
|
| 361 |
+
"loss": 1.36,
|
| 362 |
"step": 5400
|
| 363 |
},
|
| 364 |
{
|
| 365 |
"epoch": 1.99,
|
| 366 |
"learning_rate": 2.0585048754062841e-07,
|
| 367 |
+
"loss": 1.3627,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 368 |
"step": 5500
|
| 369 |
},
|
| 370 |
{
|
| 371 |
"epoch": 2.0,
|
| 372 |
"step": 5538,
|
| 373 |
+
"total_flos": 2.809773228766003e+16,
|
| 374 |
+
"train_loss": 1.629506731429398,
|
| 375 |
+
"train_runtime": 1781.7081,
|
| 376 |
+
"train_samples_per_second": 99.454,
|
| 377 |
+
"train_steps_per_second": 3.108
|
| 378 |
}
|
| 379 |
],
|
| 380 |
"max_steps": 5538,
|
| 381 |
"num_train_epochs": 2,
|
| 382 |
+
"total_flos": 2.809773228766003e+16,
|
| 383 |
"trial_name": null,
|
| 384 |
"trial_params": null
|
| 385 |
}
|