Training in progress, step 1000, checkpoint
Browse files
last-checkpoint/model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 966995080
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3df59285835e0cc93eed8a07997106068a6025a1eecf29ca1883050640082bf5
|
| 3 |
size 966995080
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1925064044
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ba8b8aca45c706933df0ac865eb25d0a5f5734f04a35d51d8b5f659916db1f8f
|
| 3 |
size 1925064044
|
last-checkpoint/rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14244
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d1f09b1f1f9b06ad2afb12e89fc8695073b76afcf9ea0b3552c7069932117824
|
| 3 |
size 14244
|
last-checkpoint/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1064
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5c2e7a67c9c301b36183def727305bd60ef4c597b197ad54cdb0001ffc36e45a
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,1177 +1,310 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric":
|
| 3 |
-
"best_model_checkpoint": "./whisper-small-taiwanese/checkpoint-
|
| 4 |
-
"epoch":
|
| 5 |
"eval_steps": 1000,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
-
"epoch": 0.
|
| 13 |
-
"grad_norm":
|
| 14 |
-
"learning_rate": 5.
|
| 15 |
-
"loss":
|
| 16 |
"step": 25
|
| 17 |
},
|
| 18 |
{
|
| 19 |
-
"epoch": 0.
|
| 20 |
-
"grad_norm":
|
| 21 |
-
"learning_rate": 1.
|
| 22 |
-
"loss": 5.
|
| 23 |
"step": 50
|
| 24 |
},
|
| 25 |
{
|
| 26 |
-
"epoch": 0.
|
| 27 |
-
"grad_norm": 32.
|
| 28 |
-
"learning_rate": 1.
|
| 29 |
-
"loss": 3.
|
| 30 |
"step": 75
|
| 31 |
},
|
| 32 |
{
|
| 33 |
-
"epoch": 0.
|
| 34 |
-
"grad_norm":
|
| 35 |
-
"learning_rate": 2.
|
| 36 |
-
"loss": 2.
|
| 37 |
"step": 100
|
| 38 |
},
|
| 39 |
{
|
| 40 |
-
"epoch": 0.
|
| 41 |
-
"grad_norm":
|
| 42 |
-
"learning_rate": 2.
|
| 43 |
-
"loss": 2.
|
| 44 |
"step": 125
|
| 45 |
},
|
| 46 |
{
|
| 47 |
-
"epoch": 0.
|
| 48 |
-
"grad_norm":
|
| 49 |
-
"learning_rate":
|
| 50 |
-
"loss": 2.
|
| 51 |
"step": 150
|
| 52 |
},
|
| 53 |
{
|
| 54 |
-
"epoch": 0.
|
| 55 |
-
"grad_norm":
|
| 56 |
-
"learning_rate": 3.
|
| 57 |
-
"loss":
|
| 58 |
"step": 175
|
| 59 |
},
|
| 60 |
{
|
| 61 |
-
"epoch": 0.
|
| 62 |
-
"grad_norm":
|
| 63 |
-
"learning_rate": 4.
|
| 64 |
-
"loss": 1.
|
| 65 |
"step": 200
|
| 66 |
},
|
| 67 |
{
|
| 68 |
-
"epoch": 0.
|
| 69 |
-
"grad_norm": 18.
|
| 70 |
-
"learning_rate": 4.
|
| 71 |
-
"loss": 1.
|
| 72 |
"step": 225
|
| 73 |
},
|
| 74 |
{
|
| 75 |
-
"epoch": 0.
|
| 76 |
-
"grad_norm": 15.
|
| 77 |
-
"learning_rate":
|
| 78 |
-
"loss": 1.
|
| 79 |
"step": 250
|
| 80 |
},
|
| 81 |
{
|
| 82 |
-
"epoch": 0.
|
| 83 |
-
"grad_norm":
|
| 84 |
-
"learning_rate": 5.
|
| 85 |
-
"loss": 1.
|
| 86 |
"step": 275
|
| 87 |
},
|
| 88 |
{
|
| 89 |
-
"epoch": 0.
|
| 90 |
-
"grad_norm": 15.
|
| 91 |
-
"learning_rate":
|
| 92 |
-
"loss": 1.
|
| 93 |
"step": 300
|
| 94 |
},
|
| 95 |
{
|
| 96 |
-
"epoch": 0.
|
| 97 |
-
"grad_norm":
|
| 98 |
-
"learning_rate": 6.
|
| 99 |
-
"loss": 1.
|
| 100 |
"step": 325
|
| 101 |
},
|
| 102 |
{
|
| 103 |
-
"epoch": 0.
|
| 104 |
-
"grad_norm": 16.
|
| 105 |
-
"learning_rate":
|
| 106 |
-
"loss": 1.
|
| 107 |
"step": 350
|
| 108 |
},
|
| 109 |
{
|
| 110 |
-
"epoch": 0.
|
| 111 |
-
"grad_norm":
|
| 112 |
-
"learning_rate":
|
| 113 |
-
"loss": 1.
|
| 114 |
"step": 375
|
| 115 |
},
|
| 116 |
{
|
| 117 |
-
"epoch": 0.
|
| 118 |
-
"grad_norm":
|
| 119 |
-
"learning_rate": 8.
|
| 120 |
-
"loss": 1.
|
| 121 |
"step": 400
|
| 122 |
},
|
| 123 |
{
|
| 124 |
-
"epoch": 0.
|
| 125 |
-
"grad_norm": 14.
|
| 126 |
-
"learning_rate":
|
| 127 |
-
"loss": 1.
|
| 128 |
"step": 425
|
| 129 |
},
|
| 130 |
{
|
| 131 |
-
"epoch": 0.
|
| 132 |
-
"grad_norm": 13.
|
| 133 |
-
"learning_rate":
|
| 134 |
-
"loss": 1.
|
| 135 |
"step": 450
|
| 136 |
},
|
| 137 |
{
|
| 138 |
-
"epoch": 0.
|
| 139 |
-
"grad_norm":
|
| 140 |
-
"learning_rate": 9.
|
| 141 |
-
"loss": 1.
|
| 142 |
"step": 475
|
| 143 |
},
|
| 144 |
{
|
| 145 |
-
"epoch": 0.
|
| 146 |
-
"grad_norm": 14.
|
| 147 |
-
"learning_rate":
|
| 148 |
-
"loss": 1.
|
| 149 |
"step": 500
|
| 150 |
},
|
| 151 |
{
|
| 152 |
-
"epoch": 0.
|
| 153 |
-
"grad_norm":
|
| 154 |
-
"learning_rate": 9.
|
| 155 |
-
"loss": 1.
|
| 156 |
"step": 525
|
| 157 |
},
|
| 158 |
{
|
| 159 |
-
"epoch": 0.
|
| 160 |
-
"grad_norm":
|
| 161 |
-
"learning_rate": 9.
|
| 162 |
-
"loss": 1.
|
| 163 |
"step": 550
|
| 164 |
},
|
| 165 |
{
|
| 166 |
-
"epoch": 0.
|
| 167 |
-
"grad_norm":
|
| 168 |
-
"learning_rate": 9.
|
| 169 |
-
"loss": 1.
|
| 170 |
"step": 575
|
| 171 |
},
|
| 172 |
{
|
| 173 |
-
"epoch": 0.
|
| 174 |
-
"grad_norm":
|
| 175 |
-
"learning_rate": 9.
|
| 176 |
-
"loss": 1.
|
| 177 |
"step": 600
|
| 178 |
},
|
| 179 |
{
|
| 180 |
-
"epoch": 0.
|
| 181 |
-
"grad_norm": 15.
|
| 182 |
-
"learning_rate": 9.
|
| 183 |
-
"loss": 1.
|
| 184 |
"step": 625
|
| 185 |
},
|
| 186 |
{
|
| 187 |
-
"epoch": 0.
|
| 188 |
-
"grad_norm": 13.
|
| 189 |
-
"learning_rate": 9.
|
| 190 |
-
"loss": 1.
|
| 191 |
"step": 650
|
| 192 |
},
|
| 193 |
{
|
| 194 |
-
"epoch": 0.
|
| 195 |
-
"grad_norm":
|
| 196 |
-
"learning_rate": 9.
|
| 197 |
-
"loss": 1.
|
| 198 |
"step": 675
|
| 199 |
},
|
| 200 |
{
|
| 201 |
-
"epoch": 0.
|
| 202 |
-
"grad_norm": 14.
|
| 203 |
-
"learning_rate": 9.
|
| 204 |
-
"loss": 1.
|
| 205 |
"step": 700
|
| 206 |
},
|
| 207 |
{
|
| 208 |
-
"epoch": 0.
|
| 209 |
-
"grad_norm":
|
| 210 |
-
"learning_rate": 9.
|
| 211 |
-
"loss": 1.
|
| 212 |
"step": 725
|
| 213 |
},
|
| 214 |
{
|
| 215 |
-
"epoch": 0.
|
| 216 |
-
"grad_norm":
|
| 217 |
-
"learning_rate": 9.
|
| 218 |
-
"loss": 1.
|
| 219 |
"step": 750
|
| 220 |
},
|
| 221 |
{
|
| 222 |
-
"epoch": 0.
|
| 223 |
-
"grad_norm":
|
| 224 |
-
"learning_rate": 9.
|
| 225 |
-
"loss": 1.
|
| 226 |
"step": 775
|
| 227 |
},
|
| 228 |
{
|
| 229 |
-
"epoch": 0.
|
| 230 |
-
"grad_norm":
|
| 231 |
-
"learning_rate": 9.
|
| 232 |
-
"loss": 1.
|
| 233 |
"step": 800
|
| 234 |
},
|
| 235 |
{
|
| 236 |
-
"epoch": 0.
|
| 237 |
-
"grad_norm":
|
| 238 |
-
"learning_rate": 9.
|
| 239 |
-
"loss": 1.
|
| 240 |
"step": 825
|
| 241 |
},
|
| 242 |
{
|
| 243 |
-
"epoch": 0.
|
| 244 |
-
"grad_norm": 15.
|
| 245 |
-
"learning_rate": 9.
|
| 246 |
-
"loss":
|
| 247 |
"step": 850
|
| 248 |
},
|
| 249 |
{
|
| 250 |
-
"epoch": 0.
|
| 251 |
-
"grad_norm":
|
| 252 |
-
"learning_rate": 9.
|
| 253 |
-
"loss": 1.
|
| 254 |
"step": 875
|
| 255 |
},
|
| 256 |
{
|
| 257 |
-
"epoch": 0.
|
| 258 |
-
"grad_norm":
|
| 259 |
-
"learning_rate":
|
| 260 |
-
"loss":
|
| 261 |
"step": 900
|
| 262 |
},
|
| 263 |
{
|
| 264 |
-
"epoch": 0.
|
| 265 |
-
"grad_norm":
|
| 266 |
-
"learning_rate":
|
| 267 |
-
"loss": 1.
|
| 268 |
"step": 925
|
| 269 |
},
|
| 270 |
{
|
| 271 |
-
"epoch": 0.
|
| 272 |
-
"grad_norm":
|
| 273 |
-
"learning_rate": 8.
|
| 274 |
-
"loss":
|
| 275 |
"step": 950
|
| 276 |
},
|
| 277 |
{
|
| 278 |
-
"epoch": 0.
|
| 279 |
-
"grad_norm": 13.
|
| 280 |
-
"learning_rate": 8.
|
| 281 |
-
"loss":
|
| 282 |
"step": 975
|
| 283 |
},
|
| 284 |
{
|
| 285 |
-
"epoch": 0.
|
| 286 |
-
"grad_norm":
|
| 287 |
-
"learning_rate": 8.
|
| 288 |
-
"loss":
|
| 289 |
"step": 1000
|
| 290 |
},
|
| 291 |
{
|
| 292 |
-
"epoch": 0.
|
| 293 |
-
"eval_cer":
|
| 294 |
-
"eval_loss":
|
| 295 |
-
"eval_runtime":
|
| 296 |
-
"eval_samples_per_second": 2.
|
| 297 |
-
"eval_steps_per_second": 0.
|
| 298 |
"step": 1000
|
| 299 |
-
},
|
| 300 |
-
{
|
| 301 |
-
"epoch": 0.6612903225806451,
|
| 302 |
-
"grad_norm": 11.495616912841797,
|
| 303 |
-
"learning_rate": 8.66188769414576e-06,
|
| 304 |
-
"loss": 0.9695,
|
| 305 |
-
"step": 1025
|
| 306 |
-
},
|
| 307 |
-
{
|
| 308 |
-
"epoch": 0.6774193548387096,
|
| 309 |
-
"grad_norm": 15.224388122558594,
|
| 310 |
-
"learning_rate": 8.602150537634409e-06,
|
| 311 |
-
"loss": 0.9488,
|
| 312 |
-
"step": 1050
|
| 313 |
-
},
|
| 314 |
-
{
|
| 315 |
-
"epoch": 0.6935483870967742,
|
| 316 |
-
"grad_norm": 13.824469566345215,
|
| 317 |
-
"learning_rate": 8.54241338112306e-06,
|
| 318 |
-
"loss": 1.0474,
|
| 319 |
-
"step": 1075
|
| 320 |
-
},
|
| 321 |
-
{
|
| 322 |
-
"epoch": 0.7096774193548387,
|
| 323 |
-
"grad_norm": 14.53409194946289,
|
| 324 |
-
"learning_rate": 8.48267622461171e-06,
|
| 325 |
-
"loss": 0.9866,
|
| 326 |
-
"step": 1100
|
| 327 |
-
},
|
| 328 |
-
{
|
| 329 |
-
"epoch": 0.7258064516129032,
|
| 330 |
-
"grad_norm": 12.956225395202637,
|
| 331 |
-
"learning_rate": 8.422939068100358e-06,
|
| 332 |
-
"loss": 0.9072,
|
| 333 |
-
"step": 1125
|
| 334 |
-
},
|
| 335 |
-
{
|
| 336 |
-
"epoch": 0.7419354838709677,
|
| 337 |
-
"grad_norm": 13.533162117004395,
|
| 338 |
-
"learning_rate": 8.36320191158901e-06,
|
| 339 |
-
"loss": 0.9428,
|
| 340 |
-
"step": 1150
|
| 341 |
-
},
|
| 342 |
-
{
|
| 343 |
-
"epoch": 0.7580645161290323,
|
| 344 |
-
"grad_norm": 14.72665023803711,
|
| 345 |
-
"learning_rate": 8.303464755077659e-06,
|
| 346 |
-
"loss": 0.9387,
|
| 347 |
-
"step": 1175
|
| 348 |
-
},
|
| 349 |
-
{
|
| 350 |
-
"epoch": 0.7741935483870968,
|
| 351 |
-
"grad_norm": 12.921445846557617,
|
| 352 |
-
"learning_rate": 8.24372759856631e-06,
|
| 353 |
-
"loss": 0.9442,
|
| 354 |
-
"step": 1200
|
| 355 |
-
},
|
| 356 |
-
{
|
| 357 |
-
"epoch": 0.7903225806451613,
|
| 358 |
-
"grad_norm": 11.652874946594238,
|
| 359 |
-
"learning_rate": 8.18399044205496e-06,
|
| 360 |
-
"loss": 0.9359,
|
| 361 |
-
"step": 1225
|
| 362 |
-
},
|
| 363 |
-
{
|
| 364 |
-
"epoch": 0.8064516129032258,
|
| 365 |
-
"grad_norm": 15.415846824645996,
|
| 366 |
-
"learning_rate": 8.124253285543608e-06,
|
| 367 |
-
"loss": 0.916,
|
| 368 |
-
"step": 1250
|
| 369 |
-
},
|
| 370 |
-
{
|
| 371 |
-
"epoch": 0.8225806451612904,
|
| 372 |
-
"grad_norm": 18.422143936157227,
|
| 373 |
-
"learning_rate": 8.064516129032258e-06,
|
| 374 |
-
"loss": 0.9608,
|
| 375 |
-
"step": 1275
|
| 376 |
-
},
|
| 377 |
-
{
|
| 378 |
-
"epoch": 0.8387096774193549,
|
| 379 |
-
"grad_norm": 11.93355941772461,
|
| 380 |
-
"learning_rate": 8.004778972520909e-06,
|
| 381 |
-
"loss": 0.9297,
|
| 382 |
-
"step": 1300
|
| 383 |
-
},
|
| 384 |
-
{
|
| 385 |
-
"epoch": 0.8548387096774194,
|
| 386 |
-
"grad_norm": 16.42209243774414,
|
| 387 |
-
"learning_rate": 7.945041816009559e-06,
|
| 388 |
-
"loss": 0.8933,
|
| 389 |
-
"step": 1325
|
| 390 |
-
},
|
| 391 |
-
{
|
| 392 |
-
"epoch": 0.8709677419354839,
|
| 393 |
-
"grad_norm": 14.272250175476074,
|
| 394 |
-
"learning_rate": 7.88530465949821e-06,
|
| 395 |
-
"loss": 0.9185,
|
| 396 |
-
"step": 1350
|
| 397 |
-
},
|
| 398 |
-
{
|
| 399 |
-
"epoch": 0.8870967741935484,
|
| 400 |
-
"grad_norm": 12.172361373901367,
|
| 401 |
-
"learning_rate": 7.825567502986858e-06,
|
| 402 |
-
"loss": 0.8476,
|
| 403 |
-
"step": 1375
|
| 404 |
-
},
|
| 405 |
-
{
|
| 406 |
-
"epoch": 0.9032258064516129,
|
| 407 |
-
"grad_norm": 14.475882530212402,
|
| 408 |
-
"learning_rate": 7.765830346475508e-06,
|
| 409 |
-
"loss": 0.925,
|
| 410 |
-
"step": 1400
|
| 411 |
-
},
|
| 412 |
-
{
|
| 413 |
-
"epoch": 0.9193548387096774,
|
| 414 |
-
"grad_norm": 14.247998237609863,
|
| 415 |
-
"learning_rate": 7.706093189964159e-06,
|
| 416 |
-
"loss": 0.888,
|
| 417 |
-
"step": 1425
|
| 418 |
-
},
|
| 419 |
-
{
|
| 420 |
-
"epoch": 0.9354838709677419,
|
| 421 |
-
"grad_norm": 12.855352401733398,
|
| 422 |
-
"learning_rate": 7.646356033452809e-06,
|
| 423 |
-
"loss": 0.888,
|
| 424 |
-
"step": 1450
|
| 425 |
-
},
|
| 426 |
-
{
|
| 427 |
-
"epoch": 0.9516129032258065,
|
| 428 |
-
"grad_norm": 14.016806602478027,
|
| 429 |
-
"learning_rate": 7.586618876941458e-06,
|
| 430 |
-
"loss": 0.9237,
|
| 431 |
-
"step": 1475
|
| 432 |
-
},
|
| 433 |
-
{
|
| 434 |
-
"epoch": 0.967741935483871,
|
| 435 |
-
"grad_norm": 13.113448143005371,
|
| 436 |
-
"learning_rate": 7.526881720430108e-06,
|
| 437 |
-
"loss": 0.8767,
|
| 438 |
-
"step": 1500
|
| 439 |
-
},
|
| 440 |
-
{
|
| 441 |
-
"epoch": 0.9838709677419355,
|
| 442 |
-
"grad_norm": 15.823156356811523,
|
| 443 |
-
"learning_rate": 7.467144563918758e-06,
|
| 444 |
-
"loss": 0.8561,
|
| 445 |
-
"step": 1525
|
| 446 |
-
},
|
| 447 |
-
{
|
| 448 |
-
"epoch": 1.0,
|
| 449 |
-
"grad_norm": 16.72173309326172,
|
| 450 |
-
"learning_rate": 7.4074074074074075e-06,
|
| 451 |
-
"loss": 0.853,
|
| 452 |
-
"step": 1550
|
| 453 |
-
},
|
| 454 |
-
{
|
| 455 |
-
"epoch": 1.0161290322580645,
|
| 456 |
-
"grad_norm": 11.821678161621094,
|
| 457 |
-
"learning_rate": 7.347670250896059e-06,
|
| 458 |
-
"loss": 0.6325,
|
| 459 |
-
"step": 1575
|
| 460 |
-
},
|
| 461 |
-
{
|
| 462 |
-
"epoch": 1.032258064516129,
|
| 463 |
-
"grad_norm": 10.003717422485352,
|
| 464 |
-
"learning_rate": 7.287933094384708e-06,
|
| 465 |
-
"loss": 0.5999,
|
| 466 |
-
"step": 1600
|
| 467 |
-
},
|
| 468 |
-
{
|
| 469 |
-
"epoch": 1.0483870967741935,
|
| 470 |
-
"grad_norm": 11.098932266235352,
|
| 471 |
-
"learning_rate": 7.2281959378733575e-06,
|
| 472 |
-
"loss": 0.6439,
|
| 473 |
-
"step": 1625
|
| 474 |
-
},
|
| 475 |
-
{
|
| 476 |
-
"epoch": 1.064516129032258,
|
| 477 |
-
"grad_norm": 13.39173412322998,
|
| 478 |
-
"learning_rate": 7.168458781362008e-06,
|
| 479 |
-
"loss": 0.6171,
|
| 480 |
-
"step": 1650
|
| 481 |
-
},
|
| 482 |
-
{
|
| 483 |
-
"epoch": 1.0806451612903225,
|
| 484 |
-
"grad_norm": 12.478330612182617,
|
| 485 |
-
"learning_rate": 7.108721624850657e-06,
|
| 486 |
-
"loss": 0.6296,
|
| 487 |
-
"step": 1675
|
| 488 |
-
},
|
| 489 |
-
{
|
| 490 |
-
"epoch": 1.096774193548387,
|
| 491 |
-
"grad_norm": 11.143562316894531,
|
| 492 |
-
"learning_rate": 7.048984468339307e-06,
|
| 493 |
-
"loss": 0.6252,
|
| 494 |
-
"step": 1700
|
| 495 |
-
},
|
| 496 |
-
{
|
| 497 |
-
"epoch": 1.1129032258064515,
|
| 498 |
-
"grad_norm": 9.06653118133545,
|
| 499 |
-
"learning_rate": 6.989247311827958e-06,
|
| 500 |
-
"loss": 0.627,
|
| 501 |
-
"step": 1725
|
| 502 |
-
},
|
| 503 |
-
{
|
| 504 |
-
"epoch": 1.129032258064516,
|
| 505 |
-
"grad_norm": 12.985542297363281,
|
| 506 |
-
"learning_rate": 6.929510155316607e-06,
|
| 507 |
-
"loss": 0.6145,
|
| 508 |
-
"step": 1750
|
| 509 |
-
},
|
| 510 |
-
{
|
| 511 |
-
"epoch": 1.1451612903225807,
|
| 512 |
-
"grad_norm": 12.124594688415527,
|
| 513 |
-
"learning_rate": 6.869772998805258e-06,
|
| 514 |
-
"loss": 0.601,
|
| 515 |
-
"step": 1775
|
| 516 |
-
},
|
| 517 |
-
{
|
| 518 |
-
"epoch": 1.1612903225806452,
|
| 519 |
-
"grad_norm": 11.50346851348877,
|
| 520 |
-
"learning_rate": 6.810035842293907e-06,
|
| 521 |
-
"loss": 0.5787,
|
| 522 |
-
"step": 1800
|
| 523 |
-
},
|
| 524 |
-
{
|
| 525 |
-
"epoch": 1.1774193548387097,
|
| 526 |
-
"grad_norm": 11.256744384765625,
|
| 527 |
-
"learning_rate": 6.7502986857825566e-06,
|
| 528 |
-
"loss": 0.5949,
|
| 529 |
-
"step": 1825
|
| 530 |
-
},
|
| 531 |
-
{
|
| 532 |
-
"epoch": 1.1935483870967742,
|
| 533 |
-
"grad_norm": 12.568142890930176,
|
| 534 |
-
"learning_rate": 6.690561529271207e-06,
|
| 535 |
-
"loss": 0.6396,
|
| 536 |
-
"step": 1850
|
| 537 |
-
},
|
| 538 |
-
{
|
| 539 |
-
"epoch": 1.2096774193548387,
|
| 540 |
-
"grad_norm": 11.688636779785156,
|
| 541 |
-
"learning_rate": 6.630824372759857e-06,
|
| 542 |
-
"loss": 0.6106,
|
| 543 |
-
"step": 1875
|
| 544 |
-
},
|
| 545 |
-
{
|
| 546 |
-
"epoch": 1.2258064516129032,
|
| 547 |
-
"grad_norm": 13.135574340820312,
|
| 548 |
-
"learning_rate": 6.5710872162485075e-06,
|
| 549 |
-
"loss": 0.6197,
|
| 550 |
-
"step": 1900
|
| 551 |
-
},
|
| 552 |
-
{
|
| 553 |
-
"epoch": 1.2419354838709677,
|
| 554 |
-
"grad_norm": 14.128840446472168,
|
| 555 |
-
"learning_rate": 6.511350059737157e-06,
|
| 556 |
-
"loss": 0.6474,
|
| 557 |
-
"step": 1925
|
| 558 |
-
},
|
| 559 |
-
{
|
| 560 |
-
"epoch": 1.2580645161290323,
|
| 561 |
-
"grad_norm": 11.889117240905762,
|
| 562 |
-
"learning_rate": 6.451612903225806e-06,
|
| 563 |
-
"loss": 0.5966,
|
| 564 |
-
"step": 1950
|
| 565 |
-
},
|
| 566 |
-
{
|
| 567 |
-
"epoch": 1.2741935483870968,
|
| 568 |
-
"grad_norm": 12.298087120056152,
|
| 569 |
-
"learning_rate": 6.391875746714457e-06,
|
| 570 |
-
"loss": 0.6007,
|
| 571 |
-
"step": 1975
|
| 572 |
-
},
|
| 573 |
-
{
|
| 574 |
-
"epoch": 1.2903225806451613,
|
| 575 |
-
"grad_norm": 13.969961166381836,
|
| 576 |
-
"learning_rate": 6.332138590203107e-06,
|
| 577 |
-
"loss": 0.61,
|
| 578 |
-
"step": 2000
|
| 579 |
-
},
|
| 580 |
-
{
|
| 581 |
-
"epoch": 1.2903225806451613,
|
| 582 |
-
"eval_cer": 53.38840841616109,
|
| 583 |
-
"eval_loss": 0.753625750541687,
|
| 584 |
-
"eval_runtime": 951.328,
|
| 585 |
-
"eval_samples_per_second": 2.398,
|
| 586 |
-
"eval_steps_per_second": 0.301,
|
| 587 |
-
"step": 2000
|
| 588 |
-
},
|
| 589 |
-
{
|
| 590 |
-
"epoch": 1.3064516129032258,
|
| 591 |
-
"grad_norm": 9.99063777923584,
|
| 592 |
-
"learning_rate": 6.272401433691757e-06,
|
| 593 |
-
"loss": 0.588,
|
| 594 |
-
"step": 2025
|
| 595 |
-
},
|
| 596 |
-
{
|
| 597 |
-
"epoch": 1.3225806451612903,
|
| 598 |
-
"grad_norm": 13.123091697692871,
|
| 599 |
-
"learning_rate": 6.212664277180407e-06,
|
| 600 |
-
"loss": 0.5886,
|
| 601 |
-
"step": 2050
|
| 602 |
-
},
|
| 603 |
-
{
|
| 604 |
-
"epoch": 1.3387096774193548,
|
| 605 |
-
"grad_norm": 10.930394172668457,
|
| 606 |
-
"learning_rate": 6.152927120669057e-06,
|
| 607 |
-
"loss": 0.6117,
|
| 608 |
-
"step": 2075
|
| 609 |
-
},
|
| 610 |
-
{
|
| 611 |
-
"epoch": 1.3548387096774195,
|
| 612 |
-
"grad_norm": 12.531543731689453,
|
| 613 |
-
"learning_rate": 6.0931899641577065e-06,
|
| 614 |
-
"loss": 0.5931,
|
| 615 |
-
"step": 2100
|
| 616 |
-
},
|
| 617 |
-
{
|
| 618 |
-
"epoch": 1.370967741935484,
|
| 619 |
-
"grad_norm": 13.16308307647705,
|
| 620 |
-
"learning_rate": 6.033452807646356e-06,
|
| 621 |
-
"loss": 0.598,
|
| 622 |
-
"step": 2125
|
| 623 |
-
},
|
| 624 |
-
{
|
| 625 |
-
"epoch": 1.3870967741935485,
|
| 626 |
-
"grad_norm": 11.17799186706543,
|
| 627 |
-
"learning_rate": 5.973715651135007e-06,
|
| 628 |
-
"loss": 0.6141,
|
| 629 |
-
"step": 2150
|
| 630 |
-
},
|
| 631 |
-
{
|
| 632 |
-
"epoch": 1.403225806451613,
|
| 633 |
-
"grad_norm": 10.640506744384766,
|
| 634 |
-
"learning_rate": 5.9139784946236566e-06,
|
| 635 |
-
"loss": 0.5682,
|
| 636 |
-
"step": 2175
|
| 637 |
-
},
|
| 638 |
-
{
|
| 639 |
-
"epoch": 1.4193548387096775,
|
| 640 |
-
"grad_norm": 11.789594650268555,
|
| 641 |
-
"learning_rate": 5.854241338112307e-06,
|
| 642 |
-
"loss": 0.5598,
|
| 643 |
-
"step": 2200
|
| 644 |
-
},
|
| 645 |
-
{
|
| 646 |
-
"epoch": 1.435483870967742,
|
| 647 |
-
"grad_norm": 11.937474250793457,
|
| 648 |
-
"learning_rate": 5.794504181600956e-06,
|
| 649 |
-
"loss": 0.6344,
|
| 650 |
-
"step": 2225
|
| 651 |
-
},
|
| 652 |
-
{
|
| 653 |
-
"epoch": 1.4516129032258065,
|
| 654 |
-
"grad_norm": 14.106030464172363,
|
| 655 |
-
"learning_rate": 5.734767025089606e-06,
|
| 656 |
-
"loss": 0.5783,
|
| 657 |
-
"step": 2250
|
| 658 |
-
},
|
| 659 |
-
{
|
| 660 |
-
"epoch": 1.467741935483871,
|
| 661 |
-
"grad_norm": 12.365781784057617,
|
| 662 |
-
"learning_rate": 5.675029868578256e-06,
|
| 663 |
-
"loss": 0.6335,
|
| 664 |
-
"step": 2275
|
| 665 |
-
},
|
| 666 |
-
{
|
| 667 |
-
"epoch": 1.4838709677419355,
|
| 668 |
-
"grad_norm": 14.670917510986328,
|
| 669 |
-
"learning_rate": 5.615292712066906e-06,
|
| 670 |
-
"loss": 0.5988,
|
| 671 |
-
"step": 2300
|
| 672 |
-
},
|
| 673 |
-
{
|
| 674 |
-
"epoch": 1.5,
|
| 675 |
-
"grad_norm": 10.45535659790039,
|
| 676 |
-
"learning_rate": 5.555555555555557e-06,
|
| 677 |
-
"loss": 0.5912,
|
| 678 |
-
"step": 2325
|
| 679 |
-
},
|
| 680 |
-
{
|
| 681 |
-
"epoch": 1.5161290322580645,
|
| 682 |
-
"grad_norm": 15.059216499328613,
|
| 683 |
-
"learning_rate": 5.495818399044206e-06,
|
| 684 |
-
"loss": 0.5405,
|
| 685 |
-
"step": 2350
|
| 686 |
-
},
|
| 687 |
-
{
|
| 688 |
-
"epoch": 1.532258064516129,
|
| 689 |
-
"grad_norm": 12.705628395080566,
|
| 690 |
-
"learning_rate": 5.436081242532856e-06,
|
| 691 |
-
"loss": 0.5816,
|
| 692 |
-
"step": 2375
|
| 693 |
-
},
|
| 694 |
-
{
|
| 695 |
-
"epoch": 1.5483870967741935,
|
| 696 |
-
"grad_norm": 14.382452964782715,
|
| 697 |
-
"learning_rate": 5.376344086021506e-06,
|
| 698 |
-
"loss": 0.5437,
|
| 699 |
-
"step": 2400
|
| 700 |
-
},
|
| 701 |
-
{
|
| 702 |
-
"epoch": 1.564516129032258,
|
| 703 |
-
"grad_norm": 10.80752944946289,
|
| 704 |
-
"learning_rate": 5.316606929510155e-06,
|
| 705 |
-
"loss": 0.5975,
|
| 706 |
-
"step": 2425
|
| 707 |
-
},
|
| 708 |
-
{
|
| 709 |
-
"epoch": 1.5806451612903225,
|
| 710 |
-
"grad_norm": 12.146509170532227,
|
| 711 |
-
"learning_rate": 5.2568697729988065e-06,
|
| 712 |
-
"loss": 0.599,
|
| 713 |
-
"step": 2450
|
| 714 |
-
},
|
| 715 |
-
{
|
| 716 |
-
"epoch": 1.596774193548387,
|
| 717 |
-
"grad_norm": 12.145088195800781,
|
| 718 |
-
"learning_rate": 5.197132616487456e-06,
|
| 719 |
-
"loss": 0.6506,
|
| 720 |
-
"step": 2475
|
| 721 |
-
},
|
| 722 |
-
{
|
| 723 |
-
"epoch": 1.6129032258064515,
|
| 724 |
-
"grad_norm": 13.103174209594727,
|
| 725 |
-
"learning_rate": 5.137395459976105e-06,
|
| 726 |
-
"loss": 0.5649,
|
| 727 |
-
"step": 2500
|
| 728 |
-
},
|
| 729 |
-
{
|
| 730 |
-
"epoch": 1.629032258064516,
|
| 731 |
-
"grad_norm": 13.602423667907715,
|
| 732 |
-
"learning_rate": 5.077658303464756e-06,
|
| 733 |
-
"loss": 0.5424,
|
| 734 |
-
"step": 2525
|
| 735 |
-
},
|
| 736 |
-
{
|
| 737 |
-
"epoch": 1.6451612903225805,
|
| 738 |
-
"grad_norm": 14.787790298461914,
|
| 739 |
-
"learning_rate": 5.017921146953405e-06,
|
| 740 |
-
"loss": 0.5628,
|
| 741 |
-
"step": 2550
|
| 742 |
-
},
|
| 743 |
-
{
|
| 744 |
-
"epoch": 1.661290322580645,
|
| 745 |
-
"grad_norm": 11.559283256530762,
|
| 746 |
-
"learning_rate": 4.9581839904420555e-06,
|
| 747 |
-
"loss": 0.6216,
|
| 748 |
-
"step": 2575
|
| 749 |
-
},
|
| 750 |
-
{
|
| 751 |
-
"epoch": 1.6774193548387095,
|
| 752 |
-
"grad_norm": 13.20376968383789,
|
| 753 |
-
"learning_rate": 4.898446833930705e-06,
|
| 754 |
-
"loss": 0.5694,
|
| 755 |
-
"step": 2600
|
| 756 |
-
},
|
| 757 |
-
{
|
| 758 |
-
"epoch": 1.6935483870967742,
|
| 759 |
-
"grad_norm": 9.632781982421875,
|
| 760 |
-
"learning_rate": 4.838709677419355e-06,
|
| 761 |
-
"loss": 0.5808,
|
| 762 |
-
"step": 2625
|
| 763 |
-
},
|
| 764 |
-
{
|
| 765 |
-
"epoch": 1.7096774193548387,
|
| 766 |
-
"grad_norm": 12.304398536682129,
|
| 767 |
-
"learning_rate": 4.7789725209080055e-06,
|
| 768 |
-
"loss": 0.5777,
|
| 769 |
-
"step": 2650
|
| 770 |
-
},
|
| 771 |
-
{
|
| 772 |
-
"epoch": 1.7258064516129032,
|
| 773 |
-
"grad_norm": 11.025238990783691,
|
| 774 |
-
"learning_rate": 4.719235364396655e-06,
|
| 775 |
-
"loss": 0.5964,
|
| 776 |
-
"step": 2675
|
| 777 |
-
},
|
| 778 |
-
{
|
| 779 |
-
"epoch": 1.7419354838709677,
|
| 780 |
-
"grad_norm": 13.640275955200195,
|
| 781 |
-
"learning_rate": 4.659498207885305e-06,
|
| 782 |
-
"loss": 0.5936,
|
| 783 |
-
"step": 2700
|
| 784 |
-
},
|
| 785 |
-
{
|
| 786 |
-
"epoch": 1.7580645161290323,
|
| 787 |
-
"grad_norm": 14.28750991821289,
|
| 788 |
-
"learning_rate": 4.599761051373955e-06,
|
| 789 |
-
"loss": 0.5814,
|
| 790 |
-
"step": 2725
|
| 791 |
-
},
|
| 792 |
-
{
|
| 793 |
-
"epoch": 1.7741935483870968,
|
| 794 |
-
"grad_norm": 14.228248596191406,
|
| 795 |
-
"learning_rate": 4.540023894862605e-06,
|
| 796 |
-
"loss": 0.5881,
|
| 797 |
-
"step": 2750
|
| 798 |
-
},
|
| 799 |
-
{
|
| 800 |
-
"epoch": 1.7903225806451613,
|
| 801 |
-
"grad_norm": 12.126937866210938,
|
| 802 |
-
"learning_rate": 4.480286738351255e-06,
|
| 803 |
-
"loss": 0.5568,
|
| 804 |
-
"step": 2775
|
| 805 |
-
},
|
| 806 |
-
{
|
| 807 |
-
"epoch": 1.8064516129032258,
|
| 808 |
-
"grad_norm": 12.653525352478027,
|
| 809 |
-
"learning_rate": 4.420549581839905e-06,
|
| 810 |
-
"loss": 0.5988,
|
| 811 |
-
"step": 2800
|
| 812 |
-
},
|
| 813 |
-
{
|
| 814 |
-
"epoch": 1.8225806451612905,
|
| 815 |
-
"grad_norm": 10.851930618286133,
|
| 816 |
-
"learning_rate": 4.360812425328555e-06,
|
| 817 |
-
"loss": 0.6073,
|
| 818 |
-
"step": 2825
|
| 819 |
-
},
|
| 820 |
-
{
|
| 821 |
-
"epoch": 1.838709677419355,
|
| 822 |
-
"grad_norm": 12.00724983215332,
|
| 823 |
-
"learning_rate": 4.3010752688172045e-06,
|
| 824 |
-
"loss": 0.5739,
|
| 825 |
-
"step": 2850
|
| 826 |
-
},
|
| 827 |
-
{
|
| 828 |
-
"epoch": 1.8548387096774195,
|
| 829 |
-
"grad_norm": 10.997614860534668,
|
| 830 |
-
"learning_rate": 4.241338112305855e-06,
|
| 831 |
-
"loss": 0.5663,
|
| 832 |
-
"step": 2875
|
| 833 |
-
},
|
| 834 |
-
{
|
| 835 |
-
"epoch": 1.870967741935484,
|
| 836 |
-
"grad_norm": 12.384391784667969,
|
| 837 |
-
"learning_rate": 4.181600955794505e-06,
|
| 838 |
-
"loss": 0.5325,
|
| 839 |
-
"step": 2900
|
| 840 |
-
},
|
| 841 |
-
{
|
| 842 |
-
"epoch": 1.8870967741935485,
|
| 843 |
-
"grad_norm": 10.200772285461426,
|
| 844 |
-
"learning_rate": 4.121863799283155e-06,
|
| 845 |
-
"loss": 0.5918,
|
| 846 |
-
"step": 2925
|
| 847 |
-
},
|
| 848 |
-
{
|
| 849 |
-
"epoch": 1.903225806451613,
|
| 850 |
-
"grad_norm": 13.224651336669922,
|
| 851 |
-
"learning_rate": 4.062126642771804e-06,
|
| 852 |
-
"loss": 0.5399,
|
| 853 |
-
"step": 2950
|
| 854 |
-
},
|
| 855 |
-
{
|
| 856 |
-
"epoch": 1.9193548387096775,
|
| 857 |
-
"grad_norm": 10.611023902893066,
|
| 858 |
-
"learning_rate": 4.002389486260454e-06,
|
| 859 |
-
"loss": 0.5593,
|
| 860 |
-
"step": 2975
|
| 861 |
-
},
|
| 862 |
-
{
|
| 863 |
-
"epoch": 1.935483870967742,
|
| 864 |
-
"grad_norm": 10.110644340515137,
|
| 865 |
-
"learning_rate": 3.942652329749105e-06,
|
| 866 |
-
"loss": 0.5611,
|
| 867 |
-
"step": 3000
|
| 868 |
-
},
|
| 869 |
-
{
|
| 870 |
-
"epoch": 1.935483870967742,
|
| 871 |
-
"eval_cer": 51.336001032657805,
|
| 872 |
-
"eval_loss": 0.6702780723571777,
|
| 873 |
-
"eval_runtime": 963.0475,
|
| 874 |
-
"eval_samples_per_second": 2.369,
|
| 875 |
-
"eval_steps_per_second": 0.297,
|
| 876 |
-
"step": 3000
|
| 877 |
-
},
|
| 878 |
-
{
|
| 879 |
-
"epoch": 1.9516129032258065,
|
| 880 |
-
"grad_norm": 10.685456275939941,
|
| 881 |
-
"learning_rate": 3.882915173237754e-06,
|
| 882 |
-
"loss": 0.5326,
|
| 883 |
-
"step": 3025
|
| 884 |
-
},
|
| 885 |
-
{
|
| 886 |
-
"epoch": 1.967741935483871,
|
| 887 |
-
"grad_norm": 14.404354095458984,
|
| 888 |
-
"learning_rate": 3.823178016726404e-06,
|
| 889 |
-
"loss": 0.5828,
|
| 890 |
-
"step": 3050
|
| 891 |
-
},
|
| 892 |
-
{
|
| 893 |
-
"epoch": 1.9838709677419355,
|
| 894 |
-
"grad_norm": 13.997696876525879,
|
| 895 |
-
"learning_rate": 3.763440860215054e-06,
|
| 896 |
-
"loss": 0.5394,
|
| 897 |
-
"step": 3075
|
| 898 |
-
},
|
| 899 |
-
{
|
| 900 |
-
"epoch": 2.0,
|
| 901 |
-
"grad_norm": 16.121444702148438,
|
| 902 |
-
"learning_rate": 3.7037037037037037e-06,
|
| 903 |
-
"loss": 0.5635,
|
| 904 |
-
"step": 3100
|
| 905 |
-
},
|
| 906 |
-
{
|
| 907 |
-
"epoch": 2.0161290322580645,
|
| 908 |
-
"grad_norm": 9.237725257873535,
|
| 909 |
-
"learning_rate": 3.643966547192354e-06,
|
| 910 |
-
"loss": 0.3737,
|
| 911 |
-
"step": 3125
|
| 912 |
-
},
|
| 913 |
-
{
|
| 914 |
-
"epoch": 2.032258064516129,
|
| 915 |
-
"grad_norm": 11.313372611999512,
|
| 916 |
-
"learning_rate": 3.584229390681004e-06,
|
| 917 |
-
"loss": 0.3934,
|
| 918 |
-
"step": 3150
|
| 919 |
-
},
|
| 920 |
-
{
|
| 921 |
-
"epoch": 2.0483870967741935,
|
| 922 |
-
"grad_norm": 9.819090843200684,
|
| 923 |
-
"learning_rate": 3.5244922341696534e-06,
|
| 924 |
-
"loss": 0.3494,
|
| 925 |
-
"step": 3175
|
| 926 |
-
},
|
| 927 |
-
{
|
| 928 |
-
"epoch": 2.064516129032258,
|
| 929 |
-
"grad_norm": 9.302324295043945,
|
| 930 |
-
"learning_rate": 3.4647550776583037e-06,
|
| 931 |
-
"loss": 0.3691,
|
| 932 |
-
"step": 3200
|
| 933 |
-
},
|
| 934 |
-
{
|
| 935 |
-
"epoch": 2.0806451612903225,
|
| 936 |
-
"grad_norm": 11.517475128173828,
|
| 937 |
-
"learning_rate": 3.4050179211469536e-06,
|
| 938 |
-
"loss": 0.3652,
|
| 939 |
-
"step": 3225
|
| 940 |
-
},
|
| 941 |
-
{
|
| 942 |
-
"epoch": 2.096774193548387,
|
| 943 |
-
"grad_norm": 7.707530975341797,
|
| 944 |
-
"learning_rate": 3.3452807646356034e-06,
|
| 945 |
-
"loss": 0.3566,
|
| 946 |
-
"step": 3250
|
| 947 |
-
},
|
| 948 |
-
{
|
| 949 |
-
"epoch": 2.1129032258064515,
|
| 950 |
-
"grad_norm": 9.121161460876465,
|
| 951 |
-
"learning_rate": 3.2855436081242537e-06,
|
| 952 |
-
"loss": 0.3409,
|
| 953 |
-
"step": 3275
|
| 954 |
-
},
|
| 955 |
-
{
|
| 956 |
-
"epoch": 2.129032258064516,
|
| 957 |
-
"grad_norm": 10.464853286743164,
|
| 958 |
-
"learning_rate": 3.225806451612903e-06,
|
| 959 |
-
"loss": 0.33,
|
| 960 |
-
"step": 3300
|
| 961 |
-
},
|
| 962 |
-
{
|
| 963 |
-
"epoch": 2.1451612903225805,
|
| 964 |
-
"grad_norm": 8.300515174865723,
|
| 965 |
-
"learning_rate": 3.1660692951015535e-06,
|
| 966 |
-
"loss": 0.3436,
|
| 967 |
-
"step": 3325
|
| 968 |
-
},
|
| 969 |
-
{
|
| 970 |
-
"epoch": 2.161290322580645,
|
| 971 |
-
"grad_norm": 7.577033519744873,
|
| 972 |
-
"learning_rate": 3.1063321385902034e-06,
|
| 973 |
-
"loss": 0.3441,
|
| 974 |
-
"step": 3350
|
| 975 |
-
},
|
| 976 |
-
{
|
| 977 |
-
"epoch": 2.1774193548387095,
|
| 978 |
-
"grad_norm": 12.314337730407715,
|
| 979 |
-
"learning_rate": 3.0465949820788532e-06,
|
| 980 |
-
"loss": 0.387,
|
| 981 |
-
"step": 3375
|
| 982 |
-
},
|
| 983 |
-
{
|
| 984 |
-
"epoch": 2.193548387096774,
|
| 985 |
-
"grad_norm": 8.03864860534668,
|
| 986 |
-
"learning_rate": 2.9868578255675035e-06,
|
| 987 |
-
"loss": 0.3533,
|
| 988 |
-
"step": 3400
|
| 989 |
-
},
|
| 990 |
-
{
|
| 991 |
-
"epoch": 2.2096774193548385,
|
| 992 |
-
"grad_norm": 10.326530456542969,
|
| 993 |
-
"learning_rate": 2.9271206690561534e-06,
|
| 994 |
-
"loss": 0.351,
|
| 995 |
-
"step": 3425
|
| 996 |
-
},
|
| 997 |
-
{
|
| 998 |
-
"epoch": 2.225806451612903,
|
| 999 |
-
"grad_norm": 8.268649101257324,
|
| 1000 |
-
"learning_rate": 2.867383512544803e-06,
|
| 1001 |
-
"loss": 0.3437,
|
| 1002 |
-
"step": 3450
|
| 1003 |
-
},
|
| 1004 |
-
{
|
| 1005 |
-
"epoch": 2.241935483870968,
|
| 1006 |
-
"grad_norm": 9.62258529663086,
|
| 1007 |
-
"learning_rate": 2.807646356033453e-06,
|
| 1008 |
-
"loss": 0.3254,
|
| 1009 |
-
"step": 3475
|
| 1010 |
-
},
|
| 1011 |
-
{
|
| 1012 |
-
"epoch": 2.258064516129032,
|
| 1013 |
-
"grad_norm": 8.58535099029541,
|
| 1014 |
-
"learning_rate": 2.747909199522103e-06,
|
| 1015 |
-
"loss": 0.3592,
|
| 1016 |
-
"step": 3500
|
| 1017 |
-
},
|
| 1018 |
-
{
|
| 1019 |
-
"epoch": 2.274193548387097,
|
| 1020 |
-
"grad_norm": 10.211243629455566,
|
| 1021 |
-
"learning_rate": 2.688172043010753e-06,
|
| 1022 |
-
"loss": 0.3334,
|
| 1023 |
-
"step": 3525
|
| 1024 |
-
},
|
| 1025 |
-
{
|
| 1026 |
-
"epoch": 2.2903225806451615,
|
| 1027 |
-
"grad_norm": 9.174546241760254,
|
| 1028 |
-
"learning_rate": 2.6284348864994032e-06,
|
| 1029 |
-
"loss": 0.3533,
|
| 1030 |
-
"step": 3550
|
| 1031 |
-
},
|
| 1032 |
-
{
|
| 1033 |
-
"epoch": 2.306451612903226,
|
| 1034 |
-
"grad_norm": 9.889862060546875,
|
| 1035 |
-
"learning_rate": 2.5686977299880527e-06,
|
| 1036 |
-
"loss": 0.3263,
|
| 1037 |
-
"step": 3575
|
| 1038 |
-
},
|
| 1039 |
-
{
|
| 1040 |
-
"epoch": 2.3225806451612905,
|
| 1041 |
-
"grad_norm": 10.23873519897461,
|
| 1042 |
-
"learning_rate": 2.5089605734767026e-06,
|
| 1043 |
-
"loss": 0.3601,
|
| 1044 |
-
"step": 3600
|
| 1045 |
-
},
|
| 1046 |
-
{
|
| 1047 |
-
"epoch": 2.338709677419355,
|
| 1048 |
-
"grad_norm": 8.46229076385498,
|
| 1049 |
-
"learning_rate": 2.4492234169653525e-06,
|
| 1050 |
-
"loss": 0.335,
|
| 1051 |
-
"step": 3625
|
| 1052 |
-
},
|
| 1053 |
-
{
|
| 1054 |
-
"epoch": 2.3548387096774195,
|
| 1055 |
-
"grad_norm": 8.364771842956543,
|
| 1056 |
-
"learning_rate": 2.3894862604540028e-06,
|
| 1057 |
-
"loss": 0.3447,
|
| 1058 |
-
"step": 3650
|
| 1059 |
-
},
|
| 1060 |
-
{
|
| 1061 |
-
"epoch": 2.370967741935484,
|
| 1062 |
-
"grad_norm": 11.249506950378418,
|
| 1063 |
-
"learning_rate": 2.3297491039426526e-06,
|
| 1064 |
-
"loss": 0.3544,
|
| 1065 |
-
"step": 3675
|
| 1066 |
-
},
|
| 1067 |
-
{
|
| 1068 |
-
"epoch": 2.3870967741935485,
|
| 1069 |
-
"grad_norm": 8.8016996383667,
|
| 1070 |
-
"learning_rate": 2.2700119474313025e-06,
|
| 1071 |
-
"loss": 0.3277,
|
| 1072 |
-
"step": 3700
|
| 1073 |
-
},
|
| 1074 |
-
{
|
| 1075 |
-
"epoch": 2.403225806451613,
|
| 1076 |
-
"grad_norm": 9.774581909179688,
|
| 1077 |
-
"learning_rate": 2.2102747909199524e-06,
|
| 1078 |
-
"loss": 0.3346,
|
| 1079 |
-
"step": 3725
|
| 1080 |
-
},
|
| 1081 |
-
{
|
| 1082 |
-
"epoch": 2.4193548387096775,
|
| 1083 |
-
"grad_norm": 8.027830123901367,
|
| 1084 |
-
"learning_rate": 2.1505376344086023e-06,
|
| 1085 |
-
"loss": 0.3291,
|
| 1086 |
-
"step": 3750
|
| 1087 |
-
},
|
| 1088 |
-
{
|
| 1089 |
-
"epoch": 2.435483870967742,
|
| 1090 |
-
"grad_norm": 10.107059478759766,
|
| 1091 |
-
"learning_rate": 2.0908004778972526e-06,
|
| 1092 |
-
"loss": 0.3366,
|
| 1093 |
-
"step": 3775
|
| 1094 |
-
},
|
| 1095 |
-
{
|
| 1096 |
-
"epoch": 2.4516129032258065,
|
| 1097 |
-
"grad_norm": 8.280789375305176,
|
| 1098 |
-
"learning_rate": 2.031063321385902e-06,
|
| 1099 |
-
"loss": 0.3473,
|
| 1100 |
-
"step": 3800
|
| 1101 |
-
},
|
| 1102 |
-
{
|
| 1103 |
-
"epoch": 2.467741935483871,
|
| 1104 |
-
"grad_norm": 9.160382270812988,
|
| 1105 |
-
"learning_rate": 1.9713261648745523e-06,
|
| 1106 |
-
"loss": 0.3215,
|
| 1107 |
-
"step": 3825
|
| 1108 |
-
},
|
| 1109 |
-
{
|
| 1110 |
-
"epoch": 2.4838709677419355,
|
| 1111 |
-
"grad_norm": 7.922098636627197,
|
| 1112 |
-
"learning_rate": 1.911589008363202e-06,
|
| 1113 |
-
"loss": 0.3285,
|
| 1114 |
-
"step": 3850
|
| 1115 |
-
},
|
| 1116 |
-
{
|
| 1117 |
-
"epoch": 2.5,
|
| 1118 |
-
"grad_norm": 9.239423751831055,
|
| 1119 |
-
"learning_rate": 1.8518518518518519e-06,
|
| 1120 |
-
"loss": 0.3608,
|
| 1121 |
-
"step": 3875
|
| 1122 |
-
},
|
| 1123 |
-
{
|
| 1124 |
-
"epoch": 2.5161290322580645,
|
| 1125 |
-
"grad_norm": 8.667262077331543,
|
| 1126 |
-
"learning_rate": 1.792114695340502e-06,
|
| 1127 |
-
"loss": 0.3389,
|
| 1128 |
-
"step": 3900
|
| 1129 |
-
},
|
| 1130 |
-
{
|
| 1131 |
-
"epoch": 2.532258064516129,
|
| 1132 |
-
"grad_norm": 10.475480079650879,
|
| 1133 |
-
"learning_rate": 1.7323775388291518e-06,
|
| 1134 |
-
"loss": 0.3226,
|
| 1135 |
-
"step": 3925
|
| 1136 |
-
},
|
| 1137 |
-
{
|
| 1138 |
-
"epoch": 2.5483870967741935,
|
| 1139 |
-
"grad_norm": 11.079362869262695,
|
| 1140 |
-
"learning_rate": 1.6726403823178017e-06,
|
| 1141 |
-
"loss": 0.3559,
|
| 1142 |
-
"step": 3950
|
| 1143 |
-
},
|
| 1144 |
-
{
|
| 1145 |
-
"epoch": 2.564516129032258,
|
| 1146 |
-
"grad_norm": 10.680990219116211,
|
| 1147 |
-
"learning_rate": 1.6129032258064516e-06,
|
| 1148 |
-
"loss": 0.297,
|
| 1149 |
-
"step": 3975
|
| 1150 |
-
},
|
| 1151 |
-
{
|
| 1152 |
-
"epoch": 2.5806451612903225,
|
| 1153 |
-
"grad_norm": 13.101299285888672,
|
| 1154 |
-
"learning_rate": 1.5531660692951017e-06,
|
| 1155 |
-
"loss": 0.3359,
|
| 1156 |
-
"step": 4000
|
| 1157 |
-
},
|
| 1158 |
-
{
|
| 1159 |
-
"epoch": 2.5806451612903225,
|
| 1160 |
-
"eval_cer": 48.63818252226668,
|
| 1161 |
-
"eval_loss": 0.6473900675773621,
|
| 1162 |
-
"eval_runtime": 969.6918,
|
| 1163 |
-
"eval_samples_per_second": 2.352,
|
| 1164 |
-
"eval_steps_per_second": 0.295,
|
| 1165 |
-
"step": 4000
|
| 1166 |
}
|
| 1167 |
],
|
| 1168 |
"logging_steps": 25,
|
| 1169 |
-
"max_steps":
|
| 1170 |
"num_input_tokens_seen": 0,
|
| 1171 |
-
"num_train_epochs":
|
| 1172 |
"save_steps": 1000,
|
| 1173 |
-
"total_flos":
|
| 1174 |
-
"train_batch_size":
|
| 1175 |
"trial_name": null,
|
| 1176 |
"trial_params": null
|
| 1177 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 61.346116219917825,
|
| 3 |
+
"best_model_checkpoint": "./whisper-small-taiwanese/checkpoint-1000",
|
| 4 |
+
"epoch": 0.40024014408645187,
|
| 5 |
"eval_steps": 1000,
|
| 6 |
+
"global_step": 1000,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
+
"epoch": 0.010006003602161296,
|
| 13 |
+
"grad_norm": 197.02195739746094,
|
| 14 |
+
"learning_rate": 5.000000000000001e-07,
|
| 15 |
+
"loss": 7.3864,
|
| 16 |
"step": 25
|
| 17 |
},
|
| 18 |
{
|
| 19 |
+
"epoch": 0.020012007204322592,
|
| 20 |
+
"grad_norm": 41.90484619140625,
|
| 21 |
+
"learning_rate": 1.0000000000000002e-06,
|
| 22 |
+
"loss": 5.5537,
|
| 23 |
"step": 50
|
| 24 |
},
|
| 25 |
{
|
| 26 |
+
"epoch": 0.03001801080648389,
|
| 27 |
+
"grad_norm": 32.08101272583008,
|
| 28 |
+
"learning_rate": 1.5e-06,
|
| 29 |
+
"loss": 3.7704,
|
| 30 |
"step": 75
|
| 31 |
},
|
| 32 |
{
|
| 33 |
+
"epoch": 0.040024014408645184,
|
| 34 |
+
"grad_norm": 28.216585159301758,
|
| 35 |
+
"learning_rate": 2.0000000000000003e-06,
|
| 36 |
+
"loss": 2.6296,
|
| 37 |
"step": 100
|
| 38 |
},
|
| 39 |
{
|
| 40 |
+
"epoch": 0.05003001801080648,
|
| 41 |
+
"grad_norm": 23.683828353881836,
|
| 42 |
+
"learning_rate": 2.5e-06,
|
| 43 |
+
"loss": 2.4474,
|
| 44 |
"step": 125
|
| 45 |
},
|
| 46 |
{
|
| 47 |
+
"epoch": 0.06003602161296778,
|
| 48 |
+
"grad_norm": 21.883520126342773,
|
| 49 |
+
"learning_rate": 3e-06,
|
| 50 |
+
"loss": 2.2054,
|
| 51 |
"step": 150
|
| 52 |
},
|
| 53 |
{
|
| 54 |
+
"epoch": 0.07004202521512908,
|
| 55 |
+
"grad_norm": 24.772098541259766,
|
| 56 |
+
"learning_rate": 3.5e-06,
|
| 57 |
+
"loss": 2.0695,
|
| 58 |
"step": 175
|
| 59 |
},
|
| 60 |
{
|
| 61 |
+
"epoch": 0.08004802881729037,
|
| 62 |
+
"grad_norm": 24.105548858642578,
|
| 63 |
+
"learning_rate": 4.000000000000001e-06,
|
| 64 |
+
"loss": 1.9116,
|
| 65 |
"step": 200
|
| 66 |
},
|
| 67 |
{
|
| 68 |
+
"epoch": 0.09005403241945167,
|
| 69 |
+
"grad_norm": 18.805519104003906,
|
| 70 |
+
"learning_rate": 4.5e-06,
|
| 71 |
+
"loss": 1.7643,
|
| 72 |
"step": 225
|
| 73 |
},
|
| 74 |
{
|
| 75 |
+
"epoch": 0.10006003602161297,
|
| 76 |
+
"grad_norm": 15.599541664123535,
|
| 77 |
+
"learning_rate": 5e-06,
|
| 78 |
+
"loss": 1.6394,
|
| 79 |
"step": 250
|
| 80 |
},
|
| 81 |
{
|
| 82 |
+
"epoch": 0.11006603962377426,
|
| 83 |
+
"grad_norm": 15.514196395874023,
|
| 84 |
+
"learning_rate": 5.500000000000001e-06,
|
| 85 |
+
"loss": 1.6016,
|
| 86 |
"step": 275
|
| 87 |
},
|
| 88 |
{
|
| 89 |
+
"epoch": 0.12007204322593557,
|
| 90 |
+
"grad_norm": 15.5431547164917,
|
| 91 |
+
"learning_rate": 6e-06,
|
| 92 |
+
"loss": 1.5851,
|
| 93 |
"step": 300
|
| 94 |
},
|
| 95 |
{
|
| 96 |
+
"epoch": 0.13007804682809687,
|
| 97 |
+
"grad_norm": 16.450502395629883,
|
| 98 |
+
"learning_rate": 6.5000000000000004e-06,
|
| 99 |
+
"loss": 1.5076,
|
| 100 |
"step": 325
|
| 101 |
},
|
| 102 |
{
|
| 103 |
+
"epoch": 0.14008405043025815,
|
| 104 |
+
"grad_norm": 16.393997192382812,
|
| 105 |
+
"learning_rate": 7e-06,
|
| 106 |
+
"loss": 1.5487,
|
| 107 |
"step": 350
|
| 108 |
},
|
| 109 |
{
|
| 110 |
+
"epoch": 0.15009005403241946,
|
| 111 |
+
"grad_norm": 14.165709495544434,
|
| 112 |
+
"learning_rate": 7.500000000000001e-06,
|
| 113 |
+
"loss": 1.5365,
|
| 114 |
"step": 375
|
| 115 |
},
|
| 116 |
{
|
| 117 |
+
"epoch": 0.16009605763458074,
|
| 118 |
+
"grad_norm": 15.929381370544434,
|
| 119 |
+
"learning_rate": 8.000000000000001e-06,
|
| 120 |
+
"loss": 1.5023,
|
| 121 |
"step": 400
|
| 122 |
},
|
| 123 |
{
|
| 124 |
+
"epoch": 0.17010206123674204,
|
| 125 |
+
"grad_norm": 14.422001838684082,
|
| 126 |
+
"learning_rate": 8.5e-06,
|
| 127 |
+
"loss": 1.3558,
|
| 128 |
"step": 425
|
| 129 |
},
|
| 130 |
{
|
| 131 |
+
"epoch": 0.18010806483890335,
|
| 132 |
+
"grad_norm": 13.510339736938477,
|
| 133 |
+
"learning_rate": 9e-06,
|
| 134 |
+
"loss": 1.3898,
|
| 135 |
"step": 450
|
| 136 |
},
|
| 137 |
{
|
| 138 |
+
"epoch": 0.19011406844106463,
|
| 139 |
+
"grad_norm": 14.485660552978516,
|
| 140 |
+
"learning_rate": 9.5e-06,
|
| 141 |
+
"loss": 1.4279,
|
| 142 |
"step": 475
|
| 143 |
},
|
| 144 |
{
|
| 145 |
+
"epoch": 0.20012007204322593,
|
| 146 |
+
"grad_norm": 14.117327690124512,
|
| 147 |
+
"learning_rate": 1e-05,
|
| 148 |
+
"loss": 1.3455,
|
| 149 |
"step": 500
|
| 150 |
},
|
| 151 |
{
|
| 152 |
+
"epoch": 0.21012607564538724,
|
| 153 |
+
"grad_norm": 16.12464714050293,
|
| 154 |
+
"learning_rate": 9.944395017793596e-06,
|
| 155 |
+
"loss": 1.4,
|
| 156 |
"step": 525
|
| 157 |
},
|
| 158 |
{
|
| 159 |
+
"epoch": 0.22013207924754852,
|
| 160 |
+
"grad_norm": 15.304022789001465,
|
| 161 |
+
"learning_rate": 9.888790035587188e-06,
|
| 162 |
+
"loss": 1.4159,
|
| 163 |
"step": 550
|
| 164 |
},
|
| 165 |
{
|
| 166 |
+
"epoch": 0.23013808284970982,
|
| 167 |
+
"grad_norm": 14.668664932250977,
|
| 168 |
+
"learning_rate": 9.833185053380784e-06,
|
| 169 |
+
"loss": 1.3445,
|
| 170 |
"step": 575
|
| 171 |
},
|
| 172 |
{
|
| 173 |
+
"epoch": 0.24014408645187113,
|
| 174 |
+
"grad_norm": 13.041420936584473,
|
| 175 |
+
"learning_rate": 9.777580071174379e-06,
|
| 176 |
+
"loss": 1.3622,
|
| 177 |
"step": 600
|
| 178 |
},
|
| 179 |
{
|
| 180 |
+
"epoch": 0.25015009005403244,
|
| 181 |
+
"grad_norm": 15.908055305480957,
|
| 182 |
+
"learning_rate": 9.721975088967973e-06,
|
| 183 |
+
"loss": 1.3234,
|
| 184 |
"step": 625
|
| 185 |
},
|
| 186 |
{
|
| 187 |
+
"epoch": 0.26015609365619374,
|
| 188 |
+
"grad_norm": 13.73078727722168,
|
| 189 |
+
"learning_rate": 9.666370106761567e-06,
|
| 190 |
+
"loss": 1.2332,
|
| 191 |
"step": 650
|
| 192 |
},
|
| 193 |
{
|
| 194 |
+
"epoch": 0.270162097258355,
|
| 195 |
+
"grad_norm": 14.327301979064941,
|
| 196 |
+
"learning_rate": 9.610765124555162e-06,
|
| 197 |
+
"loss": 1.3042,
|
| 198 |
"step": 675
|
| 199 |
},
|
| 200 |
{
|
| 201 |
+
"epoch": 0.2801681008605163,
|
| 202 |
+
"grad_norm": 14.390907287597656,
|
| 203 |
+
"learning_rate": 9.555160142348756e-06,
|
| 204 |
+
"loss": 1.3216,
|
| 205 |
"step": 700
|
| 206 |
},
|
| 207 |
{
|
| 208 |
+
"epoch": 0.2901741044626776,
|
| 209 |
+
"grad_norm": 13.917515754699707,
|
| 210 |
+
"learning_rate": 9.49955516014235e-06,
|
| 211 |
+
"loss": 1.2931,
|
| 212 |
"step": 725
|
| 213 |
},
|
| 214 |
{
|
| 215 |
+
"epoch": 0.3001801080648389,
|
| 216 |
+
"grad_norm": 15.108023643493652,
|
| 217 |
+
"learning_rate": 9.443950177935945e-06,
|
| 218 |
+
"loss": 1.3286,
|
| 219 |
"step": 750
|
| 220 |
},
|
| 221 |
{
|
| 222 |
+
"epoch": 0.3101861116670002,
|
| 223 |
+
"grad_norm": 13.692678451538086,
|
| 224 |
+
"learning_rate": 9.388345195729539e-06,
|
| 225 |
+
"loss": 1.3057,
|
| 226 |
"step": 775
|
| 227 |
},
|
| 228 |
{
|
| 229 |
+
"epoch": 0.32019211526916147,
|
| 230 |
+
"grad_norm": 13.685354232788086,
|
| 231 |
+
"learning_rate": 9.332740213523132e-06,
|
| 232 |
+
"loss": 1.2402,
|
| 233 |
"step": 800
|
| 234 |
},
|
| 235 |
{
|
| 236 |
+
"epoch": 0.3301981188713228,
|
| 237 |
+
"grad_norm": 14.591761589050293,
|
| 238 |
+
"learning_rate": 9.277135231316726e-06,
|
| 239 |
+
"loss": 1.2688,
|
| 240 |
"step": 825
|
| 241 |
},
|
| 242 |
{
|
| 243 |
+
"epoch": 0.3402041224734841,
|
| 244 |
+
"grad_norm": 15.677751541137695,
|
| 245 |
+
"learning_rate": 9.221530249110321e-06,
|
| 246 |
+
"loss": 1.3076,
|
| 247 |
"step": 850
|
| 248 |
},
|
| 249 |
{
|
| 250 |
+
"epoch": 0.3502101260756454,
|
| 251 |
+
"grad_norm": 15.109577178955078,
|
| 252 |
+
"learning_rate": 9.165925266903915e-06,
|
| 253 |
+
"loss": 1.2141,
|
| 254 |
"step": 875
|
| 255 |
},
|
| 256 |
{
|
| 257 |
+
"epoch": 0.3602161296778067,
|
| 258 |
+
"grad_norm": 10.552845001220703,
|
| 259 |
+
"learning_rate": 9.110320284697509e-06,
|
| 260 |
+
"loss": 1.2393,
|
| 261 |
"step": 900
|
| 262 |
},
|
| 263 |
{
|
| 264 |
+
"epoch": 0.370222133279968,
|
| 265 |
+
"grad_norm": 12.321894645690918,
|
| 266 |
+
"learning_rate": 9.054715302491104e-06,
|
| 267 |
+
"loss": 1.2417,
|
| 268 |
"step": 925
|
| 269 |
},
|
| 270 |
{
|
| 271 |
+
"epoch": 0.38022813688212925,
|
| 272 |
+
"grad_norm": 13.729790687561035,
|
| 273 |
+
"learning_rate": 8.999110320284698e-06,
|
| 274 |
+
"loss": 1.2082,
|
| 275 |
"step": 950
|
| 276 |
},
|
| 277 |
{
|
| 278 |
+
"epoch": 0.39023414048429056,
|
| 279 |
+
"grad_norm": 13.137016296386719,
|
| 280 |
+
"learning_rate": 8.943505338078292e-06,
|
| 281 |
+
"loss": 1.2048,
|
| 282 |
"step": 975
|
| 283 |
},
|
| 284 |
{
|
| 285 |
+
"epoch": 0.40024014408645187,
|
| 286 |
+
"grad_norm": 12.194613456726074,
|
| 287 |
+
"learning_rate": 8.887900355871887e-06,
|
| 288 |
+
"loss": 1.2739,
|
| 289 |
"step": 1000
|
| 290 |
},
|
| 291 |
{
|
| 292 |
+
"epoch": 0.40024014408645187,
|
| 293 |
+
"eval_cer": 61.346116219917825,
|
| 294 |
+
"eval_loss": 1.169872522354126,
|
| 295 |
+
"eval_runtime": 1744.6409,
|
| 296 |
+
"eval_samples_per_second": 2.274,
|
| 297 |
+
"eval_steps_per_second": 0.284,
|
| 298 |
"step": 1000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 299 |
}
|
| 300 |
],
|
| 301 |
"logging_steps": 25,
|
| 302 |
+
"max_steps": 4996,
|
| 303 |
"num_input_tokens_seen": 0,
|
| 304 |
+
"num_train_epochs": 2,
|
| 305 |
"save_steps": 1000,
|
| 306 |
+
"total_flos": 4.61736640512e+18,
|
| 307 |
+
"train_batch_size": 8,
|
| 308 |
"trial_name": null,
|
| 309 |
"trial_params": null
|
| 310 |
}
|
last-checkpoint/training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 5176
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:dc814e8346759fc832d3d40fb2efbd9f7f5bf91489499603abb8463206368d6b
|
| 3 |
size 5176
|