Training in progress, epoch 1
Browse files- last-checkpoint/optimizer.pt +1 -1
- last-checkpoint/pytorch_model.bin +1 -1
- last-checkpoint/rng_state.pth +1 -1
- last-checkpoint/scaler.pt +0 -3
- last-checkpoint/scheduler.pt +1 -1
- last-checkpoint/trainer_state.json +81 -993
- last-checkpoint/training_args.bin +1 -1
- pytorch_model.bin +1 -1
- training_args.bin +1 -1
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4549998883
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:68aa5551b0387599aaa908a00bb0cd2d550d1ed17567d64e070fc46832b8b6d0
|
| 3 |
size 4549998883
|
last-checkpoint/pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 2279605745
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8bccecd638f394225805df88227fbba5c325f109230acde5f20671f48f6d0a30
|
| 3 |
size 2279605745
|
last-checkpoint/rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14503
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:851c351fb9daa3bd758c1ce47434d6740e9ca3af64f0e3d1baed4a9d37634e2f
|
| 3 |
size 14503
|
last-checkpoint/scaler.pt
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:5df0168a751b1257361283b5aa0c288d7c712011edfd861c618859611a2278ec
|
| 3 |
-
size 559
|
|
|
|
|
|
|
|
|
|
|
|
last-checkpoint/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 623
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5dcb519a7fb58c443c13a2e7708175f56534804717be0ab04c3804a0ce65ff81
|
| 3 |
size 623
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,1162 +1,250 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch":
|
| 5 |
-
"global_step":
|
| 6 |
"is_hyper_param_search": false,
|
| 7 |
"is_local_process_zero": true,
|
| 8 |
"is_world_process_zero": true,
|
| 9 |
"log_history": [
|
| 10 |
{
|
| 11 |
"epoch": 0.03,
|
| 12 |
-
"learning_rate": 1.
|
| 13 |
-
"loss": 2.
|
| 14 |
"step": 500
|
| 15 |
},
|
| 16 |
{
|
| 17 |
"epoch": 0.05,
|
| 18 |
-
"learning_rate": 1.
|
| 19 |
-
"loss": 2.
|
| 20 |
"step": 1000
|
| 21 |
},
|
| 22 |
{
|
| 23 |
"epoch": 0.08,
|
| 24 |
-
"learning_rate": 1.
|
| 25 |
-
"loss": 2.
|
| 26 |
"step": 1500
|
| 27 |
},
|
| 28 |
{
|
| 29 |
"epoch": 0.1,
|
| 30 |
-
"learning_rate": 1.
|
| 31 |
-
"loss": 2.
|
| 32 |
"step": 2000
|
| 33 |
},
|
| 34 |
{
|
| 35 |
"epoch": 0.13,
|
| 36 |
-
"learning_rate": 1.
|
| 37 |
-
"loss": 2.
|
| 38 |
"step": 2500
|
| 39 |
},
|
| 40 |
{
|
| 41 |
"epoch": 0.15,
|
| 42 |
-
"learning_rate": 1.
|
| 43 |
-
"loss": 2.
|
| 44 |
"step": 3000
|
| 45 |
},
|
| 46 |
{
|
| 47 |
"epoch": 0.18,
|
| 48 |
-
"learning_rate": 1.
|
| 49 |
-
"loss": 2.
|
| 50 |
"step": 3500
|
| 51 |
},
|
| 52 |
{
|
| 53 |
"epoch": 0.2,
|
| 54 |
-
"learning_rate": 1.
|
| 55 |
-
"loss": 2.
|
| 56 |
"step": 4000
|
| 57 |
},
|
| 58 |
{
|
| 59 |
"epoch": 0.23,
|
| 60 |
-
"learning_rate": 1.
|
| 61 |
-
"loss": 2.
|
| 62 |
"step": 4500
|
| 63 |
},
|
| 64 |
{
|
| 65 |
"epoch": 0.25,
|
| 66 |
-
"learning_rate": 1.
|
| 67 |
-
"loss": 2.
|
| 68 |
"step": 5000
|
| 69 |
},
|
| 70 |
{
|
| 71 |
"epoch": 0.28,
|
| 72 |
-
"learning_rate": 1.
|
| 73 |
-
"loss": 2.
|
| 74 |
"step": 5500
|
| 75 |
},
|
| 76 |
{
|
| 77 |
"epoch": 0.31,
|
| 78 |
-
"learning_rate": 1.
|
| 79 |
-
"loss": 2.
|
| 80 |
"step": 6000
|
| 81 |
},
|
| 82 |
{
|
| 83 |
"epoch": 0.33,
|
| 84 |
-
"learning_rate": 1.
|
| 85 |
-
"loss": 2.
|
| 86 |
"step": 6500
|
| 87 |
},
|
| 88 |
{
|
| 89 |
"epoch": 0.36,
|
| 90 |
-
"learning_rate": 1.
|
| 91 |
-
"loss": 2.
|
| 92 |
"step": 7000
|
| 93 |
},
|
| 94 |
{
|
| 95 |
"epoch": 0.38,
|
| 96 |
-
"learning_rate": 1.
|
| 97 |
-
"loss": 2.
|
| 98 |
"step": 7500
|
| 99 |
},
|
| 100 |
{
|
| 101 |
"epoch": 0.41,
|
| 102 |
-
"learning_rate": 1.
|
| 103 |
-
"loss": 2.
|
| 104 |
"step": 8000
|
| 105 |
},
|
| 106 |
{
|
| 107 |
"epoch": 0.43,
|
| 108 |
-
"learning_rate": 1.
|
| 109 |
-
"loss": 2.
|
| 110 |
"step": 8500
|
| 111 |
},
|
| 112 |
{
|
| 113 |
"epoch": 0.46,
|
| 114 |
-
"learning_rate": 1.
|
| 115 |
-
"loss": 2.
|
| 116 |
"step": 9000
|
| 117 |
},
|
| 118 |
{
|
| 119 |
"epoch": 0.48,
|
| 120 |
-
"learning_rate": 1.
|
| 121 |
-
"loss": 2.
|
| 122 |
"step": 9500
|
| 123 |
},
|
| 124 |
{
|
| 125 |
"epoch": 0.51,
|
| 126 |
-
"learning_rate": 1.
|
| 127 |
-
"loss": 2.
|
| 128 |
"step": 10000
|
| 129 |
},
|
| 130 |
{
|
| 131 |
"epoch": 0.53,
|
| 132 |
-
"learning_rate": 1.
|
| 133 |
-
"loss": 2.
|
| 134 |
"step": 10500
|
| 135 |
},
|
| 136 |
{
|
| 137 |
"epoch": 0.56,
|
| 138 |
-
"learning_rate": 1.
|
| 139 |
-
"loss": 2.
|
| 140 |
"step": 11000
|
| 141 |
},
|
| 142 |
{
|
| 143 |
"epoch": 0.59,
|
| 144 |
-
"learning_rate": 1.
|
| 145 |
-
"loss": 2.
|
| 146 |
"step": 11500
|
| 147 |
},
|
| 148 |
{
|
| 149 |
"epoch": 0.61,
|
| 150 |
-
"learning_rate": 1.
|
| 151 |
-
"loss": 2.
|
| 152 |
"step": 12000
|
| 153 |
},
|
| 154 |
{
|
| 155 |
"epoch": 0.64,
|
| 156 |
-
"learning_rate": 1.
|
| 157 |
-
"loss": 2.
|
| 158 |
"step": 12500
|
| 159 |
},
|
| 160 |
{
|
| 161 |
"epoch": 0.66,
|
| 162 |
-
"learning_rate": 1.
|
| 163 |
-
"loss": 2.
|
| 164 |
"step": 13000
|
| 165 |
},
|
| 166 |
{
|
| 167 |
"epoch": 0.69,
|
| 168 |
-
"learning_rate": 1.
|
| 169 |
-
"loss": 2.
|
| 170 |
"step": 13500
|
| 171 |
},
|
| 172 |
{
|
| 173 |
"epoch": 0.71,
|
| 174 |
-
"learning_rate": 1.
|
| 175 |
-
"loss": 2.
|
| 176 |
"step": 14000
|
| 177 |
},
|
| 178 |
{
|
| 179 |
"epoch": 0.74,
|
| 180 |
-
"learning_rate": 1.
|
| 181 |
-
"loss": 2.
|
| 182 |
"step": 14500
|
| 183 |
},
|
| 184 |
{
|
| 185 |
"epoch": 0.76,
|
| 186 |
-
"learning_rate": 1.
|
| 187 |
-
"loss": 2.
|
| 188 |
"step": 15000
|
| 189 |
},
|
| 190 |
{
|
| 191 |
"epoch": 0.79,
|
| 192 |
-
"learning_rate": 1.
|
| 193 |
-
"loss": 2.
|
| 194 |
"step": 15500
|
| 195 |
},
|
| 196 |
{
|
| 197 |
"epoch": 0.81,
|
| 198 |
-
"learning_rate": 1.
|
| 199 |
-
"loss": 2.
|
| 200 |
"step": 16000
|
| 201 |
},
|
| 202 |
{
|
| 203 |
"epoch": 0.84,
|
| 204 |
-
"learning_rate": 1.
|
| 205 |
-
"loss": 2.
|
| 206 |
"step": 16500
|
| 207 |
},
|
| 208 |
{
|
| 209 |
"epoch": 0.87,
|
| 210 |
-
"learning_rate": 1.
|
| 211 |
-
"loss": 2.
|
| 212 |
"step": 17000
|
| 213 |
},
|
| 214 |
{
|
| 215 |
"epoch": 0.89,
|
| 216 |
-
"learning_rate": 1.
|
| 217 |
-
"loss": 2.
|
| 218 |
"step": 17500
|
| 219 |
},
|
| 220 |
{
|
| 221 |
"epoch": 0.92,
|
| 222 |
-
"learning_rate": 1.
|
| 223 |
-
"loss": 2.
|
| 224 |
"step": 18000
|
| 225 |
},
|
| 226 |
{
|
| 227 |
"epoch": 0.94,
|
| 228 |
-
"learning_rate": 1.
|
| 229 |
-
"loss": 2.
|
| 230 |
"step": 18500
|
| 231 |
},
|
| 232 |
{
|
| 233 |
"epoch": 0.97,
|
| 234 |
-
"learning_rate": 1.
|
| 235 |
-
"loss": 2.
|
| 236 |
"step": 19000
|
| 237 |
},
|
| 238 |
{
|
| 239 |
"epoch": 0.99,
|
| 240 |
-
"learning_rate": 1.
|
| 241 |
-
"loss": 2.
|
| 242 |
"step": 19500
|
| 243 |
-
},
|
| 244 |
-
{
|
| 245 |
-
"epoch": 1.02,
|
| 246 |
-
"learning_rate": 1.8984314524343046e-05,
|
| 247 |
-
"loss": 2.4043,
|
| 248 |
-
"step": 20000
|
| 249 |
-
},
|
| 250 |
-
{
|
| 251 |
-
"epoch": 1.04,
|
| 252 |
-
"learning_rate": 1.8958902016704016e-05,
|
| 253 |
-
"loss": 2.3932,
|
| 254 |
-
"step": 20500
|
| 255 |
-
},
|
| 256 |
-
{
|
| 257 |
-
"epoch": 1.07,
|
| 258 |
-
"learning_rate": 1.8933540435934e-05,
|
| 259 |
-
"loss": 2.3954,
|
| 260 |
-
"step": 21000
|
| 261 |
-
},
|
| 262 |
-
{
|
| 263 |
-
"epoch": 1.09,
|
| 264 |
-
"learning_rate": 1.890812792829497e-05,
|
| 265 |
-
"loss": 2.4121,
|
| 266 |
-
"step": 21500
|
| 267 |
-
},
|
| 268 |
-
{
|
| 269 |
-
"epoch": 1.12,
|
| 270 |
-
"learning_rate": 1.8882664493786923e-05,
|
| 271 |
-
"loss": 2.3912,
|
| 272 |
-
"step": 22000
|
| 273 |
-
},
|
| 274 |
-
{
|
| 275 |
-
"epoch": 1.15,
|
| 276 |
-
"learning_rate": 1.8857302913016907e-05,
|
| 277 |
-
"loss": 2.4108,
|
| 278 |
-
"step": 22500
|
| 279 |
-
},
|
| 280 |
-
{
|
| 281 |
-
"epoch": 1.17,
|
| 282 |
-
"learning_rate": 1.8831839478508863e-05,
|
| 283 |
-
"loss": 2.4006,
|
| 284 |
-
"step": 23000
|
| 285 |
-
},
|
| 286 |
-
{
|
| 287 |
-
"epoch": 1.2,
|
| 288 |
-
"learning_rate": 1.8806426970869833e-05,
|
| 289 |
-
"loss": 2.4146,
|
| 290 |
-
"step": 23500
|
| 291 |
-
},
|
| 292 |
-
{
|
| 293 |
-
"epoch": 1.22,
|
| 294 |
-
"learning_rate": 1.8781065390099817e-05,
|
| 295 |
-
"loss": 2.4157,
|
| 296 |
-
"step": 24000
|
| 297 |
-
},
|
| 298 |
-
{
|
| 299 |
-
"epoch": 1.25,
|
| 300 |
-
"learning_rate": 1.8755652882460787e-05,
|
| 301 |
-
"loss": 2.3924,
|
| 302 |
-
"step": 24500
|
| 303 |
-
},
|
| 304 |
-
{
|
| 305 |
-
"epoch": 1.27,
|
| 306 |
-
"learning_rate": 1.8730240374821757e-05,
|
| 307 |
-
"loss": 2.4002,
|
| 308 |
-
"step": 25000
|
| 309 |
-
},
|
| 310 |
-
{
|
| 311 |
-
"epoch": 1.3,
|
| 312 |
-
"learning_rate": 1.8704776940313713e-05,
|
| 313 |
-
"loss": 2.411,
|
| 314 |
-
"step": 25500
|
| 315 |
-
},
|
| 316 |
-
{
|
| 317 |
-
"epoch": 1.32,
|
| 318 |
-
"learning_rate": 1.867936443267468e-05,
|
| 319 |
-
"loss": 2.3815,
|
| 320 |
-
"step": 26000
|
| 321 |
-
},
|
| 322 |
-
{
|
| 323 |
-
"epoch": 1.35,
|
| 324 |
-
"learning_rate": 1.8653900998166635e-05,
|
| 325 |
-
"loss": 2.3954,
|
| 326 |
-
"step": 26500
|
| 327 |
-
},
|
| 328 |
-
{
|
| 329 |
-
"epoch": 1.38,
|
| 330 |
-
"learning_rate": 1.8628590344265634e-05,
|
| 331 |
-
"loss": 2.3856,
|
| 332 |
-
"step": 27000
|
| 333 |
-
},
|
| 334 |
-
{
|
| 335 |
-
"epoch": 1.4,
|
| 336 |
-
"learning_rate": 1.8603177836626604e-05,
|
| 337 |
-
"loss": 2.4047,
|
| 338 |
-
"step": 27500
|
| 339 |
-
},
|
| 340 |
-
{
|
| 341 |
-
"epoch": 1.43,
|
| 342 |
-
"learning_rate": 1.8577816255856592e-05,
|
| 343 |
-
"loss": 2.3984,
|
| 344 |
-
"step": 28000
|
| 345 |
-
},
|
| 346 |
-
{
|
| 347 |
-
"epoch": 1.45,
|
| 348 |
-
"learning_rate": 1.8552454675086577e-05,
|
| 349 |
-
"loss": 2.3936,
|
| 350 |
-
"step": 28500
|
| 351 |
-
},
|
| 352 |
-
{
|
| 353 |
-
"epoch": 1.48,
|
| 354 |
-
"learning_rate": 1.8527042167447547e-05,
|
| 355 |
-
"loss": 2.4003,
|
| 356 |
-
"step": 29000
|
| 357 |
-
},
|
| 358 |
-
{
|
| 359 |
-
"epoch": 1.5,
|
| 360 |
-
"learning_rate": 1.8501782440415564e-05,
|
| 361 |
-
"loss": 2.3949,
|
| 362 |
-
"step": 29500
|
| 363 |
-
},
|
| 364 |
-
{
|
| 365 |
-
"epoch": 1.53,
|
| 366 |
-
"learning_rate": 1.8476369932776534e-05,
|
| 367 |
-
"loss": 2.4058,
|
| 368 |
-
"step": 30000
|
| 369 |
-
},
|
| 370 |
-
{
|
| 371 |
-
"epoch": 1.55,
|
| 372 |
-
"learning_rate": 1.8450957425137504e-05,
|
| 373 |
-
"loss": 2.3996,
|
| 374 |
-
"step": 30500
|
| 375 |
-
},
|
| 376 |
-
{
|
| 377 |
-
"epoch": 1.58,
|
| 378 |
-
"learning_rate": 1.8425646771236507e-05,
|
| 379 |
-
"loss": 2.4084,
|
| 380 |
-
"step": 31000
|
| 381 |
-
},
|
| 382 |
-
{
|
| 383 |
-
"epoch": 1.6,
|
| 384 |
-
"learning_rate": 1.840018333672846e-05,
|
| 385 |
-
"loss": 2.3996,
|
| 386 |
-
"step": 31500
|
| 387 |
-
},
|
| 388 |
-
{
|
| 389 |
-
"epoch": 1.63,
|
| 390 |
-
"learning_rate": 1.837477082908943e-05,
|
| 391 |
-
"loss": 2.3979,
|
| 392 |
-
"step": 32000
|
| 393 |
-
},
|
| 394 |
-
{
|
| 395 |
-
"epoch": 1.66,
|
| 396 |
-
"learning_rate": 1.834930739458138e-05,
|
| 397 |
-
"loss": 2.4039,
|
| 398 |
-
"step": 32500
|
| 399 |
-
},
|
| 400 |
-
{
|
| 401 |
-
"epoch": 1.68,
|
| 402 |
-
"learning_rate": 1.832394581381137e-05,
|
| 403 |
-
"loss": 2.3954,
|
| 404 |
-
"step": 33000
|
| 405 |
-
},
|
| 406 |
-
{
|
| 407 |
-
"epoch": 1.71,
|
| 408 |
-
"learning_rate": 1.8298635159910368e-05,
|
| 409 |
-
"loss": 2.387,
|
| 410 |
-
"step": 33500
|
| 411 |
-
},
|
| 412 |
-
{
|
| 413 |
-
"epoch": 1.73,
|
| 414 |
-
"learning_rate": 1.827322265227134e-05,
|
| 415 |
-
"loss": 2.403,
|
| 416 |
-
"step": 34000
|
| 417 |
-
},
|
| 418 |
-
{
|
| 419 |
-
"epoch": 1.76,
|
| 420 |
-
"learning_rate": 1.8247759217763293e-05,
|
| 421 |
-
"loss": 2.4083,
|
| 422 |
-
"step": 34500
|
| 423 |
-
},
|
| 424 |
-
{
|
| 425 |
-
"epoch": 1.78,
|
| 426 |
-
"learning_rate": 1.8222550417600325e-05,
|
| 427 |
-
"loss": 2.3993,
|
| 428 |
-
"step": 35000
|
| 429 |
-
},
|
| 430 |
-
{
|
| 431 |
-
"epoch": 1.81,
|
| 432 |
-
"learning_rate": 1.819708698309228e-05,
|
| 433 |
-
"loss": 2.3954,
|
| 434 |
-
"step": 35500
|
| 435 |
-
},
|
| 436 |
-
{
|
| 437 |
-
"epoch": 1.83,
|
| 438 |
-
"learning_rate": 1.8171725402322265e-05,
|
| 439 |
-
"loss": 2.4023,
|
| 440 |
-
"step": 36000
|
| 441 |
-
},
|
| 442 |
-
{
|
| 443 |
-
"epoch": 1.86,
|
| 444 |
-
"learning_rate": 1.814626196781422e-05,
|
| 445 |
-
"loss": 2.3961,
|
| 446 |
-
"step": 36500
|
| 447 |
-
},
|
| 448 |
-
{
|
| 449 |
-
"epoch": 1.88,
|
| 450 |
-
"learning_rate": 1.8120798533306173e-05,
|
| 451 |
-
"loss": 2.3885,
|
| 452 |
-
"step": 37000
|
| 453 |
-
},
|
| 454 |
-
{
|
| 455 |
-
"epoch": 1.91,
|
| 456 |
-
"learning_rate": 1.809543695253616e-05,
|
| 457 |
-
"loss": 2.4144,
|
| 458 |
-
"step": 37500
|
| 459 |
-
},
|
| 460 |
-
{
|
| 461 |
-
"epoch": 1.94,
|
| 462 |
-
"learning_rate": 1.8070126298635163e-05,
|
| 463 |
-
"loss": 2.4142,
|
| 464 |
-
"step": 38000
|
| 465 |
-
},
|
| 466 |
-
{
|
| 467 |
-
"epoch": 1.96,
|
| 468 |
-
"learning_rate": 1.804471379099613e-05,
|
| 469 |
-
"loss": 2.4015,
|
| 470 |
-
"step": 38500
|
| 471 |
-
},
|
| 472 |
-
{
|
| 473 |
-
"epoch": 1.99,
|
| 474 |
-
"learning_rate": 1.8019250356488085e-05,
|
| 475 |
-
"loss": 2.4066,
|
| 476 |
-
"step": 39000
|
| 477 |
-
},
|
| 478 |
-
{
|
| 479 |
-
"epoch": 2.01,
|
| 480 |
-
"learning_rate": 1.7993837848849055e-05,
|
| 481 |
-
"loss": 2.4089,
|
| 482 |
-
"step": 39500
|
| 483 |
-
},
|
| 484 |
-
{
|
| 485 |
-
"epoch": 2.04,
|
| 486 |
-
"learning_rate": 1.7968374414341007e-05,
|
| 487 |
-
"loss": 2.4034,
|
| 488 |
-
"step": 40000
|
| 489 |
-
},
|
| 490 |
-
{
|
| 491 |
-
"epoch": 2.06,
|
| 492 |
-
"learning_rate": 1.7942961906701977e-05,
|
| 493 |
-
"loss": 2.3973,
|
| 494 |
-
"step": 40500
|
| 495 |
-
},
|
| 496 |
-
{
|
| 497 |
-
"epoch": 2.09,
|
| 498 |
-
"learning_rate": 1.7917549399062947e-05,
|
| 499 |
-
"loss": 2.3887,
|
| 500 |
-
"step": 41000
|
| 501 |
-
},
|
| 502 |
-
{
|
| 503 |
-
"epoch": 2.11,
|
| 504 |
-
"learning_rate": 1.7892085964554903e-05,
|
| 505 |
-
"loss": 2.4105,
|
| 506 |
-
"step": 41500
|
| 507 |
-
},
|
| 508 |
-
{
|
| 509 |
-
"epoch": 2.14,
|
| 510 |
-
"learning_rate": 1.7866724383784887e-05,
|
| 511 |
-
"loss": 2.4074,
|
| 512 |
-
"step": 42000
|
| 513 |
-
},
|
| 514 |
-
{
|
| 515 |
-
"epoch": 2.16,
|
| 516 |
-
"learning_rate": 1.7841311876145857e-05,
|
| 517 |
-
"loss": 2.3937,
|
| 518 |
-
"step": 42500
|
| 519 |
-
},
|
| 520 |
-
{
|
| 521 |
-
"epoch": 2.19,
|
| 522 |
-
"learning_rate": 1.7815899368506827e-05,
|
| 523 |
-
"loss": 2.3901,
|
| 524 |
-
"step": 43000
|
| 525 |
-
},
|
| 526 |
-
{
|
| 527 |
-
"epoch": 2.22,
|
| 528 |
-
"learning_rate": 1.7790486860867794e-05,
|
| 529 |
-
"loss": 2.4128,
|
| 530 |
-
"step": 43500
|
| 531 |
-
},
|
| 532 |
-
{
|
| 533 |
-
"epoch": 2.24,
|
| 534 |
-
"learning_rate": 1.776502342635975e-05,
|
| 535 |
-
"loss": 2.4103,
|
| 536 |
-
"step": 44000
|
| 537 |
-
},
|
| 538 |
-
{
|
| 539 |
-
"epoch": 2.27,
|
| 540 |
-
"learning_rate": 1.77395599918517e-05,
|
| 541 |
-
"loss": 2.3942,
|
| 542 |
-
"step": 44500
|
| 543 |
-
},
|
| 544 |
-
{
|
| 545 |
-
"epoch": 2.29,
|
| 546 |
-
"learning_rate": 1.771430026481972e-05,
|
| 547 |
-
"loss": 2.4008,
|
| 548 |
-
"step": 45000
|
| 549 |
-
},
|
| 550 |
-
{
|
| 551 |
-
"epoch": 2.32,
|
| 552 |
-
"learning_rate": 1.7688836830311674e-05,
|
| 553 |
-
"loss": 2.3949,
|
| 554 |
-
"step": 45500
|
| 555 |
-
},
|
| 556 |
-
{
|
| 557 |
-
"epoch": 2.34,
|
| 558 |
-
"learning_rate": 1.7663373395803626e-05,
|
| 559 |
-
"loss": 2.3863,
|
| 560 |
-
"step": 46000
|
| 561 |
-
},
|
| 562 |
-
{
|
| 563 |
-
"epoch": 2.37,
|
| 564 |
-
"learning_rate": 1.763790996129558e-05,
|
| 565 |
-
"loss": 2.3852,
|
| 566 |
-
"step": 46500
|
| 567 |
-
},
|
| 568 |
-
{
|
| 569 |
-
"epoch": 2.39,
|
| 570 |
-
"learning_rate": 1.7612446526787534e-05,
|
| 571 |
-
"loss": 2.3973,
|
| 572 |
-
"step": 47000
|
| 573 |
-
},
|
| 574 |
-
{
|
| 575 |
-
"epoch": 2.42,
|
| 576 |
-
"learning_rate": 1.758708494601752e-05,
|
| 577 |
-
"loss": 2.4162,
|
| 578 |
-
"step": 47500
|
| 579 |
-
},
|
| 580 |
-
{
|
| 581 |
-
"epoch": 2.44,
|
| 582 |
-
"learning_rate": 1.756167243837849e-05,
|
| 583 |
-
"loss": 2.4018,
|
| 584 |
-
"step": 48000
|
| 585 |
-
},
|
| 586 |
-
{
|
| 587 |
-
"epoch": 2.47,
|
| 588 |
-
"learning_rate": 1.753625993073946e-05,
|
| 589 |
-
"loss": 2.4036,
|
| 590 |
-
"step": 48500
|
| 591 |
-
},
|
| 592 |
-
{
|
| 593 |
-
"epoch": 2.5,
|
| 594 |
-
"learning_rate": 1.751084742310043e-05,
|
| 595 |
-
"loss": 2.384,
|
| 596 |
-
"step": 49000
|
| 597 |
-
},
|
| 598 |
-
{
|
| 599 |
-
"epoch": 2.52,
|
| 600 |
-
"learning_rate": 1.7485383988592384e-05,
|
| 601 |
-
"loss": 2.381,
|
| 602 |
-
"step": 49500
|
| 603 |
-
},
|
| 604 |
-
{
|
| 605 |
-
"epoch": 2.55,
|
| 606 |
-
"learning_rate": 1.7459920554084336e-05,
|
| 607 |
-
"loss": 2.4114,
|
| 608 |
-
"step": 50000
|
| 609 |
-
},
|
| 610 |
-
{
|
| 611 |
-
"epoch": 2.57,
|
| 612 |
-
"learning_rate": 1.743455897331432e-05,
|
| 613 |
-
"loss": 2.4005,
|
| 614 |
-
"step": 50500
|
| 615 |
-
},
|
| 616 |
-
{
|
| 617 |
-
"epoch": 2.6,
|
| 618 |
-
"learning_rate": 1.7409095538806276e-05,
|
| 619 |
-
"loss": 2.4118,
|
| 620 |
-
"step": 51000
|
| 621 |
-
},
|
| 622 |
-
{
|
| 623 |
-
"epoch": 2.62,
|
| 624 |
-
"learning_rate": 1.7383632104298228e-05,
|
| 625 |
-
"loss": 2.3912,
|
| 626 |
-
"step": 51500
|
| 627 |
-
},
|
| 628 |
-
{
|
| 629 |
-
"epoch": 2.65,
|
| 630 |
-
"learning_rate": 1.7358219596659198e-05,
|
| 631 |
-
"loss": 2.4011,
|
| 632 |
-
"step": 52000
|
| 633 |
-
},
|
| 634 |
-
{
|
| 635 |
-
"epoch": 2.67,
|
| 636 |
-
"learning_rate": 1.7332807089020168e-05,
|
| 637 |
-
"loss": 2.3933,
|
| 638 |
-
"step": 52500
|
| 639 |
-
},
|
| 640 |
-
{
|
| 641 |
-
"epoch": 2.7,
|
| 642 |
-
"learning_rate": 1.730734365451212e-05,
|
| 643 |
-
"loss": 2.3921,
|
| 644 |
-
"step": 53000
|
| 645 |
-
},
|
| 646 |
-
{
|
| 647 |
-
"epoch": 2.72,
|
| 648 |
-
"learning_rate": 1.7281880220004075e-05,
|
| 649 |
-
"loss": 2.3897,
|
| 650 |
-
"step": 53500
|
| 651 |
-
},
|
| 652 |
-
{
|
| 653 |
-
"epoch": 2.75,
|
| 654 |
-
"learning_rate": 1.7256416785496027e-05,
|
| 655 |
-
"loss": 2.3906,
|
| 656 |
-
"step": 54000
|
| 657 |
-
},
|
| 658 |
-
{
|
| 659 |
-
"epoch": 2.78,
|
| 660 |
-
"learning_rate": 1.7231004277857e-05,
|
| 661 |
-
"loss": 2.3775,
|
| 662 |
-
"step": 54500
|
| 663 |
-
},
|
| 664 |
-
{
|
| 665 |
-
"epoch": 2.8,
|
| 666 |
-
"learning_rate": 1.7205540843348953e-05,
|
| 667 |
-
"loss": 2.4004,
|
| 668 |
-
"step": 55000
|
| 669 |
-
},
|
| 670 |
-
{
|
| 671 |
-
"epoch": 2.83,
|
| 672 |
-
"learning_rate": 1.718028111631697e-05,
|
| 673 |
-
"loss": 2.3924,
|
| 674 |
-
"step": 55500
|
| 675 |
-
},
|
| 676 |
-
{
|
| 677 |
-
"epoch": 2.85,
|
| 678 |
-
"learning_rate": 1.715486860867794e-05,
|
| 679 |
-
"loss": 2.4041,
|
| 680 |
-
"step": 56000
|
| 681 |
-
},
|
| 682 |
-
{
|
| 683 |
-
"epoch": 2.88,
|
| 684 |
-
"learning_rate": 1.7129405174169892e-05,
|
| 685 |
-
"loss": 2.3978,
|
| 686 |
-
"step": 56500
|
| 687 |
-
},
|
| 688 |
-
{
|
| 689 |
-
"epoch": 2.9,
|
| 690 |
-
"learning_rate": 1.7103992666530862e-05,
|
| 691 |
-
"loss": 2.3919,
|
| 692 |
-
"step": 57000
|
| 693 |
-
},
|
| 694 |
-
{
|
| 695 |
-
"epoch": 2.93,
|
| 696 |
-
"learning_rate": 1.7078529232022814e-05,
|
| 697 |
-
"loss": 2.4113,
|
| 698 |
-
"step": 57500
|
| 699 |
-
},
|
| 700 |
-
{
|
| 701 |
-
"epoch": 2.95,
|
| 702 |
-
"learning_rate": 1.705306579751477e-05,
|
| 703 |
-
"loss": 2.3998,
|
| 704 |
-
"step": 58000
|
| 705 |
-
},
|
| 706 |
-
{
|
| 707 |
-
"epoch": 2.98,
|
| 708 |
-
"learning_rate": 1.702765328987574e-05,
|
| 709 |
-
"loss": 2.3919,
|
| 710 |
-
"step": 58500
|
| 711 |
-
},
|
| 712 |
-
{
|
| 713 |
-
"epoch": 3.0,
|
| 714 |
-
"learning_rate": 1.700224078223671e-05,
|
| 715 |
-
"loss": 2.383,
|
| 716 |
-
"step": 59000
|
| 717 |
-
},
|
| 718 |
-
{
|
| 719 |
-
"epoch": 3.03,
|
| 720 |
-
"learning_rate": 1.6976879201466698e-05,
|
| 721 |
-
"loss": 2.3898,
|
| 722 |
-
"step": 59500
|
| 723 |
-
},
|
| 724 |
-
{
|
| 725 |
-
"epoch": 3.06,
|
| 726 |
-
"learning_rate": 1.695141576695865e-05,
|
| 727 |
-
"loss": 2.4015,
|
| 728 |
-
"step": 60000
|
| 729 |
-
},
|
| 730 |
-
{
|
| 731 |
-
"epoch": 3.08,
|
| 732 |
-
"learning_rate": 1.6926054186188634e-05,
|
| 733 |
-
"loss": 2.4021,
|
| 734 |
-
"step": 60500
|
| 735 |
-
},
|
| 736 |
-
{
|
| 737 |
-
"epoch": 3.11,
|
| 738 |
-
"learning_rate": 1.6900641678549604e-05,
|
| 739 |
-
"loss": 2.3994,
|
| 740 |
-
"step": 61000
|
| 741 |
-
},
|
| 742 |
-
{
|
| 743 |
-
"epoch": 3.13,
|
| 744 |
-
"learning_rate": 1.6875178244041556e-05,
|
| 745 |
-
"loss": 2.4086,
|
| 746 |
-
"step": 61500
|
| 747 |
-
},
|
| 748 |
-
{
|
| 749 |
-
"epoch": 3.16,
|
| 750 |
-
"learning_rate": 1.6849714809533512e-05,
|
| 751 |
-
"loss": 2.4037,
|
| 752 |
-
"step": 62000
|
| 753 |
-
},
|
| 754 |
-
{
|
| 755 |
-
"epoch": 3.18,
|
| 756 |
-
"learning_rate": 1.6824251375025464e-05,
|
| 757 |
-
"loss": 2.3923,
|
| 758 |
-
"step": 62500
|
| 759 |
-
},
|
| 760 |
-
{
|
| 761 |
-
"epoch": 3.21,
|
| 762 |
-
"learning_rate": 1.679878794051742e-05,
|
| 763 |
-
"loss": 2.3923,
|
| 764 |
-
"step": 63000
|
| 765 |
-
},
|
| 766 |
-
{
|
| 767 |
-
"epoch": 3.23,
|
| 768 |
-
"learning_rate": 1.677332450600937e-05,
|
| 769 |
-
"loss": 2.4055,
|
| 770 |
-
"step": 63500
|
| 771 |
-
},
|
| 772 |
-
{
|
| 773 |
-
"epoch": 3.26,
|
| 774 |
-
"learning_rate": 1.674791199837034e-05,
|
| 775 |
-
"loss": 2.3975,
|
| 776 |
-
"step": 64000
|
| 777 |
-
},
|
| 778 |
-
{
|
| 779 |
-
"epoch": 3.28,
|
| 780 |
-
"learning_rate": 1.6722550417600326e-05,
|
| 781 |
-
"loss": 2.406,
|
| 782 |
-
"step": 64500
|
| 783 |
-
},
|
| 784 |
-
{
|
| 785 |
-
"epoch": 3.31,
|
| 786 |
-
"learning_rate": 1.6697137909961296e-05,
|
| 787 |
-
"loss": 2.3995,
|
| 788 |
-
"step": 65000
|
| 789 |
-
},
|
| 790 |
-
{
|
| 791 |
-
"epoch": 3.34,
|
| 792 |
-
"learning_rate": 1.6672896720309635e-05,
|
| 793 |
-
"loss": 2.3872,
|
| 794 |
-
"step": 65500
|
| 795 |
-
},
|
| 796 |
-
{
|
| 797 |
-
"epoch": 3.36,
|
| 798 |
-
"learning_rate": 1.664743328580159e-05,
|
| 799 |
-
"loss": 0.0,
|
| 800 |
-
"step": 66000
|
| 801 |
-
},
|
| 802 |
-
{
|
| 803 |
-
"epoch": 3.39,
|
| 804 |
-
"learning_rate": 1.6621969851293543e-05,
|
| 805 |
-
"loss": 0.0,
|
| 806 |
-
"step": 66500
|
| 807 |
-
},
|
| 808 |
-
{
|
| 809 |
-
"epoch": 3.41,
|
| 810 |
-
"learning_rate": 1.65965064167855e-05,
|
| 811 |
-
"loss": 0.0,
|
| 812 |
-
"step": 67000
|
| 813 |
-
},
|
| 814 |
-
{
|
| 815 |
-
"epoch": 3.44,
|
| 816 |
-
"learning_rate": 1.657104298227745e-05,
|
| 817 |
-
"loss": 0.0,
|
| 818 |
-
"step": 67500
|
| 819 |
-
},
|
| 820 |
-
{
|
| 821 |
-
"epoch": 3.46,
|
| 822 |
-
"learning_rate": 1.6545579547769402e-05,
|
| 823 |
-
"loss": 0.0,
|
| 824 |
-
"step": 68000
|
| 825 |
-
},
|
| 826 |
-
{
|
| 827 |
-
"epoch": 3.49,
|
| 828 |
-
"learning_rate": 1.6520116113261358e-05,
|
| 829 |
-
"loss": 0.0,
|
| 830 |
-
"step": 68500
|
| 831 |
-
},
|
| 832 |
-
{
|
| 833 |
-
"epoch": 3.51,
|
| 834 |
-
"learning_rate": 1.649465267875331e-05,
|
| 835 |
-
"loss": 0.0,
|
| 836 |
-
"step": 69000
|
| 837 |
-
},
|
| 838 |
-
{
|
| 839 |
-
"epoch": 3.54,
|
| 840 |
-
"learning_rate": 1.6469189244245265e-05,
|
| 841 |
-
"loss": 0.0,
|
| 842 |
-
"step": 69500
|
| 843 |
-
},
|
| 844 |
-
{
|
| 845 |
-
"epoch": 3.56,
|
| 846 |
-
"learning_rate": 1.6443725809737217e-05,
|
| 847 |
-
"loss": 0.0,
|
| 848 |
-
"step": 70000
|
| 849 |
-
},
|
| 850 |
-
{
|
| 851 |
-
"epoch": 3.59,
|
| 852 |
-
"learning_rate": 1.6418262375229173e-05,
|
| 853 |
-
"loss": 0.0,
|
| 854 |
-
"step": 70500
|
| 855 |
-
},
|
| 856 |
-
{
|
| 857 |
-
"epoch": 3.62,
|
| 858 |
-
"learning_rate": 1.6392798940721125e-05,
|
| 859 |
-
"loss": 0.0,
|
| 860 |
-
"step": 71000
|
| 861 |
-
},
|
| 862 |
-
{
|
| 863 |
-
"epoch": 3.64,
|
| 864 |
-
"learning_rate": 1.636733550621308e-05,
|
| 865 |
-
"loss": 0.0,
|
| 866 |
-
"step": 71500
|
| 867 |
-
},
|
| 868 |
-
{
|
| 869 |
-
"epoch": 3.67,
|
| 870 |
-
"learning_rate": 1.6341872071705032e-05,
|
| 871 |
-
"loss": 0.0,
|
| 872 |
-
"step": 72000
|
| 873 |
-
},
|
| 874 |
-
{
|
| 875 |
-
"epoch": 3.69,
|
| 876 |
-
"learning_rate": 1.6316408637196984e-05,
|
| 877 |
-
"loss": 0.0,
|
| 878 |
-
"step": 72500
|
| 879 |
-
},
|
| 880 |
-
{
|
| 881 |
-
"epoch": 3.72,
|
| 882 |
-
"learning_rate": 1.629094520268894e-05,
|
| 883 |
-
"loss": 0.0,
|
| 884 |
-
"step": 73000
|
| 885 |
-
},
|
| 886 |
-
{
|
| 887 |
-
"epoch": 3.74,
|
| 888 |
-
"learning_rate": 1.6265481768180892e-05,
|
| 889 |
-
"loss": 0.0,
|
| 890 |
-
"step": 73500
|
| 891 |
-
},
|
| 892 |
-
{
|
| 893 |
-
"epoch": 3.77,
|
| 894 |
-
"learning_rate": 1.6240018333672847e-05,
|
| 895 |
-
"loss": 0.0,
|
| 896 |
-
"step": 74000
|
| 897 |
-
},
|
| 898 |
-
{
|
| 899 |
-
"epoch": 3.79,
|
| 900 |
-
"learning_rate": 1.62145548991648e-05,
|
| 901 |
-
"loss": 0.0,
|
| 902 |
-
"step": 74500
|
| 903 |
-
},
|
| 904 |
-
{
|
| 905 |
-
"epoch": 3.82,
|
| 906 |
-
"learning_rate": 1.6189091464656755e-05,
|
| 907 |
-
"loss": 0.0,
|
| 908 |
-
"step": 75000
|
| 909 |
-
},
|
| 910 |
-
{
|
| 911 |
-
"epoch": 3.84,
|
| 912 |
-
"learning_rate": 1.6163628030148707e-05,
|
| 913 |
-
"loss": 0.0,
|
| 914 |
-
"step": 75500
|
| 915 |
-
},
|
| 916 |
-
{
|
| 917 |
-
"epoch": 3.87,
|
| 918 |
-
"learning_rate": 1.6138164595640662e-05,
|
| 919 |
-
"loss": 0.0,
|
| 920 |
-
"step": 76000
|
| 921 |
-
},
|
| 922 |
-
{
|
| 923 |
-
"epoch": 3.9,
|
| 924 |
-
"learning_rate": 1.6112701161132614e-05,
|
| 925 |
-
"loss": 0.0,
|
| 926 |
-
"step": 76500
|
| 927 |
-
},
|
| 928 |
-
{
|
| 929 |
-
"epoch": 3.92,
|
| 930 |
-
"learning_rate": 1.608723772662457e-05,
|
| 931 |
-
"loss": 0.0,
|
| 932 |
-
"step": 77000
|
| 933 |
-
},
|
| 934 |
-
{
|
| 935 |
-
"epoch": 3.95,
|
| 936 |
-
"learning_rate": 1.606177429211652e-05,
|
| 937 |
-
"loss": 0.0,
|
| 938 |
-
"step": 77500
|
| 939 |
-
},
|
| 940 |
-
{
|
| 941 |
-
"epoch": 3.97,
|
| 942 |
-
"learning_rate": 1.6036310857608474e-05,
|
| 943 |
-
"loss": 0.0,
|
| 944 |
-
"step": 78000
|
| 945 |
-
},
|
| 946 |
-
{
|
| 947 |
-
"epoch": 4.0,
|
| 948 |
-
"learning_rate": 1.601084742310043e-05,
|
| 949 |
-
"loss": 0.0,
|
| 950 |
-
"step": 78500
|
| 951 |
-
},
|
| 952 |
-
{
|
| 953 |
-
"epoch": 4.02,
|
| 954 |
-
"learning_rate": 1.598538398859238e-05,
|
| 955 |
-
"loss": 0.0,
|
| 956 |
-
"step": 79000
|
| 957 |
-
},
|
| 958 |
-
{
|
| 959 |
-
"epoch": 4.05,
|
| 960 |
-
"learning_rate": 1.5959920554084337e-05,
|
| 961 |
-
"loss": 0.0,
|
| 962 |
-
"step": 79500
|
| 963 |
-
},
|
| 964 |
-
{
|
| 965 |
-
"epoch": 4.07,
|
| 966 |
-
"learning_rate": 1.593445711957629e-05,
|
| 967 |
-
"loss": 0.0,
|
| 968 |
-
"step": 80000
|
| 969 |
-
},
|
| 970 |
-
{
|
| 971 |
-
"epoch": 4.1,
|
| 972 |
-
"learning_rate": 1.5908993685068244e-05,
|
| 973 |
-
"loss": 0.0,
|
| 974 |
-
"step": 80500
|
| 975 |
-
},
|
| 976 |
-
{
|
| 977 |
-
"epoch": 4.13,
|
| 978 |
-
"learning_rate": 1.5883530250560196e-05,
|
| 979 |
-
"loss": 0.0,
|
| 980 |
-
"step": 81000
|
| 981 |
-
},
|
| 982 |
-
{
|
| 983 |
-
"epoch": 4.15,
|
| 984 |
-
"learning_rate": 1.585806681605215e-05,
|
| 985 |
-
"loss": 0.0,
|
| 986 |
-
"step": 81500
|
| 987 |
-
},
|
| 988 |
-
{
|
| 989 |
-
"epoch": 4.18,
|
| 990 |
-
"learning_rate": 1.5832603381544104e-05,
|
| 991 |
-
"loss": 0.0,
|
| 992 |
-
"step": 82000
|
| 993 |
-
},
|
| 994 |
-
{
|
| 995 |
-
"epoch": 4.2,
|
| 996 |
-
"learning_rate": 1.5807139947036056e-05,
|
| 997 |
-
"loss": 0.0,
|
| 998 |
-
"step": 82500
|
| 999 |
-
},
|
| 1000 |
-
{
|
| 1001 |
-
"epoch": 4.23,
|
| 1002 |
-
"learning_rate": 1.578167651252801e-05,
|
| 1003 |
-
"loss": 0.0,
|
| 1004 |
-
"step": 83000
|
| 1005 |
-
},
|
| 1006 |
-
{
|
| 1007 |
-
"epoch": 4.25,
|
| 1008 |
-
"learning_rate": 1.5756213078019963e-05,
|
| 1009 |
-
"loss": 0.0,
|
| 1010 |
-
"step": 83500
|
| 1011 |
-
},
|
| 1012 |
-
{
|
| 1013 |
-
"epoch": 4.28,
|
| 1014 |
-
"learning_rate": 1.573074964351192e-05,
|
| 1015 |
-
"loss": 0.0,
|
| 1016 |
-
"step": 84000
|
| 1017 |
-
},
|
| 1018 |
-
{
|
| 1019 |
-
"epoch": 4.3,
|
| 1020 |
-
"learning_rate": 1.570528620900387e-05,
|
| 1021 |
-
"loss": 0.0,
|
| 1022 |
-
"step": 84500
|
| 1023 |
-
},
|
| 1024 |
-
{
|
| 1025 |
-
"epoch": 4.33,
|
| 1026 |
-
"learning_rate": 1.5679822774495826e-05,
|
| 1027 |
-
"loss": 0.0,
|
| 1028 |
-
"step": 85000
|
| 1029 |
-
},
|
| 1030 |
-
{
|
| 1031 |
-
"epoch": 4.35,
|
| 1032 |
-
"learning_rate": 1.5654359339987778e-05,
|
| 1033 |
-
"loss": 0.0,
|
| 1034 |
-
"step": 85500
|
| 1035 |
-
},
|
| 1036 |
-
{
|
| 1037 |
-
"epoch": 4.38,
|
| 1038 |
-
"learning_rate": 1.5628895905479733e-05,
|
| 1039 |
-
"loss": 0.0,
|
| 1040 |
-
"step": 86000
|
| 1041 |
-
},
|
| 1042 |
-
{
|
| 1043 |
-
"epoch": 4.41,
|
| 1044 |
-
"learning_rate": 1.5603432470971686e-05,
|
| 1045 |
-
"loss": 0.0,
|
| 1046 |
-
"step": 86500
|
| 1047 |
-
},
|
| 1048 |
-
{
|
| 1049 |
-
"epoch": 4.43,
|
| 1050 |
-
"learning_rate": 1.557796903646364e-05,
|
| 1051 |
-
"loss": 0.0,
|
| 1052 |
-
"step": 87000
|
| 1053 |
-
},
|
| 1054 |
-
{
|
| 1055 |
-
"epoch": 4.46,
|
| 1056 |
-
"learning_rate": 1.5552505601955593e-05,
|
| 1057 |
-
"loss": 0.0,
|
| 1058 |
-
"step": 87500
|
| 1059 |
-
},
|
| 1060 |
-
{
|
| 1061 |
-
"epoch": 4.48,
|
| 1062 |
-
"learning_rate": 1.5527042167447545e-05,
|
| 1063 |
-
"loss": 0.0,
|
| 1064 |
-
"step": 88000
|
| 1065 |
-
},
|
| 1066 |
-
{
|
| 1067 |
-
"epoch": 4.51,
|
| 1068 |
-
"learning_rate": 1.55015787329395e-05,
|
| 1069 |
-
"loss": 0.0,
|
| 1070 |
-
"step": 88500
|
| 1071 |
-
},
|
| 1072 |
-
{
|
| 1073 |
-
"epoch": 4.53,
|
| 1074 |
-
"learning_rate": 1.5476115298431452e-05,
|
| 1075 |
-
"loss": 0.0,
|
| 1076 |
-
"step": 89000
|
| 1077 |
-
},
|
| 1078 |
-
{
|
| 1079 |
-
"epoch": 4.56,
|
| 1080 |
-
"learning_rate": 1.5450651863923408e-05,
|
| 1081 |
-
"loss": 0.0,
|
| 1082 |
-
"step": 89500
|
| 1083 |
-
},
|
| 1084 |
-
{
|
| 1085 |
-
"epoch": 4.58,
|
| 1086 |
-
"learning_rate": 1.542518842941536e-05,
|
| 1087 |
-
"loss": 0.0,
|
| 1088 |
-
"step": 90000
|
| 1089 |
-
},
|
| 1090 |
-
{
|
| 1091 |
-
"epoch": 4.61,
|
| 1092 |
-
"learning_rate": 1.5399724994907315e-05,
|
| 1093 |
-
"loss": 0.0,
|
| 1094 |
-
"step": 90500
|
| 1095 |
-
},
|
| 1096 |
-
{
|
| 1097 |
-
"epoch": 4.63,
|
| 1098 |
-
"learning_rate": 1.5374261560399267e-05,
|
| 1099 |
-
"loss": 0.0,
|
| 1100 |
-
"step": 91000
|
| 1101 |
-
},
|
| 1102 |
-
{
|
| 1103 |
-
"epoch": 4.66,
|
| 1104 |
-
"learning_rate": 1.5348798125891223e-05,
|
| 1105 |
-
"loss": 0.0,
|
| 1106 |
-
"step": 91500
|
| 1107 |
-
},
|
| 1108 |
-
{
|
| 1109 |
-
"epoch": 4.69,
|
| 1110 |
-
"learning_rate": 1.5323334691383175e-05,
|
| 1111 |
-
"loss": 0.0,
|
| 1112 |
-
"step": 92000
|
| 1113 |
-
},
|
| 1114 |
-
{
|
| 1115 |
-
"epoch": 4.71,
|
| 1116 |
-
"learning_rate": 1.5297871256875127e-05,
|
| 1117 |
-
"loss": 0.0,
|
| 1118 |
-
"step": 92500
|
| 1119 |
-
},
|
| 1120 |
-
{
|
| 1121 |
-
"epoch": 4.74,
|
| 1122 |
-
"learning_rate": 1.5272407822367082e-05,
|
| 1123 |
-
"loss": 0.0,
|
| 1124 |
-
"step": 93000
|
| 1125 |
-
},
|
| 1126 |
-
{
|
| 1127 |
-
"epoch": 4.76,
|
| 1128 |
-
"learning_rate": 1.5246944387859034e-05,
|
| 1129 |
-
"loss": 0.0,
|
| 1130 |
-
"step": 93500
|
| 1131 |
-
},
|
| 1132 |
-
{
|
| 1133 |
-
"epoch": 4.79,
|
| 1134 |
-
"learning_rate": 1.5221480953350988e-05,
|
| 1135 |
-
"loss": 0.0,
|
| 1136 |
-
"step": 94000
|
| 1137 |
-
},
|
| 1138 |
-
{
|
| 1139 |
-
"epoch": 4.81,
|
| 1140 |
-
"learning_rate": 1.5196017518842942e-05,
|
| 1141 |
-
"loss": 0.0,
|
| 1142 |
-
"step": 94500
|
| 1143 |
-
},
|
| 1144 |
-
{
|
| 1145 |
-
"epoch": 4.84,
|
| 1146 |
-
"learning_rate": 1.5170554084334896e-05,
|
| 1147 |
-
"loss": 0.0,
|
| 1148 |
-
"step": 95000
|
| 1149 |
-
},
|
| 1150 |
-
{
|
| 1151 |
-
"epoch": 4.86,
|
| 1152 |
-
"learning_rate": 1.514509064982685e-05,
|
| 1153 |
-
"loss": 0.0,
|
| 1154 |
-
"step": 95500
|
| 1155 |
}
|
| 1156 |
],
|
| 1157 |
"max_steps": 392720,
|
| 1158 |
"num_train_epochs": 20,
|
| 1159 |
-
"total_flos":
|
| 1160 |
"trial_name": null,
|
| 1161 |
"trial_params": null
|
| 1162 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.0,
|
| 5 |
+
"global_step": 19636,
|
| 6 |
"is_hyper_param_search": false,
|
| 7 |
"is_local_process_zero": true,
|
| 8 |
"is_world_process_zero": true,
|
| 9 |
"log_history": [
|
| 10 |
{
|
| 11 |
"epoch": 0.03,
|
| 12 |
+
"learning_rate": 1.9974536565491957e-05,
|
| 13 |
+
"loss": 2.4856,
|
| 14 |
"step": 500
|
| 15 |
},
|
| 16 |
{
|
| 17 |
"epoch": 0.05,
|
| 18 |
+
"learning_rate": 1.994907313098391e-05,
|
| 19 |
+
"loss": 2.2585,
|
| 20 |
"step": 1000
|
| 21 |
},
|
| 22 |
{
|
| 23 |
"epoch": 0.08,
|
| 24 |
+
"learning_rate": 1.9923609696475865e-05,
|
| 25 |
+
"loss": 2.2295,
|
| 26 |
"step": 1500
|
| 27 |
},
|
| 28 |
{
|
| 29 |
"epoch": 0.1,
|
| 30 |
+
"learning_rate": 1.9898146261967817e-05,
|
| 31 |
+
"loss": 2.1886,
|
| 32 |
"step": 2000
|
| 33 |
},
|
| 34 |
{
|
| 35 |
"epoch": 0.13,
|
| 36 |
+
"learning_rate": 1.987268282745977e-05,
|
| 37 |
+
"loss": 2.1856,
|
| 38 |
"step": 2500
|
| 39 |
},
|
| 40 |
{
|
| 41 |
"epoch": 0.15,
|
| 42 |
+
"learning_rate": 1.984721939295172e-05,
|
| 43 |
+
"loss": 2.1617,
|
| 44 |
"step": 3000
|
| 45 |
},
|
| 46 |
{
|
| 47 |
"epoch": 0.18,
|
| 48 |
+
"learning_rate": 1.9821755958443676e-05,
|
| 49 |
+
"loss": 2.1527,
|
| 50 |
"step": 3500
|
| 51 |
},
|
| 52 |
{
|
| 53 |
"epoch": 0.2,
|
| 54 |
+
"learning_rate": 1.9796292523935628e-05,
|
| 55 |
+
"loss": 2.1478,
|
| 56 |
"step": 4000
|
| 57 |
},
|
| 58 |
{
|
| 59 |
"epoch": 0.23,
|
| 60 |
+
"learning_rate": 1.9770829089427584e-05,
|
| 61 |
+
"loss": 2.1309,
|
| 62 |
"step": 4500
|
| 63 |
},
|
| 64 |
{
|
| 65 |
"epoch": 0.25,
|
| 66 |
+
"learning_rate": 1.9745365654919536e-05,
|
| 67 |
+
"loss": 2.1347,
|
| 68 |
"step": 5000
|
| 69 |
},
|
| 70 |
{
|
| 71 |
"epoch": 0.28,
|
| 72 |
+
"learning_rate": 1.971990222041149e-05,
|
| 73 |
+
"loss": 2.1224,
|
| 74 |
"step": 5500
|
| 75 |
},
|
| 76 |
{
|
| 77 |
"epoch": 0.31,
|
| 78 |
+
"learning_rate": 1.9694438785903443e-05,
|
| 79 |
+
"loss": 2.1263,
|
| 80 |
"step": 6000
|
| 81 |
},
|
| 82 |
{
|
| 83 |
"epoch": 0.33,
|
| 84 |
+
"learning_rate": 1.96689753513954e-05,
|
| 85 |
+
"loss": 2.1101,
|
| 86 |
"step": 6500
|
| 87 |
},
|
| 88 |
{
|
| 89 |
"epoch": 0.36,
|
| 90 |
+
"learning_rate": 1.9643511916887354e-05,
|
| 91 |
+
"loss": 2.0876,
|
| 92 |
"step": 7000
|
| 93 |
},
|
| 94 |
{
|
| 95 |
"epoch": 0.38,
|
| 96 |
+
"learning_rate": 1.9618048482379303e-05,
|
| 97 |
+
"loss": 2.0956,
|
| 98 |
"step": 7500
|
| 99 |
},
|
| 100 |
{
|
| 101 |
"epoch": 0.41,
|
| 102 |
+
"learning_rate": 1.9592585047871258e-05,
|
| 103 |
+
"loss": 2.0911,
|
| 104 |
"step": 8000
|
| 105 |
},
|
| 106 |
{
|
| 107 |
"epoch": 0.43,
|
| 108 |
+
"learning_rate": 1.956712161336321e-05,
|
| 109 |
+
"loss": 2.0745,
|
| 110 |
"step": 8500
|
| 111 |
},
|
| 112 |
{
|
| 113 |
"epoch": 0.46,
|
| 114 |
+
"learning_rate": 1.9541658178855165e-05,
|
| 115 |
+
"loss": 2.0619,
|
| 116 |
"step": 9000
|
| 117 |
},
|
| 118 |
{
|
| 119 |
"epoch": 0.48,
|
| 120 |
+
"learning_rate": 1.9516194744347117e-05,
|
| 121 |
+
"loss": 2.0685,
|
| 122 |
"step": 9500
|
| 123 |
},
|
| 124 |
{
|
| 125 |
"epoch": 0.51,
|
| 126 |
+
"learning_rate": 1.9490731309839073e-05,
|
| 127 |
+
"loss": 2.0746,
|
| 128 |
"step": 10000
|
| 129 |
},
|
| 130 |
{
|
| 131 |
"epoch": 0.53,
|
| 132 |
+
"learning_rate": 1.9465267875331025e-05,
|
| 133 |
+
"loss": 2.0701,
|
| 134 |
"step": 10500
|
| 135 |
},
|
| 136 |
{
|
| 137 |
"epoch": 0.56,
|
| 138 |
+
"learning_rate": 1.943980444082298e-05,
|
| 139 |
+
"loss": 2.0793,
|
| 140 |
"step": 11000
|
| 141 |
},
|
| 142 |
{
|
| 143 |
"epoch": 0.59,
|
| 144 |
+
"learning_rate": 1.9414341006314932e-05,
|
| 145 |
+
"loss": 2.063,
|
| 146 |
"step": 11500
|
| 147 |
},
|
| 148 |
{
|
| 149 |
"epoch": 0.61,
|
| 150 |
+
"learning_rate": 1.9388877571806888e-05,
|
| 151 |
+
"loss": 2.0668,
|
| 152 |
"step": 12000
|
| 153 |
},
|
| 154 |
{
|
| 155 |
"epoch": 0.64,
|
| 156 |
+
"learning_rate": 1.936341413729884e-05,
|
| 157 |
+
"loss": 2.0458,
|
| 158 |
"step": 12500
|
| 159 |
},
|
| 160 |
{
|
| 161 |
"epoch": 0.66,
|
| 162 |
+
"learning_rate": 1.9337950702790792e-05,
|
| 163 |
+
"loss": 2.053,
|
| 164 |
"step": 13000
|
| 165 |
},
|
| 166 |
{
|
| 167 |
"epoch": 0.69,
|
| 168 |
+
"learning_rate": 1.9312487268282747e-05,
|
| 169 |
+
"loss": 2.043,
|
| 170 |
"step": 13500
|
| 171 |
},
|
| 172 |
{
|
| 173 |
"epoch": 0.71,
|
| 174 |
+
"learning_rate": 1.92870238337747e-05,
|
| 175 |
+
"loss": 2.0302,
|
| 176 |
"step": 14000
|
| 177 |
},
|
| 178 |
{
|
| 179 |
"epoch": 0.74,
|
| 180 |
+
"learning_rate": 1.9261560399266655e-05,
|
| 181 |
+
"loss": 2.0444,
|
| 182 |
"step": 14500
|
| 183 |
},
|
| 184 |
{
|
| 185 |
"epoch": 0.76,
|
| 186 |
+
"learning_rate": 1.9236096964758607e-05,
|
| 187 |
+
"loss": 2.0476,
|
| 188 |
"step": 15000
|
| 189 |
},
|
| 190 |
{
|
| 191 |
"epoch": 0.79,
|
| 192 |
+
"learning_rate": 1.9210633530250562e-05,
|
| 193 |
+
"loss": 2.0317,
|
| 194 |
"step": 15500
|
| 195 |
},
|
| 196 |
{
|
| 197 |
"epoch": 0.81,
|
| 198 |
+
"learning_rate": 1.9185170095742514e-05,
|
| 199 |
+
"loss": 2.0349,
|
| 200 |
"step": 16000
|
| 201 |
},
|
| 202 |
{
|
| 203 |
"epoch": 0.84,
|
| 204 |
+
"learning_rate": 1.915970666123447e-05,
|
| 205 |
+
"loss": 2.0391,
|
| 206 |
"step": 16500
|
| 207 |
},
|
| 208 |
{
|
| 209 |
"epoch": 0.87,
|
| 210 |
+
"learning_rate": 1.9134243226726422e-05,
|
| 211 |
+
"loss": 2.0293,
|
| 212 |
"step": 17000
|
| 213 |
},
|
| 214 |
{
|
| 215 |
"epoch": 0.89,
|
| 216 |
+
"learning_rate": 1.9108779792218374e-05,
|
| 217 |
+
"loss": 2.0164,
|
| 218 |
"step": 17500
|
| 219 |
},
|
| 220 |
{
|
| 221 |
"epoch": 0.92,
|
| 222 |
+
"learning_rate": 1.908331635771033e-05,
|
| 223 |
+
"loss": 2.0337,
|
| 224 |
"step": 18000
|
| 225 |
},
|
| 226 |
{
|
| 227 |
"epoch": 0.94,
|
| 228 |
+
"learning_rate": 1.905785292320228e-05,
|
| 229 |
+
"loss": 2.0427,
|
| 230 |
"step": 18500
|
| 231 |
},
|
| 232 |
{
|
| 233 |
"epoch": 0.97,
|
| 234 |
+
"learning_rate": 1.9032389488694237e-05,
|
| 235 |
+
"loss": 2.0337,
|
| 236 |
"step": 19000
|
| 237 |
},
|
| 238 |
{
|
| 239 |
"epoch": 0.99,
|
| 240 |
+
"learning_rate": 1.900692605418619e-05,
|
| 241 |
+
"loss": 2.0304,
|
| 242 |
"step": 19500
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 243 |
}
|
| 244 |
],
|
| 245 |
"max_steps": 392720,
|
| 246 |
"num_train_epochs": 20,
|
| 247 |
+
"total_flos": 7.512462439371571e+16,
|
| 248 |
"trial_name": null,
|
| 249 |
"trial_params": null
|
| 250 |
}
|
last-checkpoint/training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 3311
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:69d9fd3ad18563c1684c7b33cda0775c0390d2503688d8228ae9022dc5e43c7b
|
| 3 |
size 3311
|
pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 2279605745
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8bccecd638f394225805df88227fbba5c325f109230acde5f20671f48f6d0a30
|
| 3 |
size 2279605745
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 3311
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:69d9fd3ad18563c1684c7b33cda0775c0390d2503688d8228ae9022dc5e43c7b
|
| 3 |
size 3311
|