train with metrics
Browse files- all_results.json +14 -4
- eval_results.json +13 -0
- pytorch_model.bin +1 -1
- tokenizer.json +0 -0
- train_results.json +4 -4
- trainer_state.json +122 -122
- training_args.bin +1 -1
all_results.json
CHANGED
|
@@ -1,8 +1,18 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 2.0,
|
| 3 |
-
"
|
| 4 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 5 |
"train_samples": 87599,
|
| 6 |
-
"train_samples_per_second":
|
| 7 |
-
"train_steps_per_second":
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 2.0,
|
| 3 |
+
"eval_gen_len": 9.2974,
|
| 4 |
+
"eval_loss": 0.5938774347305298,
|
| 5 |
+
"eval_rouge1": 17.5052,
|
| 6 |
+
"eval_rouge2": 5.8714,
|
| 7 |
+
"eval_rougeL": 17.4487,
|
| 8 |
+
"eval_rougeLsum": 17.4528,
|
| 9 |
+
"eval_runtime": 905.9638,
|
| 10 |
+
"eval_samples": 10570,
|
| 11 |
+
"eval_samples_per_second": 11.667,
|
| 12 |
+
"eval_steps_per_second": 3.89,
|
| 13 |
+
"train_loss": 0.1063714675380759,
|
| 14 |
+
"train_runtime": 3374.8833,
|
| 15 |
"train_samples": 87599,
|
| 16 |
+
"train_samples_per_second": 51.912,
|
| 17 |
+
"train_steps_per_second": 17.304
|
| 18 |
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.0,
|
| 3 |
+
"eval_gen_len": 9.2974,
|
| 4 |
+
"eval_loss": 0.5938774347305298,
|
| 5 |
+
"eval_rouge1": 17.5052,
|
| 6 |
+
"eval_rouge2": 5.8714,
|
| 7 |
+
"eval_rougeL": 17.4487,
|
| 8 |
+
"eval_rougeLsum": 17.4528,
|
| 9 |
+
"eval_runtime": 905.9638,
|
| 10 |
+
"eval_samples": 10570,
|
| 11 |
+
"eval_samples_per_second": 11.667,
|
| 12 |
+
"eval_steps_per_second": 3.89
|
| 13 |
+
}
|
pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 990284749
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:befb20cfeecf2f9507ccac39fe16381da87a9ebc601a837d7145634b51041857
|
| 3 |
size 990284749
|
tokenizer.json
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|
train_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 2.0,
|
| 3 |
-
"train_loss": 0.
|
| 4 |
-
"train_runtime":
|
| 5 |
"train_samples": 87599,
|
| 6 |
-
"train_samples_per_second":
|
| 7 |
-
"train_steps_per_second":
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 2.0,
|
| 3 |
+
"train_loss": 0.1063714675380759,
|
| 4 |
+
"train_runtime": 3374.8833,
|
| 5 |
"train_samples": 87599,
|
| 6 |
+
"train_samples_per_second": 51.912,
|
| 7 |
+
"train_steps_per_second": 17.304
|
| 8 |
}
|
trainer_state.json
CHANGED
|
@@ -10,712 +10,712 @@
|
|
| 10 |
{
|
| 11 |
"epoch": 0.02,
|
| 12 |
"learning_rate": 2.974315068493151e-05,
|
| 13 |
-
"loss": 2.
|
| 14 |
"step": 500
|
| 15 |
},
|
| 16 |
{
|
| 17 |
"epoch": 0.03,
|
| 18 |
"learning_rate": 2.9486301369863017e-05,
|
| 19 |
-
"loss": 1.
|
| 20 |
"step": 1000
|
| 21 |
},
|
| 22 |
{
|
| 23 |
"epoch": 0.05,
|
| 24 |
"learning_rate": 2.922945205479452e-05,
|
| 25 |
-
"loss":
|
| 26 |
"step": 1500
|
| 27 |
},
|
| 28 |
{
|
| 29 |
"epoch": 0.07,
|
| 30 |
"learning_rate": 2.8972602739726026e-05,
|
| 31 |
-
"loss": 0.
|
| 32 |
"step": 2000
|
| 33 |
},
|
| 34 |
{
|
| 35 |
"epoch": 0.09,
|
| 36 |
"learning_rate": 2.8715753424657534e-05,
|
| 37 |
-
"loss": 0.
|
| 38 |
"step": 2500
|
| 39 |
},
|
| 40 |
{
|
| 41 |
"epoch": 0.1,
|
| 42 |
"learning_rate": 2.8458904109589042e-05,
|
| 43 |
-
"loss": 0.
|
| 44 |
"step": 3000
|
| 45 |
},
|
| 46 |
{
|
| 47 |
"epoch": 0.12,
|
| 48 |
"learning_rate": 2.820205479452055e-05,
|
| 49 |
-
"loss": 0.
|
| 50 |
"step": 3500
|
| 51 |
},
|
| 52 |
{
|
| 53 |
"epoch": 0.14,
|
| 54 |
"learning_rate": 2.7945205479452054e-05,
|
| 55 |
-
"loss": 0.
|
| 56 |
"step": 4000
|
| 57 |
},
|
| 58 |
{
|
| 59 |
"epoch": 0.15,
|
| 60 |
"learning_rate": 2.7688356164383562e-05,
|
| 61 |
-
"loss": 0.
|
| 62 |
"step": 4500
|
| 63 |
},
|
| 64 |
{
|
| 65 |
"epoch": 0.17,
|
| 66 |
"learning_rate": 2.743150684931507e-05,
|
| 67 |
-
"loss": 0.
|
| 68 |
"step": 5000
|
| 69 |
},
|
| 70 |
{
|
| 71 |
"epoch": 0.19,
|
| 72 |
"learning_rate": 2.7174657534246575e-05,
|
| 73 |
-
"loss": 0.
|
| 74 |
"step": 5500
|
| 75 |
},
|
| 76 |
{
|
| 77 |
"epoch": 0.21,
|
| 78 |
"learning_rate": 2.6917808219178083e-05,
|
| 79 |
-
"loss": 0.
|
| 80 |
"step": 6000
|
| 81 |
},
|
| 82 |
{
|
| 83 |
"epoch": 0.22,
|
| 84 |
"learning_rate": 2.666095890410959e-05,
|
| 85 |
-
"loss": 0.
|
| 86 |
"step": 6500
|
| 87 |
},
|
| 88 |
{
|
| 89 |
"epoch": 0.24,
|
| 90 |
"learning_rate": 2.6404109589041096e-05,
|
| 91 |
-
"loss": 0.
|
| 92 |
"step": 7000
|
| 93 |
},
|
| 94 |
{
|
| 95 |
"epoch": 0.26,
|
| 96 |
"learning_rate": 2.6147260273972604e-05,
|
| 97 |
-
"loss": 0.
|
| 98 |
"step": 7500
|
| 99 |
},
|
| 100 |
{
|
| 101 |
"epoch": 0.27,
|
| 102 |
"learning_rate": 2.589041095890411e-05,
|
| 103 |
-
"loss": 0.
|
| 104 |
"step": 8000
|
| 105 |
},
|
| 106 |
{
|
| 107 |
"epoch": 0.29,
|
| 108 |
"learning_rate": 2.563356164383562e-05,
|
| 109 |
-
"loss": 0.
|
| 110 |
"step": 8500
|
| 111 |
},
|
| 112 |
{
|
| 113 |
"epoch": 0.31,
|
| 114 |
"learning_rate": 2.5376712328767124e-05,
|
| 115 |
-
"loss": 0.
|
| 116 |
"step": 9000
|
| 117 |
},
|
| 118 |
{
|
| 119 |
"epoch": 0.33,
|
| 120 |
"learning_rate": 2.511986301369863e-05,
|
| 121 |
-
"loss": 0.
|
| 122 |
"step": 9500
|
| 123 |
},
|
| 124 |
{
|
| 125 |
"epoch": 0.34,
|
| 126 |
"learning_rate": 2.4863013698630137e-05,
|
| 127 |
-
"loss": 0.
|
| 128 |
"step": 10000
|
| 129 |
},
|
| 130 |
{
|
| 131 |
"epoch": 0.36,
|
| 132 |
"learning_rate": 2.4606164383561645e-05,
|
| 133 |
-
"loss": 0.
|
| 134 |
"step": 10500
|
| 135 |
},
|
| 136 |
{
|
| 137 |
"epoch": 0.38,
|
| 138 |
"learning_rate": 2.4349315068493153e-05,
|
| 139 |
-
"loss": 0.
|
| 140 |
"step": 11000
|
| 141 |
},
|
| 142 |
{
|
| 143 |
"epoch": 0.39,
|
| 144 |
"learning_rate": 2.409246575342466e-05,
|
| 145 |
-
"loss": 0.
|
| 146 |
"step": 11500
|
| 147 |
},
|
| 148 |
{
|
| 149 |
"epoch": 0.41,
|
| 150 |
"learning_rate": 2.3835616438356165e-05,
|
| 151 |
-
"loss": 0.
|
| 152 |
"step": 12000
|
| 153 |
},
|
| 154 |
{
|
| 155 |
"epoch": 0.43,
|
| 156 |
"learning_rate": 2.357876712328767e-05,
|
| 157 |
-
"loss": 0.
|
| 158 |
"step": 12500
|
| 159 |
},
|
| 160 |
{
|
| 161 |
"epoch": 0.45,
|
| 162 |
"learning_rate": 2.3321917808219178e-05,
|
| 163 |
-
"loss": 0.
|
| 164 |
"step": 13000
|
| 165 |
},
|
| 166 |
{
|
| 167 |
"epoch": 0.46,
|
| 168 |
"learning_rate": 2.3065068493150686e-05,
|
| 169 |
-
"loss": 0.
|
| 170 |
"step": 13500
|
| 171 |
},
|
| 172 |
{
|
| 173 |
"epoch": 0.48,
|
| 174 |
"learning_rate": 2.2808219178082194e-05,
|
| 175 |
-
"loss": 0.
|
| 176 |
"step": 14000
|
| 177 |
},
|
| 178 |
{
|
| 179 |
"epoch": 0.5,
|
| 180 |
"learning_rate": 2.25513698630137e-05,
|
| 181 |
-
"loss": 0.
|
| 182 |
"step": 14500
|
| 183 |
},
|
| 184 |
{
|
| 185 |
"epoch": 0.51,
|
| 186 |
"learning_rate": 2.2294520547945206e-05,
|
| 187 |
-
"loss": 0.
|
| 188 |
"step": 15000
|
| 189 |
},
|
| 190 |
{
|
| 191 |
"epoch": 0.53,
|
| 192 |
"learning_rate": 2.203767123287671e-05,
|
| 193 |
-
"loss": 0.
|
| 194 |
"step": 15500
|
| 195 |
},
|
| 196 |
{
|
| 197 |
"epoch": 0.55,
|
| 198 |
"learning_rate": 2.178082191780822e-05,
|
| 199 |
-
"loss": 0.
|
| 200 |
"step": 16000
|
| 201 |
},
|
| 202 |
{
|
| 203 |
"epoch": 0.57,
|
| 204 |
"learning_rate": 2.1523972602739727e-05,
|
| 205 |
-
"loss": 0.
|
| 206 |
"step": 16500
|
| 207 |
},
|
| 208 |
{
|
| 209 |
"epoch": 0.58,
|
| 210 |
"learning_rate": 2.126712328767123e-05,
|
| 211 |
-
"loss": 0.
|
| 212 |
"step": 17000
|
| 213 |
},
|
| 214 |
{
|
| 215 |
"epoch": 0.6,
|
| 216 |
"learning_rate": 2.101027397260274e-05,
|
| 217 |
-
"loss": 0.
|
| 218 |
"step": 17500
|
| 219 |
},
|
| 220 |
{
|
| 221 |
"epoch": 0.62,
|
| 222 |
"learning_rate": 2.0753424657534248e-05,
|
| 223 |
-
"loss": 0.
|
| 224 |
"step": 18000
|
| 225 |
},
|
| 226 |
{
|
| 227 |
"epoch": 0.63,
|
| 228 |
"learning_rate": 2.0496575342465756e-05,
|
| 229 |
-
"loss": 0.
|
| 230 |
"step": 18500
|
| 231 |
},
|
| 232 |
{
|
| 233 |
"epoch": 0.65,
|
| 234 |
"learning_rate": 2.023972602739726e-05,
|
| 235 |
-
"loss": 0.
|
| 236 |
"step": 19000
|
| 237 |
},
|
| 238 |
{
|
| 239 |
"epoch": 0.67,
|
| 240 |
"learning_rate": 1.9982876712328768e-05,
|
| 241 |
-
"loss": 0.
|
| 242 |
"step": 19500
|
| 243 |
},
|
| 244 |
{
|
| 245 |
"epoch": 0.68,
|
| 246 |
"learning_rate": 1.9726027397260273e-05,
|
| 247 |
-
"loss": 0.
|
| 248 |
"step": 20000
|
| 249 |
},
|
| 250 |
{
|
| 251 |
"epoch": 0.7,
|
| 252 |
"learning_rate": 1.946917808219178e-05,
|
| 253 |
-
"loss": 0.
|
| 254 |
"step": 20500
|
| 255 |
},
|
| 256 |
{
|
| 257 |
"epoch": 0.72,
|
| 258 |
"learning_rate": 1.921232876712329e-05,
|
| 259 |
-
"loss": 0.
|
| 260 |
"step": 21000
|
| 261 |
},
|
| 262 |
{
|
| 263 |
"epoch": 0.74,
|
| 264 |
"learning_rate": 1.8955479452054797e-05,
|
| 265 |
-
"loss": 0.
|
| 266 |
"step": 21500
|
| 267 |
},
|
| 268 |
{
|
| 269 |
"epoch": 0.75,
|
| 270 |
"learning_rate": 1.8698630136986305e-05,
|
| 271 |
-
"loss": 0.
|
| 272 |
"step": 22000
|
| 273 |
},
|
| 274 |
{
|
| 275 |
"epoch": 0.77,
|
| 276 |
"learning_rate": 1.8441780821917806e-05,
|
| 277 |
-
"loss": 0.
|
| 278 |
"step": 22500
|
| 279 |
},
|
| 280 |
{
|
| 281 |
"epoch": 0.79,
|
| 282 |
"learning_rate": 1.8184931506849314e-05,
|
| 283 |
-
"loss": 0.
|
| 284 |
"step": 23000
|
| 285 |
},
|
| 286 |
{
|
| 287 |
"epoch": 0.8,
|
| 288 |
"learning_rate": 1.7928082191780822e-05,
|
| 289 |
-
"loss": 0.
|
| 290 |
"step": 23500
|
| 291 |
},
|
| 292 |
{
|
| 293 |
"epoch": 0.82,
|
| 294 |
"learning_rate": 1.767123287671233e-05,
|
| 295 |
-
"loss": 0.
|
| 296 |
"step": 24000
|
| 297 |
},
|
| 298 |
{
|
| 299 |
"epoch": 0.84,
|
| 300 |
"learning_rate": 1.7414383561643838e-05,
|
| 301 |
-
"loss": 0.
|
| 302 |
"step": 24500
|
| 303 |
},
|
| 304 |
{
|
| 305 |
"epoch": 0.86,
|
| 306 |
"learning_rate": 1.7157534246575342e-05,
|
| 307 |
-
"loss": 0.
|
| 308 |
"step": 25000
|
| 309 |
},
|
| 310 |
{
|
| 311 |
"epoch": 0.87,
|
| 312 |
"learning_rate": 1.690068493150685e-05,
|
| 313 |
-
"loss": 0.
|
| 314 |
"step": 25500
|
| 315 |
},
|
| 316 |
{
|
| 317 |
"epoch": 0.89,
|
| 318 |
"learning_rate": 1.6643835616438355e-05,
|
| 319 |
-
"loss": 0.
|
| 320 |
"step": 26000
|
| 321 |
},
|
| 322 |
{
|
| 323 |
"epoch": 0.91,
|
| 324 |
"learning_rate": 1.6386986301369863e-05,
|
| 325 |
-
"loss": 0.
|
| 326 |
"step": 26500
|
| 327 |
},
|
| 328 |
{
|
| 329 |
"epoch": 0.92,
|
| 330 |
"learning_rate": 1.613013698630137e-05,
|
| 331 |
-
"loss": 0.
|
| 332 |
"step": 27000
|
| 333 |
},
|
| 334 |
{
|
| 335 |
"epoch": 0.94,
|
| 336 |
"learning_rate": 1.5873287671232876e-05,
|
| 337 |
-
"loss": 0.
|
| 338 |
"step": 27500
|
| 339 |
},
|
| 340 |
{
|
| 341 |
"epoch": 0.96,
|
| 342 |
"learning_rate": 1.5616438356164384e-05,
|
| 343 |
-
"loss": 0.
|
| 344 |
"step": 28000
|
| 345 |
},
|
| 346 |
{
|
| 347 |
"epoch": 0.98,
|
| 348 |
"learning_rate": 1.535958904109589e-05,
|
| 349 |
-
"loss": 0.
|
| 350 |
"step": 28500
|
| 351 |
},
|
| 352 |
{
|
| 353 |
"epoch": 0.99,
|
| 354 |
"learning_rate": 1.5102739726027398e-05,
|
| 355 |
-
"loss": 0.
|
| 356 |
"step": 29000
|
| 357 |
},
|
| 358 |
{
|
| 359 |
"epoch": 1.01,
|
| 360 |
"learning_rate": 1.4845890410958904e-05,
|
| 361 |
-
"loss": 0.
|
| 362 |
"step": 29500
|
| 363 |
},
|
| 364 |
{
|
| 365 |
"epoch": 1.03,
|
| 366 |
"learning_rate": 1.4589041095890412e-05,
|
| 367 |
-
"loss": 0.
|
| 368 |
"step": 30000
|
| 369 |
},
|
| 370 |
{
|
| 371 |
"epoch": 1.04,
|
| 372 |
"learning_rate": 1.4332191780821918e-05,
|
| 373 |
-
"loss": 0.
|
| 374 |
"step": 30500
|
| 375 |
},
|
| 376 |
{
|
| 377 |
"epoch": 1.06,
|
| 378 |
"learning_rate": 1.4075342465753425e-05,
|
| 379 |
-
"loss": 0.
|
| 380 |
"step": 31000
|
| 381 |
},
|
| 382 |
{
|
| 383 |
"epoch": 1.08,
|
| 384 |
"learning_rate": 1.3818493150684933e-05,
|
| 385 |
-
"loss": 0.
|
| 386 |
"step": 31500
|
| 387 |
},
|
| 388 |
{
|
| 389 |
"epoch": 1.1,
|
| 390 |
"learning_rate": 1.3561643835616437e-05,
|
| 391 |
-
"loss": 0.
|
| 392 |
"step": 32000
|
| 393 |
},
|
| 394 |
{
|
| 395 |
"epoch": 1.11,
|
| 396 |
"learning_rate": 1.3304794520547945e-05,
|
| 397 |
-
"loss": 0.
|
| 398 |
"step": 32500
|
| 399 |
},
|
| 400 |
{
|
| 401 |
"epoch": 1.13,
|
| 402 |
"learning_rate": 1.3047945205479453e-05,
|
| 403 |
-
"loss": 0.
|
| 404 |
"step": 33000
|
| 405 |
},
|
| 406 |
{
|
| 407 |
"epoch": 1.15,
|
| 408 |
"learning_rate": 1.279109589041096e-05,
|
| 409 |
-
"loss": 0.
|
| 410 |
"step": 33500
|
| 411 |
},
|
| 412 |
{
|
| 413 |
"epoch": 1.16,
|
| 414 |
"learning_rate": 1.2534246575342466e-05,
|
| 415 |
-
"loss": 0.
|
| 416 |
"step": 34000
|
| 417 |
},
|
| 418 |
{
|
| 419 |
"epoch": 1.18,
|
| 420 |
"learning_rate": 1.2277397260273974e-05,
|
| 421 |
-
"loss": 0.
|
| 422 |
"step": 34500
|
| 423 |
},
|
| 424 |
{
|
| 425 |
"epoch": 1.2,
|
| 426 |
"learning_rate": 1.202054794520548e-05,
|
| 427 |
-
"loss": 0.
|
| 428 |
"step": 35000
|
| 429 |
},
|
| 430 |
{
|
| 431 |
"epoch": 1.22,
|
| 432 |
"learning_rate": 1.1763698630136986e-05,
|
| 433 |
-
"loss": 0.
|
| 434 |
"step": 35500
|
| 435 |
},
|
| 436 |
{
|
| 437 |
"epoch": 1.23,
|
| 438 |
"learning_rate": 1.1506849315068493e-05,
|
| 439 |
-
"loss": 0.
|
| 440 |
"step": 36000
|
| 441 |
},
|
| 442 |
{
|
| 443 |
"epoch": 1.25,
|
| 444 |
"learning_rate": 1.125e-05,
|
| 445 |
-
"loss": 0.
|
| 446 |
"step": 36500
|
| 447 |
},
|
| 448 |
{
|
| 449 |
"epoch": 1.27,
|
| 450 |
"learning_rate": 1.0993150684931507e-05,
|
| 451 |
-
"loss": 0.
|
| 452 |
"step": 37000
|
| 453 |
},
|
| 454 |
{
|
| 455 |
"epoch": 1.28,
|
| 456 |
"learning_rate": 1.0736301369863013e-05,
|
| 457 |
-
"loss": 0.
|
| 458 |
"step": 37500
|
| 459 |
},
|
| 460 |
{
|
| 461 |
"epoch": 1.3,
|
| 462 |
"learning_rate": 1.0479452054794521e-05,
|
| 463 |
-
"loss": 0.
|
| 464 |
"step": 38000
|
| 465 |
},
|
| 466 |
{
|
| 467 |
"epoch": 1.32,
|
| 468 |
"learning_rate": 1.0222602739726028e-05,
|
| 469 |
-
"loss": 0.
|
| 470 |
"step": 38500
|
| 471 |
},
|
| 472 |
{
|
| 473 |
"epoch": 1.34,
|
| 474 |
"learning_rate": 9.965753424657534e-06,
|
| 475 |
-
"loss": 0.
|
| 476 |
"step": 39000
|
| 477 |
},
|
| 478 |
{
|
| 479 |
"epoch": 1.35,
|
| 480 |
"learning_rate": 9.708904109589042e-06,
|
| 481 |
-
"loss": 0.
|
| 482 |
"step": 39500
|
| 483 |
},
|
| 484 |
{
|
| 485 |
"epoch": 1.37,
|
| 486 |
"learning_rate": 9.452054794520548e-06,
|
| 487 |
-
"loss": 0.
|
| 488 |
"step": 40000
|
| 489 |
},
|
| 490 |
{
|
| 491 |
"epoch": 1.39,
|
| 492 |
"learning_rate": 9.195205479452054e-06,
|
| 493 |
-
"loss": 0.
|
| 494 |
"step": 40500
|
| 495 |
},
|
| 496 |
{
|
| 497 |
"epoch": 1.4,
|
| 498 |
"learning_rate": 8.938356164383562e-06,
|
| 499 |
-
"loss": 0.
|
| 500 |
"step": 41000
|
| 501 |
},
|
| 502 |
{
|
| 503 |
"epoch": 1.42,
|
| 504 |
"learning_rate": 8.681506849315069e-06,
|
| 505 |
-
"loss": 0.
|
| 506 |
"step": 41500
|
| 507 |
},
|
| 508 |
{
|
| 509 |
"epoch": 1.44,
|
| 510 |
"learning_rate": 8.424657534246577e-06,
|
| 511 |
-
"loss": 0.
|
| 512 |
"step": 42000
|
| 513 |
},
|
| 514 |
{
|
| 515 |
"epoch": 1.46,
|
| 516 |
"learning_rate": 8.167808219178081e-06,
|
| 517 |
-
"loss": 0.
|
| 518 |
"step": 42500
|
| 519 |
},
|
| 520 |
{
|
| 521 |
"epoch": 1.47,
|
| 522 |
"learning_rate": 7.91095890410959e-06,
|
| 523 |
-
"loss": 0.
|
| 524 |
"step": 43000
|
| 525 |
},
|
| 526 |
{
|
| 527 |
"epoch": 1.49,
|
| 528 |
"learning_rate": 7.654109589041097e-06,
|
| 529 |
-
"loss": 0.
|
| 530 |
"step": 43500
|
| 531 |
},
|
| 532 |
{
|
| 533 |
"epoch": 1.51,
|
| 534 |
"learning_rate": 7.397260273972603e-06,
|
| 535 |
-
"loss": 0.
|
| 536 |
"step": 44000
|
| 537 |
},
|
| 538 |
{
|
| 539 |
"epoch": 1.52,
|
| 540 |
"learning_rate": 7.14041095890411e-06,
|
| 541 |
-
"loss": 0.
|
| 542 |
"step": 44500
|
| 543 |
},
|
| 544 |
{
|
| 545 |
"epoch": 1.54,
|
| 546 |
"learning_rate": 6.883561643835617e-06,
|
| 547 |
-
"loss": 0.
|
| 548 |
"step": 45000
|
| 549 |
},
|
| 550 |
{
|
| 551 |
"epoch": 1.56,
|
| 552 |
"learning_rate": 6.626712328767123e-06,
|
| 553 |
-
"loss": 0.
|
| 554 |
"step": 45500
|
| 555 |
},
|
| 556 |
{
|
| 557 |
"epoch": 1.58,
|
| 558 |
"learning_rate": 6.3698630136986296e-06,
|
| 559 |
-
"loss": 0.
|
| 560 |
"step": 46000
|
| 561 |
},
|
| 562 |
{
|
| 563 |
"epoch": 1.59,
|
| 564 |
"learning_rate": 6.1130136986301376e-06,
|
| 565 |
-
"loss": 0.
|
| 566 |
"step": 46500
|
| 567 |
},
|
| 568 |
{
|
| 569 |
"epoch": 1.61,
|
| 570 |
"learning_rate": 5.856164383561644e-06,
|
| 571 |
-
"loss": 0.
|
| 572 |
"step": 47000
|
| 573 |
},
|
| 574 |
{
|
| 575 |
"epoch": 1.63,
|
| 576 |
"learning_rate": 5.599315068493151e-06,
|
| 577 |
-
"loss": 0.
|
| 578 |
"step": 47500
|
| 579 |
},
|
| 580 |
{
|
| 581 |
"epoch": 1.64,
|
| 582 |
"learning_rate": 5.342465753424657e-06,
|
| 583 |
-
"loss": 0.
|
| 584 |
"step": 48000
|
| 585 |
},
|
| 586 |
{
|
| 587 |
"epoch": 1.66,
|
| 588 |
"learning_rate": 5.085616438356165e-06,
|
| 589 |
-
"loss": 0.
|
| 590 |
"step": 48500
|
| 591 |
},
|
| 592 |
{
|
| 593 |
"epoch": 1.68,
|
| 594 |
"learning_rate": 4.8287671232876716e-06,
|
| 595 |
-
"loss": 0.
|
| 596 |
"step": 49000
|
| 597 |
},
|
| 598 |
{
|
| 599 |
"epoch": 1.7,
|
| 600 |
"learning_rate": 4.571917808219178e-06,
|
| 601 |
-
"loss": 0.
|
| 602 |
"step": 49500
|
| 603 |
},
|
| 604 |
{
|
| 605 |
"epoch": 1.71,
|
| 606 |
"learning_rate": 4.315068493150685e-06,
|
| 607 |
-
"loss": 0.
|
| 608 |
"step": 50000
|
| 609 |
},
|
| 610 |
{
|
| 611 |
"epoch": 1.73,
|
| 612 |
"learning_rate": 4.058219178082192e-06,
|
| 613 |
-
"loss": 0.
|
| 614 |
"step": 50500
|
| 615 |
},
|
| 616 |
{
|
| 617 |
"epoch": 1.75,
|
| 618 |
"learning_rate": 3.801369863013699e-06,
|
| 619 |
-
"loss": 0.
|
| 620 |
"step": 51000
|
| 621 |
},
|
| 622 |
{
|
| 623 |
"epoch": 1.76,
|
| 624 |
"learning_rate": 3.5445205479452056e-06,
|
| 625 |
-
"loss": 0.
|
| 626 |
"step": 51500
|
| 627 |
},
|
| 628 |
{
|
| 629 |
"epoch": 1.78,
|
| 630 |
"learning_rate": 3.2876712328767123e-06,
|
| 631 |
-
"loss": 0.
|
| 632 |
"step": 52000
|
| 633 |
},
|
| 634 |
{
|
| 635 |
"epoch": 1.8,
|
| 636 |
"learning_rate": 3.0308219178082194e-06,
|
| 637 |
-
"loss": 0.
|
| 638 |
"step": 52500
|
| 639 |
},
|
| 640 |
{
|
| 641 |
"epoch": 1.82,
|
| 642 |
"learning_rate": 2.773972602739726e-06,
|
| 643 |
-
"loss": 0.
|
| 644 |
"step": 53000
|
| 645 |
},
|
| 646 |
{
|
| 647 |
"epoch": 1.83,
|
| 648 |
"learning_rate": 2.517123287671233e-06,
|
| 649 |
-
"loss": 0.
|
| 650 |
"step": 53500
|
| 651 |
},
|
| 652 |
{
|
| 653 |
"epoch": 1.85,
|
| 654 |
"learning_rate": 2.2602739726027396e-06,
|
| 655 |
-
"loss": 0.
|
| 656 |
"step": 54000
|
| 657 |
},
|
| 658 |
{
|
| 659 |
"epoch": 1.87,
|
| 660 |
"learning_rate": 2.0034246575342467e-06,
|
| 661 |
-
"loss": 0.
|
| 662 |
"step": 54500
|
| 663 |
},
|
| 664 |
{
|
| 665 |
"epoch": 1.88,
|
| 666 |
"learning_rate": 1.7465753424657534e-06,
|
| 667 |
-
"loss": 0.
|
| 668 |
"step": 55000
|
| 669 |
},
|
| 670 |
{
|
| 671 |
"epoch": 1.9,
|
| 672 |
"learning_rate": 1.4897260273972603e-06,
|
| 673 |
-
"loss": 0.
|
| 674 |
"step": 55500
|
| 675 |
},
|
| 676 |
{
|
| 677 |
"epoch": 1.92,
|
| 678 |
"learning_rate": 1.232876712328767e-06,
|
| 679 |
-
"loss": 0.
|
| 680 |
"step": 56000
|
| 681 |
},
|
| 682 |
{
|
| 683 |
"epoch": 1.93,
|
| 684 |
"learning_rate": 9.76027397260274e-07,
|
| 685 |
-
"loss": 0.
|
| 686 |
"step": 56500
|
| 687 |
},
|
| 688 |
{
|
| 689 |
"epoch": 1.95,
|
| 690 |
"learning_rate": 7.191780821917808e-07,
|
| 691 |
-
"loss": 0.
|
| 692 |
"step": 57000
|
| 693 |
},
|
| 694 |
{
|
| 695 |
"epoch": 1.97,
|
| 696 |
"learning_rate": 4.6232876712328767e-07,
|
| 697 |
-
"loss": 0.
|
| 698 |
"step": 57500
|
| 699 |
},
|
| 700 |
{
|
| 701 |
"epoch": 1.99,
|
| 702 |
"learning_rate": 2.0547945205479452e-07,
|
| 703 |
-
"loss": 0.
|
| 704 |
"step": 58000
|
| 705 |
},
|
| 706 |
{
|
| 707 |
"epoch": 2.0,
|
| 708 |
"step": 58400,
|
| 709 |
-
"total_flos":
|
| 710 |
-
"train_loss": 0.
|
| 711 |
-
"train_runtime":
|
| 712 |
-
"train_samples_per_second":
|
| 713 |
-
"train_steps_per_second":
|
| 714 |
}
|
| 715 |
],
|
| 716 |
"max_steps": 58400,
|
| 717 |
"num_train_epochs": 2,
|
| 718 |
-
"total_flos":
|
| 719 |
"trial_name": null,
|
| 720 |
"trial_params": null
|
| 721 |
}
|
|
|
|
| 10 |
{
|
| 11 |
"epoch": 0.02,
|
| 12 |
"learning_rate": 2.974315068493151e-05,
|
| 13 |
+
"loss": 2.7124,
|
| 14 |
"step": 500
|
| 15 |
},
|
| 16 |
{
|
| 17 |
"epoch": 0.03,
|
| 18 |
"learning_rate": 2.9486301369863017e-05,
|
| 19 |
+
"loss": 1.247,
|
| 20 |
"step": 1000
|
| 21 |
},
|
| 22 |
{
|
| 23 |
"epoch": 0.05,
|
| 24 |
"learning_rate": 2.922945205479452e-05,
|
| 25 |
+
"loss": 1.0737,
|
| 26 |
"step": 1500
|
| 27 |
},
|
| 28 |
{
|
| 29 |
"epoch": 0.07,
|
| 30 |
"learning_rate": 2.8972602739726026e-05,
|
| 31 |
+
"loss": 0.9364,
|
| 32 |
"step": 2000
|
| 33 |
},
|
| 34 |
{
|
| 35 |
"epoch": 0.09,
|
| 36 |
"learning_rate": 2.8715753424657534e-05,
|
| 37 |
+
"loss": 0.9113,
|
| 38 |
"step": 2500
|
| 39 |
},
|
| 40 |
{
|
| 41 |
"epoch": 0.1,
|
| 42 |
"learning_rate": 2.8458904109589042e-05,
|
| 43 |
+
"loss": 0.8625,
|
| 44 |
"step": 3000
|
| 45 |
},
|
| 46 |
{
|
| 47 |
"epoch": 0.12,
|
| 48 |
"learning_rate": 2.820205479452055e-05,
|
| 49 |
+
"loss": 0.8349,
|
| 50 |
"step": 3500
|
| 51 |
},
|
| 52 |
{
|
| 53 |
"epoch": 0.14,
|
| 54 |
"learning_rate": 2.7945205479452054e-05,
|
| 55 |
+
"loss": 0.7934,
|
| 56 |
"step": 4000
|
| 57 |
},
|
| 58 |
{
|
| 59 |
"epoch": 0.15,
|
| 60 |
"learning_rate": 2.7688356164383562e-05,
|
| 61 |
+
"loss": 0.7953,
|
| 62 |
"step": 4500
|
| 63 |
},
|
| 64 |
{
|
| 65 |
"epoch": 0.17,
|
| 66 |
"learning_rate": 2.743150684931507e-05,
|
| 67 |
+
"loss": 0.7222,
|
| 68 |
"step": 5000
|
| 69 |
},
|
| 70 |
{
|
| 71 |
"epoch": 0.19,
|
| 72 |
"learning_rate": 2.7174657534246575e-05,
|
| 73 |
+
"loss": 0.7703,
|
| 74 |
"step": 5500
|
| 75 |
},
|
| 76 |
{
|
| 77 |
"epoch": 0.21,
|
| 78 |
"learning_rate": 2.6917808219178083e-05,
|
| 79 |
+
"loss": 0.7948,
|
| 80 |
"step": 6000
|
| 81 |
},
|
| 82 |
{
|
| 83 |
"epoch": 0.22,
|
| 84 |
"learning_rate": 2.666095890410959e-05,
|
| 85 |
+
"loss": 0.7789,
|
| 86 |
"step": 6500
|
| 87 |
},
|
| 88 |
{
|
| 89 |
"epoch": 0.24,
|
| 90 |
"learning_rate": 2.6404109589041096e-05,
|
| 91 |
+
"loss": 0.7879,
|
| 92 |
"step": 7000
|
| 93 |
},
|
| 94 |
{
|
| 95 |
"epoch": 0.26,
|
| 96 |
"learning_rate": 2.6147260273972604e-05,
|
| 97 |
+
"loss": 0.7719,
|
| 98 |
"step": 7500
|
| 99 |
},
|
| 100 |
{
|
| 101 |
"epoch": 0.27,
|
| 102 |
"learning_rate": 2.589041095890411e-05,
|
| 103 |
+
"loss": 0.7402,
|
| 104 |
"step": 8000
|
| 105 |
},
|
| 106 |
{
|
| 107 |
"epoch": 0.29,
|
| 108 |
"learning_rate": 2.563356164383562e-05,
|
| 109 |
+
"loss": 0.7068,
|
| 110 |
"step": 8500
|
| 111 |
},
|
| 112 |
{
|
| 113 |
"epoch": 0.31,
|
| 114 |
"learning_rate": 2.5376712328767124e-05,
|
| 115 |
+
"loss": 0.6811,
|
| 116 |
"step": 9000
|
| 117 |
},
|
| 118 |
{
|
| 119 |
"epoch": 0.33,
|
| 120 |
"learning_rate": 2.511986301369863e-05,
|
| 121 |
+
"loss": 0.7144,
|
| 122 |
"step": 9500
|
| 123 |
},
|
| 124 |
{
|
| 125 |
"epoch": 0.34,
|
| 126 |
"learning_rate": 2.4863013698630137e-05,
|
| 127 |
+
"loss": 0.7241,
|
| 128 |
"step": 10000
|
| 129 |
},
|
| 130 |
{
|
| 131 |
"epoch": 0.36,
|
| 132 |
"learning_rate": 2.4606164383561645e-05,
|
| 133 |
+
"loss": 0.739,
|
| 134 |
"step": 10500
|
| 135 |
},
|
| 136 |
{
|
| 137 |
"epoch": 0.38,
|
| 138 |
"learning_rate": 2.4349315068493153e-05,
|
| 139 |
+
"loss": 0.7408,
|
| 140 |
"step": 11000
|
| 141 |
},
|
| 142 |
{
|
| 143 |
"epoch": 0.39,
|
| 144 |
"learning_rate": 2.409246575342466e-05,
|
| 145 |
+
"loss": 0.7175,
|
| 146 |
"step": 11500
|
| 147 |
},
|
| 148 |
{
|
| 149 |
"epoch": 0.41,
|
| 150 |
"learning_rate": 2.3835616438356165e-05,
|
| 151 |
+
"loss": 0.7625,
|
| 152 |
"step": 12000
|
| 153 |
},
|
| 154 |
{
|
| 155 |
"epoch": 0.43,
|
| 156 |
"learning_rate": 2.357876712328767e-05,
|
| 157 |
+
"loss": 0.6987,
|
| 158 |
"step": 12500
|
| 159 |
},
|
| 160 |
{
|
| 161 |
"epoch": 0.45,
|
| 162 |
"learning_rate": 2.3321917808219178e-05,
|
| 163 |
+
"loss": 0.684,
|
| 164 |
"step": 13000
|
| 165 |
},
|
| 166 |
{
|
| 167 |
"epoch": 0.46,
|
| 168 |
"learning_rate": 2.3065068493150686e-05,
|
| 169 |
+
"loss": 0.7049,
|
| 170 |
"step": 13500
|
| 171 |
},
|
| 172 |
{
|
| 173 |
"epoch": 0.48,
|
| 174 |
"learning_rate": 2.2808219178082194e-05,
|
| 175 |
+
"loss": 0.669,
|
| 176 |
"step": 14000
|
| 177 |
},
|
| 178 |
{
|
| 179 |
"epoch": 0.5,
|
| 180 |
"learning_rate": 2.25513698630137e-05,
|
| 181 |
+
"loss": 0.6912,
|
| 182 |
"step": 14500
|
| 183 |
},
|
| 184 |
{
|
| 185 |
"epoch": 0.51,
|
| 186 |
"learning_rate": 2.2294520547945206e-05,
|
| 187 |
+
"loss": 0.6724,
|
| 188 |
"step": 15000
|
| 189 |
},
|
| 190 |
{
|
| 191 |
"epoch": 0.53,
|
| 192 |
"learning_rate": 2.203767123287671e-05,
|
| 193 |
+
"loss": 0.6429,
|
| 194 |
"step": 15500
|
| 195 |
},
|
| 196 |
{
|
| 197 |
"epoch": 0.55,
|
| 198 |
"learning_rate": 2.178082191780822e-05,
|
| 199 |
+
"loss": 0.6582,
|
| 200 |
"step": 16000
|
| 201 |
},
|
| 202 |
{
|
| 203 |
"epoch": 0.57,
|
| 204 |
"learning_rate": 2.1523972602739727e-05,
|
| 205 |
+
"loss": 0.6778,
|
| 206 |
"step": 16500
|
| 207 |
},
|
| 208 |
{
|
| 209 |
"epoch": 0.58,
|
| 210 |
"learning_rate": 2.126712328767123e-05,
|
| 211 |
+
"loss": 0.6743,
|
| 212 |
"step": 17000
|
| 213 |
},
|
| 214 |
{
|
| 215 |
"epoch": 0.6,
|
| 216 |
"learning_rate": 2.101027397260274e-05,
|
| 217 |
+
"loss": 0.6421,
|
| 218 |
"step": 17500
|
| 219 |
},
|
| 220 |
{
|
| 221 |
"epoch": 0.62,
|
| 222 |
"learning_rate": 2.0753424657534248e-05,
|
| 223 |
+
"loss": 0.6484,
|
| 224 |
"step": 18000
|
| 225 |
},
|
| 226 |
{
|
| 227 |
"epoch": 0.63,
|
| 228 |
"learning_rate": 2.0496575342465756e-05,
|
| 229 |
+
"loss": 0.6851,
|
| 230 |
"step": 18500
|
| 231 |
},
|
| 232 |
{
|
| 233 |
"epoch": 0.65,
|
| 234 |
"learning_rate": 2.023972602739726e-05,
|
| 235 |
+
"loss": 0.7124,
|
| 236 |
"step": 19000
|
| 237 |
},
|
| 238 |
{
|
| 239 |
"epoch": 0.67,
|
| 240 |
"learning_rate": 1.9982876712328768e-05,
|
| 241 |
+
"loss": 0.6624,
|
| 242 |
"step": 19500
|
| 243 |
},
|
| 244 |
{
|
| 245 |
"epoch": 0.68,
|
| 246 |
"learning_rate": 1.9726027397260273e-05,
|
| 247 |
+
"loss": 0.6519,
|
| 248 |
"step": 20000
|
| 249 |
},
|
| 250 |
{
|
| 251 |
"epoch": 0.7,
|
| 252 |
"learning_rate": 1.946917808219178e-05,
|
| 253 |
+
"loss": 0.6528,
|
| 254 |
"step": 20500
|
| 255 |
},
|
| 256 |
{
|
| 257 |
"epoch": 0.72,
|
| 258 |
"learning_rate": 1.921232876712329e-05,
|
| 259 |
+
"loss": 0.6881,
|
| 260 |
"step": 21000
|
| 261 |
},
|
| 262 |
{
|
| 263 |
"epoch": 0.74,
|
| 264 |
"learning_rate": 1.8955479452054797e-05,
|
| 265 |
+
"loss": 0.6608,
|
| 266 |
"step": 21500
|
| 267 |
},
|
| 268 |
{
|
| 269 |
"epoch": 0.75,
|
| 270 |
"learning_rate": 1.8698630136986305e-05,
|
| 271 |
+
"loss": 0.6633,
|
| 272 |
"step": 22000
|
| 273 |
},
|
| 274 |
{
|
| 275 |
"epoch": 0.77,
|
| 276 |
"learning_rate": 1.8441780821917806e-05,
|
| 277 |
+
"loss": 0.6357,
|
| 278 |
"step": 22500
|
| 279 |
},
|
| 280 |
{
|
| 281 |
"epoch": 0.79,
|
| 282 |
"learning_rate": 1.8184931506849314e-05,
|
| 283 |
+
"loss": 0.6411,
|
| 284 |
"step": 23000
|
| 285 |
},
|
| 286 |
{
|
| 287 |
"epoch": 0.8,
|
| 288 |
"learning_rate": 1.7928082191780822e-05,
|
| 289 |
+
"loss": 0.6579,
|
| 290 |
"step": 23500
|
| 291 |
},
|
| 292 |
{
|
| 293 |
"epoch": 0.82,
|
| 294 |
"learning_rate": 1.767123287671233e-05,
|
| 295 |
+
"loss": 0.6476,
|
| 296 |
"step": 24000
|
| 297 |
},
|
| 298 |
{
|
| 299 |
"epoch": 0.84,
|
| 300 |
"learning_rate": 1.7414383561643838e-05,
|
| 301 |
+
"loss": 0.633,
|
| 302 |
"step": 24500
|
| 303 |
},
|
| 304 |
{
|
| 305 |
"epoch": 0.86,
|
| 306 |
"learning_rate": 1.7157534246575342e-05,
|
| 307 |
+
"loss": 0.7078,
|
| 308 |
"step": 25000
|
| 309 |
},
|
| 310 |
{
|
| 311 |
"epoch": 0.87,
|
| 312 |
"learning_rate": 1.690068493150685e-05,
|
| 313 |
+
"loss": 0.6676,
|
| 314 |
"step": 25500
|
| 315 |
},
|
| 316 |
{
|
| 317 |
"epoch": 0.89,
|
| 318 |
"learning_rate": 1.6643835616438355e-05,
|
| 319 |
+
"loss": 0.6343,
|
| 320 |
"step": 26000
|
| 321 |
},
|
| 322 |
{
|
| 323 |
"epoch": 0.91,
|
| 324 |
"learning_rate": 1.6386986301369863e-05,
|
| 325 |
+
"loss": 0.6723,
|
| 326 |
"step": 26500
|
| 327 |
},
|
| 328 |
{
|
| 329 |
"epoch": 0.92,
|
| 330 |
"learning_rate": 1.613013698630137e-05,
|
| 331 |
+
"loss": 0.6235,
|
| 332 |
"step": 27000
|
| 333 |
},
|
| 334 |
{
|
| 335 |
"epoch": 0.94,
|
| 336 |
"learning_rate": 1.5873287671232876e-05,
|
| 337 |
+
"loss": 0.6666,
|
| 338 |
"step": 27500
|
| 339 |
},
|
| 340 |
{
|
| 341 |
"epoch": 0.96,
|
| 342 |
"learning_rate": 1.5616438356164384e-05,
|
| 343 |
+
"loss": 0.7011,
|
| 344 |
"step": 28000
|
| 345 |
},
|
| 346 |
{
|
| 347 |
"epoch": 0.98,
|
| 348 |
"learning_rate": 1.535958904109589e-05,
|
| 349 |
+
"loss": 0.6765,
|
| 350 |
"step": 28500
|
| 351 |
},
|
| 352 |
{
|
| 353 |
"epoch": 0.99,
|
| 354 |
"learning_rate": 1.5102739726027398e-05,
|
| 355 |
+
"loss": 0.6493,
|
| 356 |
"step": 29000
|
| 357 |
},
|
| 358 |
{
|
| 359 |
"epoch": 1.01,
|
| 360 |
"learning_rate": 1.4845890410958904e-05,
|
| 361 |
+
"loss": 0.5903,
|
| 362 |
"step": 29500
|
| 363 |
},
|
| 364 |
{
|
| 365 |
"epoch": 1.03,
|
| 366 |
"learning_rate": 1.4589041095890412e-05,
|
| 367 |
+
"loss": 0.6206,
|
| 368 |
"step": 30000
|
| 369 |
},
|
| 370 |
{
|
| 371 |
"epoch": 1.04,
|
| 372 |
"learning_rate": 1.4332191780821918e-05,
|
| 373 |
+
"loss": 0.5903,
|
| 374 |
"step": 30500
|
| 375 |
},
|
| 376 |
{
|
| 377 |
"epoch": 1.06,
|
| 378 |
"learning_rate": 1.4075342465753425e-05,
|
| 379 |
+
"loss": 0.6578,
|
| 380 |
"step": 31000
|
| 381 |
},
|
| 382 |
{
|
| 383 |
"epoch": 1.08,
|
| 384 |
"learning_rate": 1.3818493150684933e-05,
|
| 385 |
+
"loss": 0.6314,
|
| 386 |
"step": 31500
|
| 387 |
},
|
| 388 |
{
|
| 389 |
"epoch": 1.1,
|
| 390 |
"learning_rate": 1.3561643835616437e-05,
|
| 391 |
+
"loss": 0.6242,
|
| 392 |
"step": 32000
|
| 393 |
},
|
| 394 |
{
|
| 395 |
"epoch": 1.11,
|
| 396 |
"learning_rate": 1.3304794520547945e-05,
|
| 397 |
+
"loss": 0.6273,
|
| 398 |
"step": 32500
|
| 399 |
},
|
| 400 |
{
|
| 401 |
"epoch": 1.13,
|
| 402 |
"learning_rate": 1.3047945205479453e-05,
|
| 403 |
+
"loss": 0.6294,
|
| 404 |
"step": 33000
|
| 405 |
},
|
| 406 |
{
|
| 407 |
"epoch": 1.15,
|
| 408 |
"learning_rate": 1.279109589041096e-05,
|
| 409 |
+
"loss": 0.6141,
|
| 410 |
"step": 33500
|
| 411 |
},
|
| 412 |
{
|
| 413 |
"epoch": 1.16,
|
| 414 |
"learning_rate": 1.2534246575342466e-05,
|
| 415 |
+
"loss": 0.5828,
|
| 416 |
"step": 34000
|
| 417 |
},
|
| 418 |
{
|
| 419 |
"epoch": 1.18,
|
| 420 |
"learning_rate": 1.2277397260273974e-05,
|
| 421 |
+
"loss": 0.577,
|
| 422 |
"step": 34500
|
| 423 |
},
|
| 424 |
{
|
| 425 |
"epoch": 1.2,
|
| 426 |
"learning_rate": 1.202054794520548e-05,
|
| 427 |
+
"loss": 0.6136,
|
| 428 |
"step": 35000
|
| 429 |
},
|
| 430 |
{
|
| 431 |
"epoch": 1.22,
|
| 432 |
"learning_rate": 1.1763698630136986e-05,
|
| 433 |
+
"loss": 0.6233,
|
| 434 |
"step": 35500
|
| 435 |
},
|
| 436 |
{
|
| 437 |
"epoch": 1.23,
|
| 438 |
"learning_rate": 1.1506849315068493e-05,
|
| 439 |
+
"loss": 0.5801,
|
| 440 |
"step": 36000
|
| 441 |
},
|
| 442 |
{
|
| 443 |
"epoch": 1.25,
|
| 444 |
"learning_rate": 1.125e-05,
|
| 445 |
+
"loss": 0.6176,
|
| 446 |
"step": 36500
|
| 447 |
},
|
| 448 |
{
|
| 449 |
"epoch": 1.27,
|
| 450 |
"learning_rate": 1.0993150684931507e-05,
|
| 451 |
+
"loss": 0.6007,
|
| 452 |
"step": 37000
|
| 453 |
},
|
| 454 |
{
|
| 455 |
"epoch": 1.28,
|
| 456 |
"learning_rate": 1.0736301369863013e-05,
|
| 457 |
+
"loss": 0.5864,
|
| 458 |
"step": 37500
|
| 459 |
},
|
| 460 |
{
|
| 461 |
"epoch": 1.3,
|
| 462 |
"learning_rate": 1.0479452054794521e-05,
|
| 463 |
+
"loss": 0.651,
|
| 464 |
"step": 38000
|
| 465 |
},
|
| 466 |
{
|
| 467 |
"epoch": 1.32,
|
| 468 |
"learning_rate": 1.0222602739726028e-05,
|
| 469 |
+
"loss": 0.5897,
|
| 470 |
"step": 38500
|
| 471 |
},
|
| 472 |
{
|
| 473 |
"epoch": 1.34,
|
| 474 |
"learning_rate": 9.965753424657534e-06,
|
| 475 |
+
"loss": 0.5907,
|
| 476 |
"step": 39000
|
| 477 |
},
|
| 478 |
{
|
| 479 |
"epoch": 1.35,
|
| 480 |
"learning_rate": 9.708904109589042e-06,
|
| 481 |
+
"loss": 0.631,
|
| 482 |
"step": 39500
|
| 483 |
},
|
| 484 |
{
|
| 485 |
"epoch": 1.37,
|
| 486 |
"learning_rate": 9.452054794520548e-06,
|
| 487 |
+
"loss": 0.6517,
|
| 488 |
"step": 40000
|
| 489 |
},
|
| 490 |
{
|
| 491 |
"epoch": 1.39,
|
| 492 |
"learning_rate": 9.195205479452054e-06,
|
| 493 |
+
"loss": 0.5801,
|
| 494 |
"step": 40500
|
| 495 |
},
|
| 496 |
{
|
| 497 |
"epoch": 1.4,
|
| 498 |
"learning_rate": 8.938356164383562e-06,
|
| 499 |
+
"loss": 0.6225,
|
| 500 |
"step": 41000
|
| 501 |
},
|
| 502 |
{
|
| 503 |
"epoch": 1.42,
|
| 504 |
"learning_rate": 8.681506849315069e-06,
|
| 505 |
+
"loss": 0.6194,
|
| 506 |
"step": 41500
|
| 507 |
},
|
| 508 |
{
|
| 509 |
"epoch": 1.44,
|
| 510 |
"learning_rate": 8.424657534246577e-06,
|
| 511 |
+
"loss": 0.5939,
|
| 512 |
"step": 42000
|
| 513 |
},
|
| 514 |
{
|
| 515 |
"epoch": 1.46,
|
| 516 |
"learning_rate": 8.167808219178081e-06,
|
| 517 |
+
"loss": 0.6029,
|
| 518 |
"step": 42500
|
| 519 |
},
|
| 520 |
{
|
| 521 |
"epoch": 1.47,
|
| 522 |
"learning_rate": 7.91095890410959e-06,
|
| 523 |
+
"loss": 0.5907,
|
| 524 |
"step": 43000
|
| 525 |
},
|
| 526 |
{
|
| 527 |
"epoch": 1.49,
|
| 528 |
"learning_rate": 7.654109589041097e-06,
|
| 529 |
+
"loss": 0.582,
|
| 530 |
"step": 43500
|
| 531 |
},
|
| 532 |
{
|
| 533 |
"epoch": 1.51,
|
| 534 |
"learning_rate": 7.397260273972603e-06,
|
| 535 |
+
"loss": 0.5766,
|
| 536 |
"step": 44000
|
| 537 |
},
|
| 538 |
{
|
| 539 |
"epoch": 1.52,
|
| 540 |
"learning_rate": 7.14041095890411e-06,
|
| 541 |
+
"loss": 0.6133,
|
| 542 |
"step": 44500
|
| 543 |
},
|
| 544 |
{
|
| 545 |
"epoch": 1.54,
|
| 546 |
"learning_rate": 6.883561643835617e-06,
|
| 547 |
+
"loss": 0.5998,
|
| 548 |
"step": 45000
|
| 549 |
},
|
| 550 |
{
|
| 551 |
"epoch": 1.56,
|
| 552 |
"learning_rate": 6.626712328767123e-06,
|
| 553 |
+
"loss": 0.6405,
|
| 554 |
"step": 45500
|
| 555 |
},
|
| 556 |
{
|
| 557 |
"epoch": 1.58,
|
| 558 |
"learning_rate": 6.3698630136986296e-06,
|
| 559 |
+
"loss": 0.5943,
|
| 560 |
"step": 46000
|
| 561 |
},
|
| 562 |
{
|
| 563 |
"epoch": 1.59,
|
| 564 |
"learning_rate": 6.1130136986301376e-06,
|
| 565 |
+
"loss": 0.6153,
|
| 566 |
"step": 46500
|
| 567 |
},
|
| 568 |
{
|
| 569 |
"epoch": 1.61,
|
| 570 |
"learning_rate": 5.856164383561644e-06,
|
| 571 |
+
"loss": 0.5876,
|
| 572 |
"step": 47000
|
| 573 |
},
|
| 574 |
{
|
| 575 |
"epoch": 1.63,
|
| 576 |
"learning_rate": 5.599315068493151e-06,
|
| 577 |
+
"loss": 0.6071,
|
| 578 |
"step": 47500
|
| 579 |
},
|
| 580 |
{
|
| 581 |
"epoch": 1.64,
|
| 582 |
"learning_rate": 5.342465753424657e-06,
|
| 583 |
+
"loss": 0.5965,
|
| 584 |
"step": 48000
|
| 585 |
},
|
| 586 |
{
|
| 587 |
"epoch": 1.66,
|
| 588 |
"learning_rate": 5.085616438356165e-06,
|
| 589 |
+
"loss": 0.586,
|
| 590 |
"step": 48500
|
| 591 |
},
|
| 592 |
{
|
| 593 |
"epoch": 1.68,
|
| 594 |
"learning_rate": 4.8287671232876716e-06,
|
| 595 |
+
"loss": 0.5812,
|
| 596 |
"step": 49000
|
| 597 |
},
|
| 598 |
{
|
| 599 |
"epoch": 1.7,
|
| 600 |
"learning_rate": 4.571917808219178e-06,
|
| 601 |
+
"loss": 0.6407,
|
| 602 |
"step": 49500
|
| 603 |
},
|
| 604 |
{
|
| 605 |
"epoch": 1.71,
|
| 606 |
"learning_rate": 4.315068493150685e-06,
|
| 607 |
+
"loss": 0.6125,
|
| 608 |
"step": 50000
|
| 609 |
},
|
| 610 |
{
|
| 611 |
"epoch": 1.73,
|
| 612 |
"learning_rate": 4.058219178082192e-06,
|
| 613 |
+
"loss": 0.5673,
|
| 614 |
"step": 50500
|
| 615 |
},
|
| 616 |
{
|
| 617 |
"epoch": 1.75,
|
| 618 |
"learning_rate": 3.801369863013699e-06,
|
| 619 |
+
"loss": 0.5724,
|
| 620 |
"step": 51000
|
| 621 |
},
|
| 622 |
{
|
| 623 |
"epoch": 1.76,
|
| 624 |
"learning_rate": 3.5445205479452056e-06,
|
| 625 |
+
"loss": 0.5494,
|
| 626 |
"step": 51500
|
| 627 |
},
|
| 628 |
{
|
| 629 |
"epoch": 1.78,
|
| 630 |
"learning_rate": 3.2876712328767123e-06,
|
| 631 |
+
"loss": 0.6055,
|
| 632 |
"step": 52000
|
| 633 |
},
|
| 634 |
{
|
| 635 |
"epoch": 1.8,
|
| 636 |
"learning_rate": 3.0308219178082194e-06,
|
| 637 |
+
"loss": 0.5904,
|
| 638 |
"step": 52500
|
| 639 |
},
|
| 640 |
{
|
| 641 |
"epoch": 1.82,
|
| 642 |
"learning_rate": 2.773972602739726e-06,
|
| 643 |
+
"loss": 0.6092,
|
| 644 |
"step": 53000
|
| 645 |
},
|
| 646 |
{
|
| 647 |
"epoch": 1.83,
|
| 648 |
"learning_rate": 2.517123287671233e-06,
|
| 649 |
+
"loss": 0.6142,
|
| 650 |
"step": 53500
|
| 651 |
},
|
| 652 |
{
|
| 653 |
"epoch": 1.85,
|
| 654 |
"learning_rate": 2.2602739726027396e-06,
|
| 655 |
+
"loss": 0.6156,
|
| 656 |
"step": 54000
|
| 657 |
},
|
| 658 |
{
|
| 659 |
"epoch": 1.87,
|
| 660 |
"learning_rate": 2.0034246575342467e-06,
|
| 661 |
+
"loss": 0.6119,
|
| 662 |
"step": 54500
|
| 663 |
},
|
| 664 |
{
|
| 665 |
"epoch": 1.88,
|
| 666 |
"learning_rate": 1.7465753424657534e-06,
|
| 667 |
+
"loss": 0.603,
|
| 668 |
"step": 55000
|
| 669 |
},
|
| 670 |
{
|
| 671 |
"epoch": 1.9,
|
| 672 |
"learning_rate": 1.4897260273972603e-06,
|
| 673 |
+
"loss": 0.6102,
|
| 674 |
"step": 55500
|
| 675 |
},
|
| 676 |
{
|
| 677 |
"epoch": 1.92,
|
| 678 |
"learning_rate": 1.232876712328767e-06,
|
| 679 |
+
"loss": 0.6271,
|
| 680 |
"step": 56000
|
| 681 |
},
|
| 682 |
{
|
| 683 |
"epoch": 1.93,
|
| 684 |
"learning_rate": 9.76027397260274e-07,
|
| 685 |
+
"loss": 0.5944,
|
| 686 |
"step": 56500
|
| 687 |
},
|
| 688 |
{
|
| 689 |
"epoch": 1.95,
|
| 690 |
"learning_rate": 7.191780821917808e-07,
|
| 691 |
+
"loss": 0.6215,
|
| 692 |
"step": 57000
|
| 693 |
},
|
| 694 |
{
|
| 695 |
"epoch": 1.97,
|
| 696 |
"learning_rate": 4.6232876712328767e-07,
|
| 697 |
+
"loss": 0.5623,
|
| 698 |
"step": 57500
|
| 699 |
},
|
| 700 |
{
|
| 701 |
"epoch": 1.99,
|
| 702 |
"learning_rate": 2.0547945205479452e-07,
|
| 703 |
+
"loss": 0.568,
|
| 704 |
"step": 58000
|
| 705 |
},
|
| 706 |
{
|
| 707 |
"epoch": 2.0,
|
| 708 |
"step": 58400,
|
| 709 |
+
"total_flos": 7.731654113627136e+16,
|
| 710 |
+
"train_loss": 0.1063714675380759,
|
| 711 |
+
"train_runtime": 3374.8833,
|
| 712 |
+
"train_samples_per_second": 51.912,
|
| 713 |
+
"train_steps_per_second": 17.304
|
| 714 |
}
|
| 715 |
],
|
| 716 |
"max_steps": 58400,
|
| 717 |
"num_train_epochs": 2,
|
| 718 |
+
"total_flos": 7.731654113627136e+16,
|
| 719 |
"trial_name": null,
|
| 720 |
"trial_params": null
|
| 721 |
}
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 2991
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:27427062fea82c900bf392207a9ffd2341e20b264fef7f86768cda41ba705db8
|
| 3 |
size 2991
|