Training in progress, epoch 1, checkpoint
Browse files
last-checkpoint/config.json
CHANGED
|
@@ -1,5 +1,5 @@
|
|
| 1 |
{
|
| 2 |
-
"_name_or_path": "
|
| 3 |
"activation_dropout": 0.0,
|
| 4 |
"activation_function": "gelu",
|
| 5 |
"apply_spec_augment": false,
|
|
|
|
| 1 |
{
|
| 2 |
+
"_name_or_path": "Baselhany/Graduation_Project_Whisper_tiny_fine_tune_Quran",
|
| 3 |
"activation_dropout": 0.0,
|
| 4 |
"activation_function": "gelu",
|
| 5 |
"apply_spec_augment": false,
|
last-checkpoint/model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 151061672
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b46e8196e00199074adc216fecc9513e28674d93785a33ee77eabc1cd8a07cd9
|
| 3 |
size 151061672
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 297616186
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0dad509a51d888adea1ba97f4ed317aab1316125f8f2d85181dbff53f6e12bb5
|
| 3 |
size 297616186
|
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:e5a0053688c21d4b223e60fe14000ea3131b6f050b39ef3f20662920444f38ad
|
| 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 sha256:959811f2b632ac28cb0d8d27e9678f9580865cbedd9c0e82696e795f9bb8c37e
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,646 +1,43 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric":
|
| 3 |
-
"best_model_checkpoint": "./Whisper_tiny_fine_tune_Quran/checkpoint-
|
| 4 |
-
"epoch":
|
| 5 |
"eval_steps": 500,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
"epoch": 0.32,
|
| 13 |
-
"grad_norm":
|
| 14 |
"learning_rate": 1e-05,
|
| 15 |
-
"loss":
|
| 16 |
"step": 50
|
| 17 |
},
|
| 18 |
{
|
| 19 |
"epoch": 0.64,
|
| 20 |
-
"grad_norm":
|
| 21 |
"learning_rate": 2e-05,
|
| 22 |
-
"loss": 0.
|
| 23 |
"step": 100
|
| 24 |
},
|
| 25 |
{
|
| 26 |
"epoch": 0.96,
|
| 27 |
-
"grad_norm":
|
| 28 |
"learning_rate": 3e-05,
|
| 29 |
-
"loss": 0.
|
| 30 |
"step": 150
|
| 31 |
},
|
| 32 |
{
|
| 33 |
"epoch": 1.0,
|
| 34 |
-
"eval_cer":
|
| 35 |
-
"eval_loss": 0.
|
| 36 |
-
"eval_runtime":
|
| 37 |
-
"eval_samples_per_second":
|
| 38 |
-
"eval_steps_per_second": 0.
|
| 39 |
-
"eval_wer":
|
| 40 |
"step": 157
|
| 41 |
-
},
|
| 42 |
-
{
|
| 43 |
-
"epoch": 1.2752,
|
| 44 |
-
"grad_norm": 183927.1875,
|
| 45 |
-
"learning_rate": 4e-05,
|
| 46 |
-
"loss": 0.0662,
|
| 47 |
-
"step": 200
|
| 48 |
-
},
|
| 49 |
-
{
|
| 50 |
-
"epoch": 1.5952,
|
| 51 |
-
"grad_norm": 234198.296875,
|
| 52 |
-
"learning_rate": 5e-05,
|
| 53 |
-
"loss": 0.0577,
|
| 54 |
-
"step": 250
|
| 55 |
-
},
|
| 56 |
-
{
|
| 57 |
-
"epoch": 1.9152,
|
| 58 |
-
"grad_norm": 209207.671875,
|
| 59 |
-
"learning_rate": 6e-05,
|
| 60 |
-
"loss": 0.0559,
|
| 61 |
-
"step": 300
|
| 62 |
-
},
|
| 63 |
-
{
|
| 64 |
-
"epoch": 2.0,
|
| 65 |
-
"eval_cer": 56.4423610340619,
|
| 66 |
-
"eval_loss": 0.037598446011543274,
|
| 67 |
-
"eval_runtime": 691.9923,
|
| 68 |
-
"eval_samples_per_second": 2.889,
|
| 69 |
-
"eval_steps_per_second": 0.181,
|
| 70 |
-
"eval_wer": 116.80792305494714,
|
| 71 |
-
"step": 314
|
| 72 |
-
},
|
| 73 |
-
{
|
| 74 |
-
"epoch": 2.2304,
|
| 75 |
-
"grad_norm": 127075.0234375,
|
| 76 |
-
"learning_rate": 7e-05,
|
| 77 |
-
"loss": 0.0456,
|
| 78 |
-
"step": 350
|
| 79 |
-
},
|
| 80 |
-
{
|
| 81 |
-
"epoch": 2.5504,
|
| 82 |
-
"grad_norm": 112266.140625,
|
| 83 |
-
"learning_rate": 8e-05,
|
| 84 |
-
"loss": 0.0389,
|
| 85 |
-
"step": 400
|
| 86 |
-
},
|
| 87 |
-
{
|
| 88 |
-
"epoch": 2.8704,
|
| 89 |
-
"grad_norm": 123542.2421875,
|
| 90 |
-
"learning_rate": 9e-05,
|
| 91 |
-
"loss": 0.0399,
|
| 92 |
-
"step": 450
|
| 93 |
-
},
|
| 94 |
-
{
|
| 95 |
-
"epoch": 3.0,
|
| 96 |
-
"eval_cer": 29.94396982136913,
|
| 97 |
-
"eval_loss": 0.03602377325296402,
|
| 98 |
-
"eval_runtime": 643.0305,
|
| 99 |
-
"eval_samples_per_second": 3.109,
|
| 100 |
-
"eval_steps_per_second": 0.194,
|
| 101 |
-
"eval_wer": 72.37405961337015,
|
| 102 |
-
"step": 471
|
| 103 |
-
},
|
| 104 |
-
{
|
| 105 |
-
"epoch": 3.1856,
|
| 106 |
-
"grad_norm": 129181.1328125,
|
| 107 |
-
"learning_rate": 0.0001,
|
| 108 |
-
"loss": 0.0326,
|
| 109 |
-
"step": 500
|
| 110 |
-
},
|
| 111 |
-
{
|
| 112 |
-
"epoch": 3.5056000000000003,
|
| 113 |
-
"grad_norm": 195835.4375,
|
| 114 |
-
"learning_rate": 9.991016468658499e-05,
|
| 115 |
-
"loss": 0.0327,
|
| 116 |
-
"step": 550
|
| 117 |
-
},
|
| 118 |
-
{
|
| 119 |
-
"epoch": 3.8256,
|
| 120 |
-
"grad_norm": 177487.765625,
|
| 121 |
-
"learning_rate": 9.964098156168142e-05,
|
| 122 |
-
"loss": 0.0322,
|
| 123 |
-
"step": 600
|
| 124 |
-
},
|
| 125 |
-
{
|
| 126 |
-
"epoch": 4.0,
|
| 127 |
-
"eval_cer": 18.302452013757904,
|
| 128 |
-
"eval_loss": 0.033060286194086075,
|
| 129 |
-
"eval_runtime": 627.0494,
|
| 130 |
-
"eval_samples_per_second": 3.188,
|
| 131 |
-
"eval_steps_per_second": 0.199,
|
| 132 |
-
"eval_wer": 46.46700314255785,
|
| 133 |
-
"step": 628
|
| 134 |
-
},
|
| 135 |
-
{
|
| 136 |
-
"epoch": 4.1408,
|
| 137 |
-
"grad_norm": 113981.9140625,
|
| 138 |
-
"learning_rate": 9.919341791130496e-05,
|
| 139 |
-
"loss": 0.0279,
|
| 140 |
-
"step": 650
|
| 141 |
-
},
|
| 142 |
-
{
|
| 143 |
-
"epoch": 4.4608,
|
| 144 |
-
"grad_norm": 98992.5859375,
|
| 145 |
-
"learning_rate": 9.85690820162878e-05,
|
| 146 |
-
"loss": 0.0234,
|
| 147 |
-
"step": 700
|
| 148 |
-
},
|
| 149 |
-
{
|
| 150 |
-
"epoch": 4.7808,
|
| 151 |
-
"grad_norm": 119039.2421875,
|
| 152 |
-
"learning_rate": 9.777021737306214e-05,
|
| 153 |
-
"loss": 0.0238,
|
| 154 |
-
"step": 750
|
| 155 |
-
},
|
| 156 |
-
{
|
| 157 |
-
"epoch": 5.0,
|
| 158 |
-
"eval_cer": 33.07777654499057,
|
| 159 |
-
"eval_loss": 0.03270672634243965,
|
| 160 |
-
"eval_runtime": 643.9344,
|
| 161 |
-
"eval_samples_per_second": 3.104,
|
| 162 |
-
"eval_steps_per_second": 0.194,
|
| 163 |
-
"eval_wer": 69.15531854109133,
|
| 164 |
-
"step": 785
|
| 165 |
-
},
|
| 166 |
-
{
|
| 167 |
-
"epoch": 5.096,
|
| 168 |
-
"grad_norm": 73677.765625,
|
| 169 |
-
"learning_rate": 9.6799694631852e-05,
|
| 170 |
-
"loss": 0.0214,
|
| 171 |
-
"step": 800
|
| 172 |
-
},
|
| 173 |
-
{
|
| 174 |
-
"epoch": 5.416,
|
| 175 |
-
"grad_norm": 103333.7421875,
|
| 176 |
-
"learning_rate": 9.56610012812427e-05,
|
| 177 |
-
"loss": 0.0168,
|
| 178 |
-
"step": 850
|
| 179 |
-
},
|
| 180 |
-
{
|
| 181 |
-
"epoch": 5.736,
|
| 182 |
-
"grad_norm": 99932.6171875,
|
| 183 |
-
"learning_rate": 9.435822911619564e-05,
|
| 184 |
-
"loss": 0.0187,
|
| 185 |
-
"step": 900
|
| 186 |
-
},
|
| 187 |
-
{
|
| 188 |
-
"epoch": 6.0,
|
| 189 |
-
"eval_cer": 21.48008432264507,
|
| 190 |
-
"eval_loss": 0.03448895364999771,
|
| 191 |
-
"eval_runtime": 624.9105,
|
| 192 |
-
"eval_samples_per_second": 3.199,
|
| 193 |
-
"eval_steps_per_second": 0.2,
|
| 194 |
-
"eval_wer": 51.385582325492805,
|
| 195 |
-
"step": 942
|
| 196 |
-
},
|
| 197 |
-
{
|
| 198 |
-
"epoch": 6.0512,
|
| 199 |
-
"grad_norm": 88239.1171875,
|
| 200 |
-
"learning_rate": 9.289605953454107e-05,
|
| 201 |
-
"loss": 0.0159,
|
| 202 |
-
"step": 950
|
| 203 |
-
},
|
| 204 |
-
{
|
| 205 |
-
"epoch": 6.3712,
|
| 206 |
-
"grad_norm": 66374.265625,
|
| 207 |
-
"learning_rate": 9.127974671478432e-05,
|
| 208 |
-
"loss": 0.012,
|
| 209 |
-
"step": 1000
|
| 210 |
-
},
|
| 211 |
-
{
|
| 212 |
-
"epoch": 6.6912,
|
| 213 |
-
"grad_norm": 64054.11328125,
|
| 214 |
-
"learning_rate": 8.951509873567499e-05,
|
| 215 |
-
"loss": 0.0125,
|
| 216 |
-
"step": 1050
|
| 217 |
-
},
|
| 218 |
-
{
|
| 219 |
-
"epoch": 7.0,
|
| 220 |
-
"eval_cer": 12.522467546876733,
|
| 221 |
-
"eval_loss": 0.03343856334686279,
|
| 222 |
-
"eval_runtime": 610.1269,
|
| 223 |
-
"eval_samples_per_second": 3.276,
|
| 224 |
-
"eval_steps_per_second": 0.205,
|
| 225 |
-
"eval_wer": 35.02523569183887,
|
| 226 |
-
"step": 1099
|
| 227 |
-
},
|
| 228 |
-
{
|
| 229 |
-
"epoch": 7.0064,
|
| 230 |
-
"grad_norm": 62407.08203125,
|
| 231 |
-
"learning_rate": 8.760845670538387e-05,
|
| 232 |
-
"loss": 0.0129,
|
| 233 |
-
"step": 1100
|
| 234 |
-
},
|
| 235 |
-
{
|
| 236 |
-
"epoch": 7.3264,
|
| 237 |
-
"grad_norm": 99814.7421875,
|
| 238 |
-
"learning_rate": 8.556667197528543e-05,
|
| 239 |
-
"loss": 0.0083,
|
| 240 |
-
"step": 1150
|
| 241 |
-
},
|
| 242 |
-
{
|
| 243 |
-
"epoch": 7.6464,
|
| 244 |
-
"grad_norm": 82683.46875,
|
| 245 |
-
"learning_rate": 8.339708152022585e-05,
|
| 246 |
-
"loss": 0.008,
|
| 247 |
-
"step": 1200
|
| 248 |
-
},
|
| 249 |
-
{
|
| 250 |
-
"epoch": 7.9664,
|
| 251 |
-
"grad_norm": 60897.98046875,
|
| 252 |
-
"learning_rate": 8.110748157374565e-05,
|
| 253 |
-
"loss": 0.0085,
|
| 254 |
-
"step": 1250
|
| 255 |
-
},
|
| 256 |
-
{
|
| 257 |
-
"epoch": 8.0,
|
| 258 |
-
"eval_cer": 17.333296349717074,
|
| 259 |
-
"eval_loss": 0.03527013212442398,
|
| 260 |
-
"eval_runtime": 621.4593,
|
| 261 |
-
"eval_samples_per_second": 3.217,
|
| 262 |
-
"eval_steps_per_second": 0.201,
|
| 263 |
-
"eval_wer": 44.2386439386725,
|
| 264 |
-
"step": 1256
|
| 265 |
-
},
|
| 266 |
-
{
|
| 267 |
-
"epoch": 8.2816,
|
| 268 |
-
"grad_norm": 69180.1875,
|
| 269 |
-
"learning_rate": 7.870609961299627e-05,
|
| 270 |
-
"loss": 0.0057,
|
| 271 |
-
"step": 1300
|
| 272 |
-
},
|
| 273 |
-
{
|
| 274 |
-
"epoch": 8.6016,
|
| 275 |
-
"grad_norm": 75431.734375,
|
| 276 |
-
"learning_rate": 7.620156479402066e-05,
|
| 277 |
-
"loss": 0.0053,
|
| 278 |
-
"step": 1350
|
| 279 |
-
},
|
| 280 |
-
{
|
| 281 |
-
"epoch": 8.9216,
|
| 282 |
-
"grad_norm": 41871.2890625,
|
| 283 |
-
"learning_rate": 7.360287694363566e-05,
|
| 284 |
-
"loss": 0.0054,
|
| 285 |
-
"step": 1400
|
| 286 |
-
},
|
| 287 |
-
{
|
| 288 |
-
"epoch": 9.0,
|
| 289 |
-
"eval_cer": 11.132808165982471,
|
| 290 |
-
"eval_loss": 0.037653908133506775,
|
| 291 |
-
"eval_runtime": 611.4108,
|
| 292 |
-
"eval_samples_per_second": 3.269,
|
| 293 |
-
"eval_steps_per_second": 0.204,
|
| 294 |
-
"eval_wer": 32.14455766117513,
|
| 295 |
-
"step": 1413
|
| 296 |
-
},
|
| 297 |
-
{
|
| 298 |
-
"epoch": 9.2368,
|
| 299 |
-
"grad_norm": 72383.8515625,
|
| 300 |
-
"learning_rate": 7.091937421934157e-05,
|
| 301 |
-
"loss": 0.0039,
|
| 302 |
-
"step": 1450
|
| 303 |
-
},
|
| 304 |
-
{
|
| 305 |
-
"epoch": 9.556799999999999,
|
| 306 |
-
"grad_norm": 58724.45703125,
|
| 307 |
-
"learning_rate": 6.816069955346985e-05,
|
| 308 |
-
"loss": 0.0036,
|
| 309 |
-
"step": 1500
|
| 310 |
-
},
|
| 311 |
-
{
|
| 312 |
-
"epoch": 9.8768,
|
| 313 |
-
"grad_norm": 55592.59765625,
|
| 314 |
-
"learning_rate": 6.533676600214928e-05,
|
| 315 |
-
"loss": 0.0031,
|
| 316 |
-
"step": 1550
|
| 317 |
-
},
|
| 318 |
-
{
|
| 319 |
-
"epoch": 10.0,
|
| 320 |
-
"eval_cer": 16.999334294907356,
|
| 321 |
-
"eval_loss": 0.038220491260290146,
|
| 322 |
-
"eval_runtime": 634.3141,
|
| 323 |
-
"eval_samples_per_second": 3.151,
|
| 324 |
-
"eval_steps_per_second": 0.197,
|
| 325 |
-
"eval_wer": 44.54813827254547,
|
| 326 |
-
"step": 1570
|
| 327 |
-
},
|
| 328 |
-
{
|
| 329 |
-
"epoch": 10.192,
|
| 330 |
-
"grad_norm": 35171.7109375,
|
| 331 |
-
"learning_rate": 6.245772112360568e-05,
|
| 332 |
-
"loss": 0.0026,
|
| 333 |
-
"step": 1600
|
| 334 |
-
},
|
| 335 |
-
{
|
| 336 |
-
"epoch": 10.512,
|
| 337 |
-
"grad_norm": 42814.4921875,
|
| 338 |
-
"learning_rate": 5.953391051379904e-05,
|
| 339 |
-
"loss": 0.0021,
|
| 340 |
-
"step": 1650
|
| 341 |
-
},
|
| 342 |
-
{
|
| 343 |
-
"epoch": 10.832,
|
| 344 |
-
"grad_norm": 35447.25390625,
|
| 345 |
-
"learning_rate": 5.6575840630429286e-05,
|
| 346 |
-
"loss": 0.0022,
|
| 347 |
-
"step": 1700
|
| 348 |
-
},
|
| 349 |
-
{
|
| 350 |
-
"epoch": 11.0,
|
| 351 |
-
"eval_cer": 13.609231110617998,
|
| 352 |
-
"eval_loss": 0.039708610624074936,
|
| 353 |
-
"eval_runtime": 620.4186,
|
| 354 |
-
"eval_samples_per_second": 3.222,
|
| 355 |
-
"eval_steps_per_second": 0.201,
|
| 356 |
-
"eval_wer": 37.786877440243785,
|
| 357 |
-
"step": 1727
|
| 358 |
-
},
|
| 359 |
-
{
|
| 360 |
-
"epoch": 11.1472,
|
| 361 |
-
"grad_norm": 55043.80078125,
|
| 362 |
-
"learning_rate": 5.359414103889947e-05,
|
| 363 |
-
"loss": 0.0017,
|
| 364 |
-
"step": 1750
|
| 365 |
-
},
|
| 366 |
-
{
|
| 367 |
-
"epoch": 11.4672,
|
| 368 |
-
"grad_norm": 36882.04296875,
|
| 369 |
-
"learning_rate": 5.059952621590216e-05,
|
| 370 |
-
"loss": 0.0014,
|
| 371 |
-
"step": 1800
|
| 372 |
-
},
|
| 373 |
-
{
|
| 374 |
-
"epoch": 11.7872,
|
| 375 |
-
"grad_norm": 36965.68359375,
|
| 376 |
-
"learning_rate": 4.7602757047884595e-05,
|
| 377 |
-
"loss": 0.0012,
|
| 378 |
-
"step": 1850
|
| 379 |
-
},
|
| 380 |
-
{
|
| 381 |
-
"epoch": 12.0,
|
| 382 |
-
"eval_cer": 12.376012426495063,
|
| 383 |
-
"eval_loss": 0.041336335241794586,
|
| 384 |
-
"eval_runtime": 613.9261,
|
| 385 |
-
"eval_samples_per_second": 3.256,
|
| 386 |
-
"eval_steps_per_second": 0.204,
|
| 387 |
-
"eval_wer": 35.02047424054852,
|
| 388 |
-
"step": 1884
|
| 389 |
-
},
|
| 390 |
-
{
|
| 391 |
-
"epoch": 12.1024,
|
| 392 |
-
"grad_norm": 35653.84375,
|
| 393 |
-
"learning_rate": 4.461460216274445e-05,
|
| 394 |
-
"loss": 0.0011,
|
| 395 |
-
"step": 1900
|
| 396 |
-
},
|
| 397 |
-
{
|
| 398 |
-
"epoch": 12.4224,
|
| 399 |
-
"grad_norm": 17012.79296875,
|
| 400 |
-
"learning_rate": 4.1645799233707284e-05,
|
| 401 |
-
"loss": 0.0008,
|
| 402 |
-
"step": 1950
|
| 403 |
-
},
|
| 404 |
-
{
|
| 405 |
-
"epoch": 12.7424,
|
| 406 |
-
"grad_norm": 25617.26171875,
|
| 407 |
-
"learning_rate": 3.870701639443698e-05,
|
| 408 |
-
"loss": 0.0008,
|
| 409 |
-
"step": 2000
|
| 410 |
-
},
|
| 411 |
-
{
|
| 412 |
-
"epoch": 13.0,
|
| 413 |
-
"eval_cer": 13.077221790746698,
|
| 414 |
-
"eval_loss": 0.04063521698117256,
|
| 415 |
-
"eval_runtime": 615.5521,
|
| 416 |
-
"eval_samples_per_second": 3.247,
|
| 417 |
-
"eval_steps_per_second": 0.203,
|
| 418 |
-
"eval_wer": 33.51585563279688,
|
| 419 |
-
"step": 2041
|
| 420 |
-
},
|
| 421 |
-
{
|
| 422 |
-
"epoch": 13.0576,
|
| 423 |
-
"grad_norm": 43133.0625,
|
| 424 |
-
"learning_rate": 3.580881390403052e-05,
|
| 425 |
-
"loss": 0.0007,
|
| 426 |
-
"step": 2050
|
| 427 |
-
},
|
| 428 |
-
{
|
| 429 |
-
"epoch": 13.3776,
|
| 430 |
-
"grad_norm": 28860.919921875,
|
| 431 |
-
"learning_rate": 3.296160619965056e-05,
|
| 432 |
-
"loss": 0.0004,
|
| 433 |
-
"step": 2100
|
| 434 |
-
},
|
| 435 |
-
{
|
| 436 |
-
"epoch": 13.6976,
|
| 437 |
-
"grad_norm": 27712.271484375,
|
| 438 |
-
"learning_rate": 3.0175624473156316e-05,
|
| 439 |
-
"loss": 0.0005,
|
| 440 |
-
"step": 2150
|
| 441 |
-
},
|
| 442 |
-
{
|
| 443 |
-
"epoch": 14.0,
|
| 444 |
-
"eval_cer": 12.631754132919118,
|
| 445 |
-
"eval_loss": 0.04265804588794708,
|
| 446 |
-
"eval_runtime": 611.9643,
|
| 447 |
-
"eval_samples_per_second": 3.267,
|
| 448 |
-
"eval_steps_per_second": 0.204,
|
| 449 |
-
"eval_wer": 33.94438624892867,
|
| 450 |
-
"step": 2198
|
| 451 |
-
},
|
| 452 |
-
{
|
| 453 |
-
"epoch": 14.0128,
|
| 454 |
-
"grad_norm": 4747.3388671875,
|
| 455 |
-
"learning_rate": 2.7460879906210487e-05,
|
| 456 |
-
"loss": 0.0004,
|
| 457 |
-
"step": 2200
|
| 458 |
-
},
|
| 459 |
-
{
|
| 460 |
-
"epoch": 14.3328,
|
| 461 |
-
"grad_norm": 10129.0400390625,
|
| 462 |
-
"learning_rate": 2.482712769597363e-05,
|
| 463 |
-
"loss": 0.0002,
|
| 464 |
-
"step": 2250
|
| 465 |
-
},
|
| 466 |
-
{
|
| 467 |
-
"epoch": 14.6528,
|
| 468 |
-
"grad_norm": 17136.42578125,
|
| 469 |
-
"learning_rate": 2.22838320006563e-05,
|
| 470 |
-
"loss": 0.0002,
|
| 471 |
-
"step": 2300
|
| 472 |
-
},
|
| 473 |
-
{
|
| 474 |
-
"epoch": 14.9728,
|
| 475 |
-
"grad_norm": 2899.75439453125,
|
| 476 |
-
"learning_rate": 1.9840131930894333e-05,
|
| 477 |
-
"loss": 0.0002,
|
| 478 |
-
"step": 2350
|
| 479 |
-
},
|
| 480 |
-
{
|
| 481 |
-
"epoch": 15.0,
|
| 482 |
-
"eval_cer": 10.566958837235104,
|
| 483 |
-
"eval_loss": 0.04313713312149048,
|
| 484 |
-
"eval_runtime": 619.263,
|
| 485 |
-
"eval_samples_per_second": 3.228,
|
| 486 |
-
"eval_steps_per_second": 0.202,
|
| 487 |
-
"eval_wer": 31.449385772783543,
|
| 488 |
-
"step": 2355
|
| 489 |
-
},
|
| 490 |
-
{
|
| 491 |
-
"epoch": 15.288,
|
| 492 |
-
"grad_norm": 1868.322021484375,
|
| 493 |
-
"learning_rate": 1.7504808709154104e-05,
|
| 494 |
-
"loss": 0.0001,
|
| 495 |
-
"step": 2400
|
| 496 |
-
},
|
| 497 |
-
{
|
| 498 |
-
"epoch": 15.608,
|
| 499 |
-
"grad_norm": 1876.469970703125,
|
| 500 |
-
"learning_rate": 1.5286254115177623e-05,
|
| 501 |
-
"loss": 0.0001,
|
| 502 |
-
"step": 2450
|
| 503 |
-
},
|
| 504 |
-
{
|
| 505 |
-
"epoch": 15.928,
|
| 506 |
-
"grad_norm": 2221.869140625,
|
| 507 |
-
"learning_rate": 1.3192440330856004e-05,
|
| 508 |
-
"loss": 0.0001,
|
| 509 |
-
"step": 2500
|
| 510 |
-
},
|
| 511 |
-
{
|
| 512 |
-
"epoch": 16.0,
|
| 513 |
-
"eval_cer": 12.524131809608344,
|
| 514 |
-
"eval_loss": 0.043873097747564316,
|
| 515 |
-
"eval_runtime": 616.7174,
|
| 516 |
-
"eval_samples_per_second": 3.241,
|
| 517 |
-
"eval_steps_per_second": 0.203,
|
| 518 |
-
"eval_wer": 35.372821636034665,
|
| 519 |
-
"step": 2512
|
| 520 |
-
},
|
| 521 |
-
{
|
| 522 |
-
"epoch": 16.2432,
|
| 523 |
-
"grad_norm": 693.1599731445312,
|
| 524 |
-
"learning_rate": 1.123089129289117e-05,
|
| 525 |
-
"loss": 0.0,
|
| 526 |
-
"step": 2550
|
| 527 |
-
},
|
| 528 |
-
{
|
| 529 |
-
"epoch": 16.5632,
|
| 530 |
-
"grad_norm": 184.430908203125,
|
| 531 |
-
"learning_rate": 9.408655656187282e-06,
|
| 532 |
-
"loss": 0.0,
|
| 533 |
-
"step": 2600
|
| 534 |
-
},
|
| 535 |
-
{
|
| 536 |
-
"epoch": 16.8832,
|
| 537 |
-
"grad_norm": 1887.3065185546875,
|
| 538 |
-
"learning_rate": 7.732281465125907e-06,
|
| 539 |
-
"loss": 0.0,
|
| 540 |
-
"step": 2650
|
| 541 |
-
},
|
| 542 |
-
{
|
| 543 |
-
"epoch": 17.0,
|
| 544 |
-
"eval_cer": 10.107067569066903,
|
| 545 |
-
"eval_loss": 0.04502426087856293,
|
| 546 |
-
"eval_runtime": 611.9114,
|
| 547 |
-
"eval_samples_per_second": 3.267,
|
| 548 |
-
"eval_steps_per_second": 0.204,
|
| 549 |
-
"eval_wer": 29.268641081801732,
|
| 550 |
-
"step": 2669
|
| 551 |
-
},
|
| 552 |
-
{
|
| 553 |
-
"epoch": 17.1984,
|
| 554 |
-
"grad_norm": 270.3798828125,
|
| 555 |
-
"learning_rate": 6.207792623741249e-06,
|
| 556 |
-
"loss": 0.0,
|
| 557 |
-
"step": 2700
|
| 558 |
-
},
|
| 559 |
-
{
|
| 560 |
-
"epoch": 17.5184,
|
| 561 |
-
"grad_norm": 287.86920166015625,
|
| 562 |
-
"learning_rate": 4.840667249347824e-06,
|
| 563 |
-
"loss": 0.0,
|
| 564 |
-
"step": 2750
|
| 565 |
-
},
|
| 566 |
-
{
|
| 567 |
-
"epoch": 17.8384,
|
| 568 |
-
"grad_norm": 200.27944946289062,
|
| 569 |
-
"learning_rate": 3.63581798740511e-06,
|
| 570 |
-
"loss": 0.0,
|
| 571 |
-
"step": 2800
|
| 572 |
-
},
|
| 573 |
-
{
|
| 574 |
-
"epoch": 18.0,
|
| 575 |
-
"eval_cer": 11.283146566071231,
|
| 576 |
-
"eval_loss": 0.04671463742852211,
|
| 577 |
-
"eval_runtime": 610.44,
|
| 578 |
-
"eval_samples_per_second": 3.275,
|
| 579 |
-
"eval_steps_per_second": 0.205,
|
| 580 |
-
"eval_wer": 30.45900390439006,
|
| 581 |
-
"step": 2826
|
| 582 |
-
},
|
| 583 |
-
{
|
| 584 |
-
"epoch": 18.1536,
|
| 585 |
-
"grad_norm": 53.246883392333984,
|
| 586 |
-
"learning_rate": 2.59757435835567e-06,
|
| 587 |
-
"loss": 0.0,
|
| 588 |
-
"step": 2850
|
| 589 |
-
},
|
| 590 |
-
{
|
| 591 |
-
"epoch": 18.4736,
|
| 592 |
-
"grad_norm": 68.43990325927734,
|
| 593 |
-
"learning_rate": 1.729667199872187e-06,
|
| 594 |
-
"loss": 0.0,
|
| 595 |
-
"step": 2900
|
| 596 |
-
},
|
| 597 |
-
{
|
| 598 |
-
"epoch": 18.7936,
|
| 599 |
-
"grad_norm": 63.084373474121094,
|
| 600 |
-
"learning_rate": 1.0352152604185428e-06,
|
| 601 |
-
"loss": 0.0,
|
| 602 |
-
"step": 2950
|
| 603 |
-
},
|
| 604 |
-
{
|
| 605 |
-
"epoch": 19.0,
|
| 606 |
-
"eval_cer": 11.1267058692999,
|
| 607 |
-
"eval_loss": 0.047411538660526276,
|
| 608 |
-
"eval_runtime": 609.5688,
|
| 609 |
-
"eval_samples_per_second": 3.279,
|
| 610 |
-
"eval_steps_per_second": 0.205,
|
| 611 |
-
"eval_wer": 30.335206170840873,
|
| 612 |
-
"step": 2983
|
| 613 |
-
},
|
| 614 |
-
{
|
| 615 |
-
"epoch": 19.1088,
|
| 616 |
-
"grad_norm": 42.04766845703125,
|
| 617 |
-
"learning_rate": 5.167139923000553e-07,
|
| 618 |
-
"loss": 0.0,
|
| 619 |
-
"step": 3000
|
| 620 |
-
},
|
| 621 |
-
{
|
| 622 |
-
"epoch": 19.4288,
|
| 623 |
-
"grad_norm": 56.3652458190918,
|
| 624 |
-
"learning_rate": 1.760265844738096e-07,
|
| 625 |
-
"loss": 0.0,
|
| 626 |
-
"step": 3050
|
| 627 |
-
},
|
| 628 |
-
{
|
| 629 |
-
"epoch": 19.7488,
|
| 630 |
-
"grad_norm": 69.01725769042969,
|
| 631 |
-
"learning_rate": 1.4377267342158274e-08,
|
| 632 |
-
"loss": 0.0,
|
| 633 |
-
"step": 3100
|
| 634 |
-
},
|
| 635 |
-
{
|
| 636 |
-
"epoch": 19.8768,
|
| 637 |
-
"eval_cer": 11.179407522467546,
|
| 638 |
-
"eval_loss": 0.047550249844789505,
|
| 639 |
-
"eval_runtime": 603.4051,
|
| 640 |
-
"eval_samples_per_second": 3.313,
|
| 641 |
-
"eval_steps_per_second": 0.207,
|
| 642 |
-
"eval_wer": 30.378059232454053,
|
| 643 |
-
"step": 3120
|
| 644 |
}
|
| 645 |
],
|
| 646 |
"logging_steps": 50,
|
|
@@ -655,12 +52,12 @@
|
|
| 655 |
"should_evaluate": false,
|
| 656 |
"should_log": false,
|
| 657 |
"should_save": true,
|
| 658 |
-
"should_training_stop":
|
| 659 |
},
|
| 660 |
"attributes": {}
|
| 661 |
}
|
| 662 |
},
|
| 663 |
-
"total_flos":
|
| 664 |
"train_batch_size": 16,
|
| 665 |
"trial_name": null,
|
| 666 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 28.53061613179697,
|
| 3 |
+
"best_model_checkpoint": "./Whisper_tiny_fine_tune_Quran/checkpoint-157",
|
| 4 |
+
"epoch": 1.0,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 157,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
"epoch": 0.32,
|
| 13 |
+
"grad_norm": 151793.484375,
|
| 14 |
"learning_rate": 1e-05,
|
| 15 |
+
"loss": 0.0293,
|
| 16 |
"step": 50
|
| 17 |
},
|
| 18 |
{
|
| 19 |
"epoch": 0.64,
|
| 20 |
+
"grad_norm": 88808.3671875,
|
| 21 |
"learning_rate": 2e-05,
|
| 22 |
+
"loss": 0.025,
|
| 23 |
"step": 100
|
| 24 |
},
|
| 25 |
{
|
| 26 |
"epoch": 0.96,
|
| 27 |
+
"grad_norm": 80433.359375,
|
| 28 |
"learning_rate": 3e-05,
|
| 29 |
+
"loss": 0.0195,
|
| 30 |
"step": 150
|
| 31 |
},
|
| 32 |
{
|
| 33 |
"epoch": 1.0,
|
| 34 |
+
"eval_cer": 9.66548319094641,
|
| 35 |
+
"eval_loss": 0.025752274319529533,
|
| 36 |
+
"eval_runtime": 658.049,
|
| 37 |
+
"eval_samples_per_second": 3.038,
|
| 38 |
+
"eval_steps_per_second": 0.19,
|
| 39 |
+
"eval_wer": 28.53061613179697,
|
| 40 |
"step": 157
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 41 |
}
|
| 42 |
],
|
| 43 |
"logging_steps": 50,
|
|
|
|
| 52 |
"should_evaluate": false,
|
| 53 |
"should_log": false,
|
| 54 |
"should_save": true,
|
| 55 |
+
"should_training_stop": false
|
| 56 |
},
|
| 57 |
"attributes": {}
|
| 58 |
}
|
| 59 |
},
|
| 60 |
+
"total_flos": 2.461888512e+17,
|
| 61 |
"train_batch_size": 16,
|
| 62 |
"trial_name": null,
|
| 63 |
"trial_params": null
|
last-checkpoint/training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 5624
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:317f0bc90b9977ce1f17eccbf424721a3903a4332b9f78d27fa2e7877abd38d0
|
| 3 |
size 5624
|