End of training
Browse files
all_results.json
CHANGED
|
@@ -1,12 +1,12 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 12.01,
|
| 3 |
-
"eval_loss": 0.
|
| 4 |
-
"eval_runtime":
|
| 5 |
-
"eval_samples_per_second":
|
| 6 |
-
"eval_steps_per_second": 0.
|
| 7 |
-
"eval_wer": 18.
|
| 8 |
-
"train_loss": 0.
|
| 9 |
-
"train_runtime":
|
| 10 |
-
"train_samples_per_second": 5.
|
| 11 |
-
"train_steps_per_second": 0.
|
| 12 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 12.01,
|
| 3 |
+
"eval_loss": 0.3812163174152374,
|
| 4 |
+
"eval_runtime": 951.9575,
|
| 5 |
+
"eval_samples_per_second": 6.924,
|
| 6 |
+
"eval_steps_per_second": 0.433,
|
| 7 |
+
"eval_wer": 18.775568066750374,
|
| 8 |
+
"train_loss": 0.106446673027426,
|
| 9 |
+
"train_runtime": 27653.1068,
|
| 10 |
+
"train_samples_per_second": 5.786,
|
| 11 |
+
"train_steps_per_second": 0.181
|
| 12 |
}
|
eval_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 12.01,
|
| 3 |
-
"eval_loss": 0.
|
| 4 |
-
"eval_runtime":
|
| 5 |
-
"eval_samples_per_second":
|
| 6 |
-
"eval_steps_per_second": 0.
|
| 7 |
-
"eval_wer": 18.
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 12.01,
|
| 3 |
+
"eval_loss": 0.3812163174152374,
|
| 4 |
+
"eval_runtime": 951.9575,
|
| 5 |
+
"eval_samples_per_second": 6.924,
|
| 6 |
+
"eval_steps_per_second": 0.433,
|
| 7 |
+
"eval_wer": 18.775568066750374
|
| 8 |
}
|
nohup.out
CHANGED
|
@@ -33935,3 +33935,25 @@ The progress bars may be unreliable.
|
|
| 33935 |
|
| 33936 |
|
| 33937 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 33935 |
|
| 33936 |
|
| 33937 |
|
| 33938 |
+
To https://huggingface.co/xezpeleta/whisper-small-eu-v2
|
| 33939 |
+
9ecd845..6d7e844 main -> main
|
| 33940 |
+
|
| 33941 |
+
07/20/2023 19:08:10 - WARNING - huggingface_hub.repository - To https://huggingface.co/xezpeleta/whisper-small-eu-v2
|
| 33942 |
+
9ecd845..6d7e844 main -> main
|
| 33943 |
+
|
| 33944 |
+
***** train metrics *****
|
| 33945 |
+
epoch = 12.01
|
| 33946 |
+
train_loss = 0.1064
|
| 33947 |
+
train_runtime = 7:40:53.10
|
| 33948 |
+
train_samples_per_second = 5.786
|
| 33949 |
+
train_steps_per_second = 0.181
|
| 33950 |
+
07/20/2023 19:08:13 - INFO - __main__ - *** Evaluate ***
|
| 33951 |
+
[INFO|trainer.py:2955] 2023-07-20 19:08:13,052 >> ***** Running Evaluation *****
|
| 33952 |
+
[INFO|trainer.py:2959] 2023-07-20 19:08:13,052 >> Num examples: Unknown
|
| 33953 |
+
[INFO|trainer.py:2960] 2023-07-20 19:08:13,052 >> Batch size = 16
|
| 33954 |
+
|
| 33955 |
+
[INFO|trainer_utils.py:689] 2023-07-20 19:08:16,031 >> The following columns in the evaluation set don't have a corresponding argument in `WhisperForConditionalGeneration.forward` and have been ignored: input_length. If input_length are not expected by `WhisperForConditionalGeneration.forward`, you can safely ignore this message.
|
| 33956 |
+
[INFO|trainer.py:2700] 2023-07-20 19:24:05,011 >> Saving model checkpoint to ./
|
| 33957 |
+
[INFO|configuration_utils.py:447] 2023-07-20 19:24:05,011 >> Configuration saved in ./config.json
|
| 33958 |
+
[INFO|modeling_utils.py:1693] 2023-07-20 19:24:06,334 >> Model weights saved in ./pytorch_model.bin
|
| 33959 |
+
[INFO|feature_extraction_utils.py:368] 2023-07-20 19:24:06,347 >> Feature extractor saved in ./preprocessor_config.json
|
runs/Jul20_11-05-05_tknadmin-System-Product-Name/events.out.tfevents.1689873845.tknadmin-System-Product-Name.2399.2
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a1eed888dec52a9e6ad526f063140bc6e06e8c499120c73eda9d1b02bfa5b501
|
| 3 |
+
size 358
|
train_results.json
CHANGED
|
@@ -1,7 +1,7 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 12.01,
|
| 3 |
-
"train_loss": 0.
|
| 4 |
-
"train_runtime":
|
| 5 |
-
"train_samples_per_second": 5.
|
| 6 |
-
"train_steps_per_second": 0.
|
| 7 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 12.01,
|
| 3 |
+
"train_loss": 0.106446673027426,
|
| 4 |
+
"train_runtime": 27653.1068,
|
| 5 |
+
"train_samples_per_second": 5.786,
|
| 6 |
+
"train_steps_per_second": 0.181
|
| 7 |
}
|
trainer_state.json
CHANGED
|
@@ -1,7 +1,7 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 18.
|
| 3 |
-
"best_model_checkpoint": "./checkpoint-
|
| 4 |
-
"epoch": 12.
|
| 5 |
"global_step": 5000,
|
| 6 |
"is_hyper_param_search": false,
|
| 7 |
"is_local_process_zero": true,
|
|
@@ -10,685 +10,685 @@
|
|
| 10 |
{
|
| 11 |
"epoch": 0.01,
|
| 12 |
"learning_rate": 4.4e-07,
|
| 13 |
-
"loss": 1.
|
| 14 |
"step": 25
|
| 15 |
},
|
| 16 |
{
|
| 17 |
"epoch": 0.01,
|
| 18 |
"learning_rate": 9.400000000000001e-07,
|
| 19 |
-
"loss": 1.
|
| 20 |
"step": 50
|
| 21 |
},
|
| 22 |
{
|
| 23 |
"epoch": 0.01,
|
| 24 |
"learning_rate": 1.44e-06,
|
| 25 |
-
"loss": 1.
|
| 26 |
"step": 75
|
| 27 |
},
|
| 28 |
{
|
| 29 |
"epoch": 0.02,
|
| 30 |
"learning_rate": 1.94e-06,
|
| 31 |
-
"loss": 0.
|
| 32 |
"step": 100
|
| 33 |
},
|
| 34 |
{
|
| 35 |
"epoch": 0.03,
|
| 36 |
"learning_rate": 2.4400000000000004e-06,
|
| 37 |
-
"loss": 0.
|
| 38 |
"step": 125
|
| 39 |
},
|
| 40 |
{
|
| 41 |
"epoch": 0.03,
|
| 42 |
"learning_rate": 2.9400000000000002e-06,
|
| 43 |
-
"loss": 0.
|
| 44 |
"step": 150
|
| 45 |
},
|
| 46 |
{
|
| 47 |
"epoch": 0.04,
|
| 48 |
"learning_rate": 3.44e-06,
|
| 49 |
-
"loss": 0.
|
| 50 |
"step": 175
|
| 51 |
},
|
| 52 |
{
|
| 53 |
"epoch": 0.04,
|
| 54 |
"learning_rate": 3.94e-06,
|
| 55 |
-
"loss": 0.
|
| 56 |
"step": 200
|
| 57 |
},
|
| 58 |
{
|
| 59 |
"epoch": 0.04,
|
| 60 |
"learning_rate": 4.440000000000001e-06,
|
| 61 |
-
"loss": 0.
|
| 62 |
"step": 225
|
| 63 |
},
|
| 64 |
{
|
| 65 |
"epoch": 0.05,
|
| 66 |
"learning_rate": 4.94e-06,
|
| 67 |
-
"loss": 0.
|
| 68 |
"step": 250
|
| 69 |
},
|
| 70 |
{
|
| 71 |
"epoch": 0.06,
|
| 72 |
"learning_rate": 5.4400000000000004e-06,
|
| 73 |
-
"loss": 0.
|
| 74 |
"step": 275
|
| 75 |
},
|
| 76 |
{
|
| 77 |
"epoch": 0.06,
|
| 78 |
"learning_rate": 5.94e-06,
|
| 79 |
-
"loss": 0.
|
| 80 |
"step": 300
|
| 81 |
},
|
| 82 |
{
|
| 83 |
"epoch": 0.07,
|
| 84 |
"learning_rate": 6.440000000000001e-06,
|
| 85 |
-
"loss": 0.
|
| 86 |
"step": 325
|
| 87 |
},
|
| 88 |
{
|
| 89 |
"epoch": 0.07,
|
| 90 |
"learning_rate": 6.9400000000000005e-06,
|
| 91 |
-
"loss": 0.
|
| 92 |
"step": 350
|
| 93 |
},
|
| 94 |
{
|
| 95 |
"epoch": 0.07,
|
| 96 |
"learning_rate": 7.440000000000001e-06,
|
| 97 |
-
"loss": 0.
|
| 98 |
"step": 375
|
| 99 |
},
|
| 100 |
{
|
| 101 |
"epoch": 0.08,
|
| 102 |
"learning_rate": 7.94e-06,
|
| 103 |
-
"loss": 0.
|
| 104 |
"step": 400
|
| 105 |
},
|
| 106 |
{
|
| 107 |
"epoch": 1.0,
|
| 108 |
"learning_rate": 8.44e-06,
|
| 109 |
-
"loss": 0.
|
| 110 |
"step": 425
|
| 111 |
},
|
| 112 |
{
|
| 113 |
"epoch": 1.01,
|
| 114 |
"learning_rate": 8.94e-06,
|
| 115 |
-
"loss": 0.
|
| 116 |
"step": 450
|
| 117 |
},
|
| 118 |
{
|
| 119 |
"epoch": 1.01,
|
| 120 |
"learning_rate": 9.440000000000001e-06,
|
| 121 |
-
"loss": 0.
|
| 122 |
"step": 475
|
| 123 |
},
|
| 124 |
{
|
| 125 |
"epoch": 1.02,
|
| 126 |
"learning_rate": 9.940000000000001e-06,
|
| 127 |
-
"loss": 0.
|
| 128 |
"step": 500
|
| 129 |
},
|
| 130 |
{
|
| 131 |
"epoch": 1.02,
|
| 132 |
"learning_rate": 9.951111111111111e-06,
|
| 133 |
-
"loss": 0.
|
| 134 |
"step": 525
|
| 135 |
},
|
| 136 |
{
|
| 137 |
"epoch": 1.03,
|
| 138 |
"learning_rate": 9.895555555555557e-06,
|
| 139 |
-
"loss": 0.
|
| 140 |
"step": 550
|
| 141 |
},
|
| 142 |
{
|
| 143 |
"epoch": 1.03,
|
| 144 |
"learning_rate": 9.84e-06,
|
| 145 |
-
"loss": 0.
|
| 146 |
"step": 575
|
| 147 |
},
|
| 148 |
{
|
| 149 |
"epoch": 1.04,
|
| 150 |
"learning_rate": 9.784444444444445e-06,
|
| 151 |
-
"loss": 0.
|
| 152 |
"step": 600
|
| 153 |
},
|
| 154 |
{
|
| 155 |
"epoch": 1.04,
|
| 156 |
"learning_rate": 9.72888888888889e-06,
|
| 157 |
-
"loss": 0.
|
| 158 |
"step": 625
|
| 159 |
},
|
| 160 |
{
|
| 161 |
"epoch": 1.05,
|
| 162 |
"learning_rate": 9.673333333333334e-06,
|
| 163 |
-
"loss": 0.
|
| 164 |
"step": 650
|
| 165 |
},
|
| 166 |
{
|
| 167 |
"epoch": 1.05,
|
| 168 |
"learning_rate": 9.617777777777778e-06,
|
| 169 |
-
"loss": 0.
|
| 170 |
"step": 675
|
| 171 |
},
|
| 172 |
{
|
| 173 |
"epoch": 1.06,
|
| 174 |
"learning_rate": 9.562222222222223e-06,
|
| 175 |
-
"loss": 0.
|
| 176 |
"step": 700
|
| 177 |
},
|
| 178 |
{
|
| 179 |
"epoch": 1.06,
|
| 180 |
"learning_rate": 9.506666666666667e-06,
|
| 181 |
-
"loss": 0.
|
| 182 |
"step": 725
|
| 183 |
},
|
| 184 |
{
|
| 185 |
"epoch": 1.07,
|
| 186 |
"learning_rate": 9.451111111111112e-06,
|
| 187 |
-
"loss": 0.
|
| 188 |
"step": 750
|
| 189 |
},
|
| 190 |
{
|
| 191 |
"epoch": 1.07,
|
| 192 |
"learning_rate": 9.395555555555556e-06,
|
| 193 |
-
"loss": 0.
|
| 194 |
"step": 775
|
| 195 |
},
|
| 196 |
{
|
| 197 |
"epoch": 1.08,
|
| 198 |
"learning_rate": 9.340000000000002e-06,
|
| 199 |
-
"loss": 0.
|
| 200 |
"step": 800
|
| 201 |
},
|
| 202 |
{
|
| 203 |
"epoch": 2.0,
|
| 204 |
"learning_rate": 9.284444444444444e-06,
|
| 205 |
-
"loss": 0.
|
| 206 |
"step": 825
|
| 207 |
},
|
| 208 |
{
|
| 209 |
"epoch": 2.01,
|
| 210 |
"learning_rate": 9.22888888888889e-06,
|
| 211 |
-
"loss": 0.
|
| 212 |
"step": 850
|
| 213 |
},
|
| 214 |
{
|
| 215 |
"epoch": 2.01,
|
| 216 |
"learning_rate": 9.173333333333334e-06,
|
| 217 |
-
"loss": 0.
|
| 218 |
"step": 875
|
| 219 |
},
|
| 220 |
{
|
| 221 |
"epoch": 2.02,
|
| 222 |
"learning_rate": 9.117777777777778e-06,
|
| 223 |
-
"loss": 0.
|
| 224 |
"step": 900
|
| 225 |
},
|
| 226 |
{
|
| 227 |
"epoch": 2.02,
|
| 228 |
"learning_rate": 9.062222222222224e-06,
|
| 229 |
-
"loss": 0.
|
| 230 |
"step": 925
|
| 231 |
},
|
| 232 |
{
|
| 233 |
"epoch": 2.03,
|
| 234 |
"learning_rate": 9.006666666666666e-06,
|
| 235 |
-
"loss": 0.
|
| 236 |
"step": 950
|
| 237 |
},
|
| 238 |
{
|
| 239 |
"epoch": 2.03,
|
| 240 |
"learning_rate": 8.951111111111112e-06,
|
| 241 |
-
"loss": 0.
|
| 242 |
"step": 975
|
| 243 |
},
|
| 244 |
{
|
| 245 |
"epoch": 2.04,
|
| 246 |
"learning_rate": 8.895555555555556e-06,
|
| 247 |
-
"loss": 0.
|
| 248 |
"step": 1000
|
| 249 |
},
|
| 250 |
{
|
| 251 |
"epoch": 2.04,
|
| 252 |
-
"eval_loss": 0.
|
| 253 |
-
"eval_runtime":
|
| 254 |
-
"eval_samples_per_second":
|
| 255 |
-
"eval_steps_per_second": 0.
|
| 256 |
-
"eval_wer": 22.
|
| 257 |
"step": 1000
|
| 258 |
},
|
| 259 |
{
|
| 260 |
"epoch": 2.04,
|
| 261 |
"learning_rate": 8.84e-06,
|
| 262 |
-
"loss": 0.
|
| 263 |
"step": 1025
|
| 264 |
},
|
| 265 |
{
|
| 266 |
"epoch": 2.05,
|
| 267 |
"learning_rate": 8.784444444444446e-06,
|
| 268 |
-
"loss": 0.
|
| 269 |
"step": 1050
|
| 270 |
},
|
| 271 |
{
|
| 272 |
"epoch": 2.05,
|
| 273 |
"learning_rate": 8.72888888888889e-06,
|
| 274 |
-
"loss": 0.
|
| 275 |
"step": 1075
|
| 276 |
},
|
| 277 |
{
|
| 278 |
"epoch": 2.06,
|
| 279 |
"learning_rate": 8.673333333333334e-06,
|
| 280 |
-
"loss": 0.
|
| 281 |
"step": 1100
|
| 282 |
},
|
| 283 |
{
|
| 284 |
"epoch": 2.06,
|
| 285 |
"learning_rate": 8.617777777777778e-06,
|
| 286 |
-
"loss": 0.
|
| 287 |
"step": 1125
|
| 288 |
},
|
| 289 |
{
|
| 290 |
"epoch": 2.07,
|
| 291 |
"learning_rate": 8.562222222222224e-06,
|
| 292 |
-
"loss": 0.
|
| 293 |
"step": 1150
|
| 294 |
},
|
| 295 |
{
|
| 296 |
"epoch": 2.07,
|
| 297 |
"learning_rate": 8.506666666666668e-06,
|
| 298 |
-
"loss": 0.
|
| 299 |
"step": 1175
|
| 300 |
},
|
| 301 |
{
|
| 302 |
"epoch": 2.08,
|
| 303 |
"learning_rate": 8.451111111111112e-06,
|
| 304 |
-
"loss": 0.
|
| 305 |
"step": 1200
|
| 306 |
},
|
| 307 |
{
|
| 308 |
"epoch": 2.08,
|
| 309 |
"learning_rate": 8.395555555555557e-06,
|
| 310 |
-
"loss": 0.
|
| 311 |
"step": 1225
|
| 312 |
},
|
| 313 |
{
|
| 314 |
"epoch": 3.0,
|
| 315 |
"learning_rate": 8.34e-06,
|
| 316 |
-
"loss": 0.
|
| 317 |
"step": 1250
|
| 318 |
},
|
| 319 |
{
|
| 320 |
"epoch": 3.01,
|
| 321 |
"learning_rate": 8.284444444444446e-06,
|
| 322 |
-
"loss": 0.
|
| 323 |
"step": 1275
|
| 324 |
},
|
| 325 |
{
|
| 326 |
"epoch": 3.01,
|
| 327 |
"learning_rate": 8.22888888888889e-06,
|
| 328 |
-
"loss": 0.
|
| 329 |
"step": 1300
|
| 330 |
},
|
| 331 |
{
|
| 332 |
"epoch": 3.02,
|
| 333 |
"learning_rate": 8.173333333333334e-06,
|
| 334 |
-
"loss": 0.
|
| 335 |
"step": 1325
|
| 336 |
},
|
| 337 |
{
|
| 338 |
"epoch": 3.02,
|
| 339 |
"learning_rate": 8.11777777777778e-06,
|
| 340 |
-
"loss": 0.
|
| 341 |
"step": 1350
|
| 342 |
},
|
| 343 |
{
|
| 344 |
"epoch": 3.03,
|
| 345 |
"learning_rate": 8.062222222222222e-06,
|
| 346 |
-
"loss": 0.
|
| 347 |
"step": 1375
|
| 348 |
},
|
| 349 |
{
|
| 350 |
"epoch": 3.03,
|
| 351 |
"learning_rate": 8.006666666666667e-06,
|
| 352 |
-
"loss": 0.
|
| 353 |
"step": 1400
|
| 354 |
},
|
| 355 |
{
|
| 356 |
"epoch": 3.04,
|
| 357 |
"learning_rate": 7.951111111111111e-06,
|
| 358 |
-
"loss": 0.
|
| 359 |
"step": 1425
|
| 360 |
},
|
| 361 |
{
|
| 362 |
"epoch": 3.04,
|
| 363 |
"learning_rate": 7.895555555555557e-06,
|
| 364 |
-
"loss": 0.
|
| 365 |
"step": 1450
|
| 366 |
},
|
| 367 |
{
|
| 368 |
"epoch": 3.05,
|
| 369 |
"learning_rate": 7.840000000000001e-06,
|
| 370 |
-
"loss": 0.
|
| 371 |
"step": 1475
|
| 372 |
},
|
| 373 |
{
|
| 374 |
"epoch": 3.05,
|
| 375 |
"learning_rate": 7.784444444444445e-06,
|
| 376 |
-
"loss": 0.
|
| 377 |
"step": 1500
|
| 378 |
},
|
| 379 |
{
|
| 380 |
"epoch": 3.06,
|
| 381 |
"learning_rate": 7.72888888888889e-06,
|
| 382 |
-
"loss": 0.
|
| 383 |
"step": 1525
|
| 384 |
},
|
| 385 |
{
|
| 386 |
"epoch": 3.06,
|
| 387 |
"learning_rate": 7.673333333333333e-06,
|
| 388 |
-
"loss": 0.
|
| 389 |
"step": 1550
|
| 390 |
},
|
| 391 |
{
|
| 392 |
"epoch": 3.07,
|
| 393 |
"learning_rate": 7.617777777777778e-06,
|
| 394 |
-
"loss": 0.
|
| 395 |
"step": 1575
|
| 396 |
},
|
| 397 |
{
|
| 398 |
"epoch": 3.07,
|
| 399 |
"learning_rate": 7.562222222222223e-06,
|
| 400 |
-
"loss": 0.
|
| 401 |
"step": 1600
|
| 402 |
},
|
| 403 |
{
|
| 404 |
"epoch": 3.08,
|
| 405 |
"learning_rate": 7.506666666666668e-06,
|
| 406 |
-
"loss": 0.
|
| 407 |
"step": 1625
|
| 408 |
},
|
| 409 |
{
|
| 410 |
"epoch": 4.0,
|
| 411 |
"learning_rate": 7.451111111111111e-06,
|
| 412 |
-
"loss": 0.
|
| 413 |
"step": 1650
|
| 414 |
},
|
| 415 |
{
|
| 416 |
"epoch": 4.01,
|
| 417 |
"learning_rate": 7.395555555555556e-06,
|
| 418 |
-
"loss": 0.
|
| 419 |
"step": 1675
|
| 420 |
},
|
| 421 |
{
|
| 422 |
"epoch": 4.01,
|
| 423 |
"learning_rate": 7.340000000000001e-06,
|
| 424 |
-
"loss": 0.
|
| 425 |
"step": 1700
|
| 426 |
},
|
| 427 |
{
|
| 428 |
"epoch": 4.02,
|
| 429 |
"learning_rate": 7.284444444444445e-06,
|
| 430 |
-
"loss": 0.
|
| 431 |
"step": 1725
|
| 432 |
},
|
| 433 |
{
|
| 434 |
"epoch": 4.02,
|
| 435 |
"learning_rate": 7.22888888888889e-06,
|
| 436 |
-
"loss": 0.
|
| 437 |
"step": 1750
|
| 438 |
},
|
| 439 |
{
|
| 440 |
"epoch": 4.03,
|
| 441 |
"learning_rate": 7.173333333333335e-06,
|
| 442 |
-
"loss": 0.
|
| 443 |
"step": 1775
|
| 444 |
},
|
| 445 |
{
|
| 446 |
"epoch": 4.03,
|
| 447 |
"learning_rate": 7.117777777777778e-06,
|
| 448 |
-
"loss": 0.
|
| 449 |
"step": 1800
|
| 450 |
},
|
| 451 |
{
|
| 452 |
"epoch": 4.04,
|
| 453 |
"learning_rate": 7.062222222222223e-06,
|
| 454 |
-
"loss": 0.
|
| 455 |
"step": 1825
|
| 456 |
},
|
| 457 |
{
|
| 458 |
"epoch": 4.04,
|
| 459 |
"learning_rate": 7.006666666666667e-06,
|
| 460 |
-
"loss": 0.
|
| 461 |
"step": 1850
|
| 462 |
},
|
| 463 |
{
|
| 464 |
"epoch": 4.05,
|
| 465 |
"learning_rate": 6.951111111111112e-06,
|
| 466 |
-
"loss": 0.
|
| 467 |
"step": 1875
|
| 468 |
},
|
| 469 |
{
|
| 470 |
"epoch": 4.05,
|
| 471 |
"learning_rate": 6.8955555555555565e-06,
|
| 472 |
-
"loss": 0.
|
| 473 |
"step": 1900
|
| 474 |
},
|
| 475 |
{
|
| 476 |
"epoch": 4.06,
|
| 477 |
"learning_rate": 6.8400000000000014e-06,
|
| 478 |
-
"loss": 0.
|
| 479 |
"step": 1925
|
| 480 |
},
|
| 481 |
{
|
| 482 |
"epoch": 4.06,
|
| 483 |
"learning_rate": 6.784444444444445e-06,
|
| 484 |
-
"loss": 0.
|
| 485 |
"step": 1950
|
| 486 |
},
|
| 487 |
{
|
| 488 |
"epoch": 4.07,
|
| 489 |
"learning_rate": 6.7288888888888895e-06,
|
| 490 |
-
"loss": 0.
|
| 491 |
"step": 1975
|
| 492 |
},
|
| 493 |
{
|
| 494 |
"epoch": 4.07,
|
| 495 |
"learning_rate": 6.6733333333333335e-06,
|
| 496 |
-
"loss": 0.
|
| 497 |
"step": 2000
|
| 498 |
},
|
| 499 |
{
|
| 500 |
"epoch": 4.07,
|
| 501 |
-
"eval_loss": 0.
|
| 502 |
-
"eval_runtime":
|
| 503 |
-
"eval_samples_per_second":
|
| 504 |
-
"eval_steps_per_second": 0.
|
| 505 |
-
"eval_wer":
|
| 506 |
"step": 2000
|
| 507 |
},
|
| 508 |
{
|
| 509 |
"epoch": 4.08,
|
| 510 |
"learning_rate": 6.617777777777778e-06,
|
| 511 |
-
"loss": 0.
|
| 512 |
"step": 2025
|
| 513 |
},
|
| 514 |
{
|
| 515 |
"epoch": 4.08,
|
| 516 |
"learning_rate": 6.562222222222223e-06,
|
| 517 |
-
"loss": 0.
|
| 518 |
"step": 2050
|
| 519 |
},
|
| 520 |
{
|
| 521 |
"epoch": 5.0,
|
| 522 |
"learning_rate": 6.5066666666666665e-06,
|
| 523 |
-
"loss": 0.
|
| 524 |
"step": 2075
|
| 525 |
},
|
| 526 |
{
|
| 527 |
"epoch": 5.01,
|
| 528 |
"learning_rate": 6.451111111111111e-06,
|
| 529 |
-
"loss": 0.
|
| 530 |
"step": 2100
|
| 531 |
},
|
| 532 |
{
|
| 533 |
"epoch": 5.01,
|
| 534 |
"learning_rate": 6.395555555555556e-06,
|
| 535 |
-
"loss": 0.
|
| 536 |
"step": 2125
|
| 537 |
},
|
| 538 |
{
|
| 539 |
"epoch": 5.02,
|
| 540 |
"learning_rate": 6.34e-06,
|
| 541 |
-
"loss": 0.
|
| 542 |
"step": 2150
|
| 543 |
},
|
| 544 |
{
|
| 545 |
"epoch": 5.02,
|
| 546 |
"learning_rate": 6.284444444444445e-06,
|
| 547 |
-
"loss": 0.
|
| 548 |
"step": 2175
|
| 549 |
},
|
| 550 |
{
|
| 551 |
"epoch": 5.03,
|
| 552 |
"learning_rate": 6.22888888888889e-06,
|
| 553 |
-
"loss": 0.
|
| 554 |
"step": 2200
|
| 555 |
},
|
| 556 |
{
|
| 557 |
"epoch": 5.03,
|
| 558 |
"learning_rate": 6.173333333333333e-06,
|
| 559 |
-
"loss": 0.
|
| 560 |
"step": 2225
|
| 561 |
},
|
| 562 |
{
|
| 563 |
"epoch": 5.04,
|
| 564 |
"learning_rate": 6.117777777777778e-06,
|
| 565 |
-
"loss": 0.
|
| 566 |
"step": 2250
|
| 567 |
},
|
| 568 |
{
|
| 569 |
"epoch": 5.04,
|
| 570 |
"learning_rate": 6.062222222222223e-06,
|
| 571 |
-
"loss": 0.
|
| 572 |
"step": 2275
|
| 573 |
},
|
| 574 |
{
|
| 575 |
"epoch": 5.05,
|
| 576 |
"learning_rate": 6.006666666666667e-06,
|
| 577 |
-
"loss": 0.
|
| 578 |
"step": 2300
|
| 579 |
},
|
| 580 |
{
|
| 581 |
"epoch": 5.05,
|
| 582 |
"learning_rate": 5.951111111111112e-06,
|
| 583 |
-
"loss": 0.
|
| 584 |
"step": 2325
|
| 585 |
},
|
| 586 |
{
|
| 587 |
"epoch": 5.06,
|
| 588 |
"learning_rate": 5.895555555555557e-06,
|
| 589 |
-
"loss": 0.
|
| 590 |
"step": 2350
|
| 591 |
},
|
| 592 |
{
|
| 593 |
"epoch": 5.06,
|
| 594 |
"learning_rate": 5.84e-06,
|
| 595 |
-
"loss": 0.
|
| 596 |
"step": 2375
|
| 597 |
},
|
| 598 |
{
|
| 599 |
"epoch": 5.07,
|
| 600 |
"learning_rate": 5.784444444444445e-06,
|
| 601 |
-
"loss": 0.
|
| 602 |
"step": 2400
|
| 603 |
},
|
| 604 |
{
|
| 605 |
"epoch": 5.07,
|
| 606 |
"learning_rate": 5.72888888888889e-06,
|
| 607 |
-
"loss": 0.
|
| 608 |
"step": 2425
|
| 609 |
},
|
| 610 |
{
|
| 611 |
"epoch": 5.08,
|
| 612 |
"learning_rate": 5.673333333333334e-06,
|
| 613 |
-
"loss": 0.
|
| 614 |
"step": 2450
|
| 615 |
},
|
| 616 |
{
|
| 617 |
"epoch": 6.0,
|
| 618 |
"learning_rate": 5.617777777777779e-06,
|
| 619 |
-
"loss": 0.
|
| 620 |
"step": 2475
|
| 621 |
},
|
| 622 |
{
|
| 623 |
"epoch": 6.01,
|
| 624 |
"learning_rate": 5.562222222222222e-06,
|
| 625 |
-
"loss": 0.
|
| 626 |
"step": 2500
|
| 627 |
},
|
| 628 |
{
|
| 629 |
"epoch": 6.01,
|
| 630 |
"learning_rate": 5.506666666666667e-06,
|
| 631 |
-
"loss": 0.
|
| 632 |
"step": 2525
|
| 633 |
},
|
| 634 |
{
|
| 635 |
"epoch": 6.02,
|
| 636 |
"learning_rate": 5.451111111111112e-06,
|
| 637 |
-
"loss": 0.
|
| 638 |
"step": 2550
|
| 639 |
},
|
| 640 |
{
|
| 641 |
"epoch": 6.02,
|
| 642 |
"learning_rate": 5.3955555555555565e-06,
|
| 643 |
-
"loss": 0.
|
| 644 |
"step": 2575
|
| 645 |
},
|
| 646 |
{
|
| 647 |
"epoch": 6.03,
|
| 648 |
"learning_rate": 5.3400000000000005e-06,
|
| 649 |
-
"loss": 0.
|
| 650 |
"step": 2600
|
| 651 |
},
|
| 652 |
{
|
| 653 |
"epoch": 6.03,
|
| 654 |
"learning_rate": 5.2844444444444454e-06,
|
| 655 |
-
"loss": 0.
|
| 656 |
"step": 2625
|
| 657 |
},
|
| 658 |
{
|
| 659 |
"epoch": 6.04,
|
| 660 |
"learning_rate": 5.228888888888889e-06,
|
| 661 |
-
"loss": 0.
|
| 662 |
"step": 2650
|
| 663 |
},
|
| 664 |
{
|
| 665 |
"epoch": 6.04,
|
| 666 |
"learning_rate": 5.1733333333333335e-06,
|
| 667 |
-
"loss": 0.
|
| 668 |
"step": 2675
|
| 669 |
},
|
| 670 |
{
|
| 671 |
"epoch": 6.05,
|
| 672 |
"learning_rate": 5.117777777777778e-06,
|
| 673 |
-
"loss": 0.
|
| 674 |
"step": 2700
|
| 675 |
},
|
| 676 |
{
|
| 677 |
"epoch": 6.05,
|
| 678 |
"learning_rate": 5.062222222222222e-06,
|
| 679 |
-
"loss": 0.
|
| 680 |
"step": 2725
|
| 681 |
},
|
| 682 |
{
|
| 683 |
"epoch": 6.06,
|
| 684 |
"learning_rate": 5.006666666666667e-06,
|
| 685 |
-
"loss": 0.
|
| 686 |
"step": 2750
|
| 687 |
},
|
| 688 |
{
|
| 689 |
"epoch": 6.06,
|
| 690 |
"learning_rate": 4.951111111111111e-06,
|
| 691 |
-
"loss": 0.
|
| 692 |
"step": 2775
|
| 693 |
},
|
| 694 |
{
|
|
@@ -700,94 +700,94 @@
|
|
| 700 |
{
|
| 701 |
"epoch": 6.07,
|
| 702 |
"learning_rate": 4.84e-06,
|
| 703 |
-
"loss": 0.
|
| 704 |
"step": 2825
|
| 705 |
},
|
| 706 |
{
|
| 707 |
"epoch": 6.08,
|
| 708 |
"learning_rate": 4.784444444444445e-06,
|
| 709 |
-
"loss": 0.
|
| 710 |
"step": 2850
|
| 711 |
},
|
| 712 |
{
|
| 713 |
"epoch": 6.08,
|
| 714 |
"learning_rate": 4.728888888888889e-06,
|
| 715 |
-
"loss": 0.
|
| 716 |
"step": 2875
|
| 717 |
},
|
| 718 |
{
|
| 719 |
"epoch": 7.0,
|
| 720 |
"learning_rate": 4.673333333333333e-06,
|
| 721 |
-
"loss": 0.
|
| 722 |
"step": 2900
|
| 723 |
},
|
| 724 |
{
|
| 725 |
"epoch": 7.01,
|
| 726 |
"learning_rate": 4.617777777777778e-06,
|
| 727 |
-
"loss": 0.
|
| 728 |
"step": 2925
|
| 729 |
},
|
| 730 |
{
|
| 731 |
"epoch": 7.01,
|
| 732 |
"learning_rate": 4.562222222222222e-06,
|
| 733 |
-
"loss": 0.
|
| 734 |
"step": 2950
|
| 735 |
},
|
| 736 |
{
|
| 737 |
"epoch": 7.02,
|
| 738 |
"learning_rate": 4.506666666666667e-06,
|
| 739 |
-
"loss": 0.
|
| 740 |
"step": 2975
|
| 741 |
},
|
| 742 |
{
|
| 743 |
"epoch": 7.02,
|
| 744 |
"learning_rate": 4.451111111111112e-06,
|
| 745 |
-
"loss": 0.
|
| 746 |
"step": 3000
|
| 747 |
},
|
| 748 |
{
|
| 749 |
"epoch": 7.02,
|
| 750 |
-
"eval_loss": 0.
|
| 751 |
-
"eval_runtime":
|
| 752 |
-
"eval_samples_per_second":
|
| 753 |
-
"eval_steps_per_second": 0.
|
| 754 |
-
"eval_wer": 18.
|
| 755 |
"step": 3000
|
| 756 |
},
|
| 757 |
{
|
| 758 |
"epoch": 7.03,
|
| 759 |
"learning_rate": 4.395555555555556e-06,
|
| 760 |
-
"loss": 0.
|
| 761 |
"step": 3025
|
| 762 |
},
|
| 763 |
{
|
| 764 |
"epoch": 7.03,
|
| 765 |
"learning_rate": 4.34e-06,
|
| 766 |
-
"loss": 0.
|
| 767 |
"step": 3050
|
| 768 |
},
|
| 769 |
{
|
| 770 |
"epoch": 7.04,
|
| 771 |
"learning_rate": 4.284444444444445e-06,
|
| 772 |
-
"loss": 0.
|
| 773 |
"step": 3075
|
| 774 |
},
|
| 775 |
{
|
| 776 |
"epoch": 7.04,
|
| 777 |
"learning_rate": 4.228888888888889e-06,
|
| 778 |
-
"loss": 0.
|
| 779 |
"step": 3100
|
| 780 |
},
|
| 781 |
{
|
| 782 |
"epoch": 7.05,
|
| 783 |
"learning_rate": 4.173333333333334e-06,
|
| 784 |
-
"loss": 0.
|
| 785 |
"step": 3125
|
| 786 |
},
|
| 787 |
{
|
| 788 |
"epoch": 7.05,
|
| 789 |
"learning_rate": 4.117777777777779e-06,
|
| 790 |
-
"loss": 0.
|
| 791 |
"step": 3150
|
| 792 |
},
|
| 793 |
{
|
|
@@ -805,37 +805,37 @@
|
|
| 805 |
{
|
| 806 |
"epoch": 7.07,
|
| 807 |
"learning_rate": 3.951111111111112e-06,
|
| 808 |
-
"loss": 0.
|
| 809 |
"step": 3225
|
| 810 |
},
|
| 811 |
{
|
| 812 |
"epoch": 7.07,
|
| 813 |
"learning_rate": 3.895555555555556e-06,
|
| 814 |
-
"loss": 0.
|
| 815 |
"step": 3250
|
| 816 |
},
|
| 817 |
{
|
| 818 |
"epoch": 7.08,
|
| 819 |
"learning_rate": 3.8400000000000005e-06,
|
| 820 |
-
"loss": 0.
|
| 821 |
"step": 3275
|
| 822 |
},
|
| 823 |
{
|
| 824 |
"epoch": 8.0,
|
| 825 |
"learning_rate": 3.784444444444445e-06,
|
| 826 |
-
"loss": 0.
|
| 827 |
"step": 3300
|
| 828 |
},
|
| 829 |
{
|
| 830 |
"epoch": 8.01,
|
| 831 |
"learning_rate": 3.728888888888889e-06,
|
| 832 |
-
"loss": 0.
|
| 833 |
"step": 3325
|
| 834 |
},
|
| 835 |
{
|
| 836 |
"epoch": 8.01,
|
| 837 |
"learning_rate": 3.673333333333334e-06,
|
| 838 |
-
"loss": 0.
|
| 839 |
"step": 3350
|
| 840 |
},
|
| 841 |
{
|
|
@@ -847,13 +847,13 @@
|
|
| 847 |
{
|
| 848 |
"epoch": 8.02,
|
| 849 |
"learning_rate": 3.5622222222222224e-06,
|
| 850 |
-
"loss": 0.
|
| 851 |
"step": 3400
|
| 852 |
},
|
| 853 |
{
|
| 854 |
"epoch": 8.03,
|
| 855 |
"learning_rate": 3.5066666666666673e-06,
|
| 856 |
-
"loss": 0.
|
| 857 |
"step": 3425
|
| 858 |
},
|
| 859 |
{
|
|
@@ -865,7 +865,7 @@
|
|
| 865 |
{
|
| 866 |
"epoch": 8.04,
|
| 867 |
"learning_rate": 3.3955555555555558e-06,
|
| 868 |
-
"loss": 0.
|
| 869 |
"step": 3475
|
| 870 |
},
|
| 871 |
{
|
|
@@ -877,101 +877,101 @@
|
|
| 877 |
{
|
| 878 |
"epoch": 8.05,
|
| 879 |
"learning_rate": 3.2844444444444447e-06,
|
| 880 |
-
"loss": 0.
|
| 881 |
"step": 3525
|
| 882 |
},
|
| 883 |
{
|
| 884 |
"epoch": 8.05,
|
| 885 |
"learning_rate": 3.228888888888889e-06,
|
| 886 |
-
"loss": 0.
|
| 887 |
"step": 3550
|
| 888 |
},
|
| 889 |
{
|
| 890 |
"epoch": 8.06,
|
| 891 |
"learning_rate": 3.173333333333334e-06,
|
| 892 |
-
"loss": 0.
|
| 893 |
"step": 3575
|
| 894 |
},
|
| 895 |
{
|
| 896 |
"epoch": 8.06,
|
| 897 |
"learning_rate": 3.117777777777778e-06,
|
| 898 |
-
"loss": 0.
|
| 899 |
"step": 3600
|
| 900 |
},
|
| 901 |
{
|
| 902 |
"epoch": 8.07,
|
| 903 |
"learning_rate": 3.0622222222222225e-06,
|
| 904 |
-
"loss": 0.
|
| 905 |
"step": 3625
|
| 906 |
},
|
| 907 |
{
|
| 908 |
"epoch": 8.07,
|
| 909 |
"learning_rate": 3.0066666666666674e-06,
|
| 910 |
-
"loss": 0.
|
| 911 |
"step": 3650
|
| 912 |
},
|
| 913 |
{
|
| 914 |
"epoch": 8.08,
|
| 915 |
"learning_rate": 2.9511111111111114e-06,
|
| 916 |
-
"loss": 0.
|
| 917 |
"step": 3675
|
| 918 |
},
|
| 919 |
{
|
| 920 |
-
"epoch":
|
| 921 |
"learning_rate": 2.895555555555556e-06,
|
| 922 |
"loss": 0.0019,
|
| 923 |
"step": 3700
|
| 924 |
},
|
| 925 |
{
|
| 926 |
-
"epoch": 9.
|
| 927 |
"learning_rate": 2.84e-06,
|
| 928 |
-
"loss": 0.
|
| 929 |
"step": 3725
|
| 930 |
},
|
| 931 |
{
|
| 932 |
"epoch": 9.01,
|
| 933 |
"learning_rate": 2.784444444444445e-06,
|
| 934 |
-
"loss": 0.
|
| 935 |
"step": 3750
|
| 936 |
},
|
| 937 |
{
|
| 938 |
-
"epoch": 9.
|
| 939 |
"learning_rate": 2.7288888888888893e-06,
|
| 940 |
-
"loss": 0.
|
| 941 |
"step": 3775
|
| 942 |
},
|
| 943 |
{
|
| 944 |
"epoch": 9.02,
|
| 945 |
"learning_rate": 2.6733333333333333e-06,
|
| 946 |
-
"loss": 0.
|
| 947 |
"step": 3800
|
| 948 |
},
|
| 949 |
{
|
| 950 |
-
"epoch": 9.
|
| 951 |
"learning_rate": 2.617777777777778e-06,
|
| 952 |
-
"loss": 0.
|
| 953 |
"step": 3825
|
| 954 |
},
|
| 955 |
{
|
| 956 |
"epoch": 9.03,
|
| 957 |
"learning_rate": 2.5622222222222226e-06,
|
| 958 |
-
"loss": 0.
|
| 959 |
"step": 3850
|
| 960 |
},
|
| 961 |
{
|
| 962 |
-
"epoch": 9.
|
| 963 |
"learning_rate": 2.5066666666666667e-06,
|
| 964 |
-
"loss": 0.
|
| 965 |
"step": 3875
|
| 966 |
},
|
| 967 |
{
|
| 968 |
"epoch": 9.04,
|
| 969 |
"learning_rate": 2.451111111111111e-06,
|
| 970 |
-
"loss": 0.
|
| 971 |
"step": 3900
|
| 972 |
},
|
| 973 |
{
|
| 974 |
-
"epoch": 9.
|
| 975 |
"learning_rate": 2.3955555555555556e-06,
|
| 976 |
"loss": 0.0017,
|
| 977 |
"step": 3925
|
|
@@ -979,11 +979,11 @@
|
|
| 979 |
{
|
| 980 |
"epoch": 9.05,
|
| 981 |
"learning_rate": 2.3400000000000005e-06,
|
| 982 |
-
"loss": 0.
|
| 983 |
"step": 3950
|
| 984 |
},
|
| 985 |
{
|
| 986 |
-
"epoch": 9.
|
| 987 |
"learning_rate": 2.2844444444444445e-06,
|
| 988 |
"loss": 0.0017,
|
| 989 |
"step": 3975
|
|
@@ -991,20 +991,20 @@
|
|
| 991 |
{
|
| 992 |
"epoch": 9.06,
|
| 993 |
"learning_rate": 2.228888888888889e-06,
|
| 994 |
-
"loss": 0.
|
| 995 |
"step": 4000
|
| 996 |
},
|
| 997 |
{
|
| 998 |
"epoch": 9.06,
|
| 999 |
-
"eval_loss": 0.
|
| 1000 |
-
"eval_runtime":
|
| 1001 |
-
"eval_samples_per_second":
|
| 1002 |
-
"eval_steps_per_second": 0.
|
| 1003 |
-
"eval_wer": 18.
|
| 1004 |
"step": 4000
|
| 1005 |
},
|
| 1006 |
{
|
| 1007 |
-
"epoch": 9.
|
| 1008 |
"learning_rate": 2.1733333333333334e-06,
|
| 1009 |
"loss": 0.0016,
|
| 1010 |
"step": 4025
|
|
@@ -1016,7 +1016,7 @@
|
|
| 1016 |
"step": 4050
|
| 1017 |
},
|
| 1018 |
{
|
| 1019 |
-
"epoch": 9.
|
| 1020 |
"learning_rate": 2.0622222222222223e-06,
|
| 1021 |
"loss": 0.0015,
|
| 1022 |
"step": 4075
|
|
@@ -1030,7 +1030,7 @@
|
|
| 1030 |
{
|
| 1031 |
"epoch": 10.0,
|
| 1032 |
"learning_rate": 1.9511111111111113e-06,
|
| 1033 |
-
"loss": 0.
|
| 1034 |
"step": 4125
|
| 1035 |
},
|
| 1036 |
{
|
|
@@ -1054,7 +1054,7 @@
|
|
| 1054 |
{
|
| 1055 |
"epoch": 10.02,
|
| 1056 |
"learning_rate": 1.728888888888889e-06,
|
| 1057 |
-
"loss": 0.
|
| 1058 |
"step": 4225
|
| 1059 |
},
|
| 1060 |
{
|
|
@@ -1066,7 +1066,7 @@
|
|
| 1066 |
{
|
| 1067 |
"epoch": 10.03,
|
| 1068 |
"learning_rate": 1.6177777777777778e-06,
|
| 1069 |
-
"loss": 0.
|
| 1070 |
"step": 4275
|
| 1071 |
},
|
| 1072 |
{
|
|
@@ -1078,7 +1078,7 @@
|
|
| 1078 |
{
|
| 1079 |
"epoch": 10.04,
|
| 1080 |
"learning_rate": 1.506666666666667e-06,
|
| 1081 |
-
"loss": 0.
|
| 1082 |
"step": 4325
|
| 1083 |
},
|
| 1084 |
{
|
|
@@ -1124,13 +1124,13 @@
|
|
| 1124 |
"step": 4500
|
| 1125 |
},
|
| 1126 |
{
|
| 1127 |
-
"epoch":
|
| 1128 |
"learning_rate": 1.0622222222222222e-06,
|
| 1129 |
-
"loss": 0.
|
| 1130 |
"step": 4525
|
| 1131 |
},
|
| 1132 |
{
|
| 1133 |
-
"epoch": 11.
|
| 1134 |
"learning_rate": 1.0066666666666668e-06,
|
| 1135 |
"loss": 0.0014,
|
| 1136 |
"step": 4550
|
|
@@ -1138,11 +1138,11 @@
|
|
| 1138 |
{
|
| 1139 |
"epoch": 11.01,
|
| 1140 |
"learning_rate": 9.511111111111111e-07,
|
| 1141 |
-
"loss": 0.
|
| 1142 |
"step": 4575
|
| 1143 |
},
|
| 1144 |
{
|
| 1145 |
-
"epoch": 11.
|
| 1146 |
"learning_rate": 8.955555555555557e-07,
|
| 1147 |
"loss": 0.0013,
|
| 1148 |
"step": 4600
|
|
@@ -1154,19 +1154,19 @@
|
|
| 1154 |
"step": 4625
|
| 1155 |
},
|
| 1156 |
{
|
| 1157 |
-
"epoch": 11.
|
| 1158 |
"learning_rate": 7.844444444444445e-07,
|
| 1159 |
-
"loss": 0.
|
| 1160 |
"step": 4650
|
| 1161 |
},
|
| 1162 |
{
|
| 1163 |
"epoch": 11.03,
|
| 1164 |
"learning_rate": 7.28888888888889e-07,
|
| 1165 |
-
"loss": 0.
|
| 1166 |
"step": 4675
|
| 1167 |
},
|
| 1168 |
{
|
| 1169 |
-
"epoch": 11.
|
| 1170 |
"learning_rate": 6.733333333333334e-07,
|
| 1171 |
"loss": 0.0013,
|
| 1172 |
"step": 4700
|
|
@@ -1178,7 +1178,7 @@
|
|
| 1178 |
"step": 4725
|
| 1179 |
},
|
| 1180 |
{
|
| 1181 |
-
"epoch": 11.
|
| 1182 |
"learning_rate": 5.622222222222223e-07,
|
| 1183 |
"loss": 0.0013,
|
| 1184 |
"step": 4750
|
|
@@ -1190,7 +1190,7 @@
|
|
| 1190 |
"step": 4775
|
| 1191 |
},
|
| 1192 |
{
|
| 1193 |
-
"epoch": 11.
|
| 1194 |
"learning_rate": 4.511111111111111e-07,
|
| 1195 |
"loss": 0.0013,
|
| 1196 |
"step": 4800
|
|
@@ -1202,19 +1202,19 @@
|
|
| 1202 |
"step": 4825
|
| 1203 |
},
|
| 1204 |
{
|
| 1205 |
-
"epoch": 11.
|
| 1206 |
"learning_rate": 3.4000000000000003e-07,
|
| 1207 |
-
"loss": 0.
|
| 1208 |
"step": 4850
|
| 1209 |
},
|
| 1210 |
{
|
| 1211 |
"epoch": 11.07,
|
| 1212 |
"learning_rate": 2.844444444444445e-07,
|
| 1213 |
-
"loss": 0.
|
| 1214 |
"step": 4875
|
| 1215 |
},
|
| 1216 |
{
|
| 1217 |
-
"epoch": 11.
|
| 1218 |
"learning_rate": 2.2888888888888892e-07,
|
| 1219 |
"loss": 0.0012,
|
| 1220 |
"step": 4900
|
|
@@ -1222,7 +1222,7 @@
|
|
| 1222 |
{
|
| 1223 |
"epoch": 11.08,
|
| 1224 |
"learning_rate": 1.7333333333333335e-07,
|
| 1225 |
-
"loss": 0.
|
| 1226 |
"step": 4925
|
| 1227 |
},
|
| 1228 |
{
|
|
@@ -1240,31 +1240,31 @@
|
|
| 1240 |
{
|
| 1241 |
"epoch": 12.01,
|
| 1242 |
"learning_rate": 6.666666666666667e-09,
|
| 1243 |
-
"loss": 0.
|
| 1244 |
"step": 5000
|
| 1245 |
},
|
| 1246 |
{
|
| 1247 |
"epoch": 12.01,
|
| 1248 |
-
"eval_loss": 0.
|
| 1249 |
-
"eval_runtime":
|
| 1250 |
-
"eval_samples_per_second":
|
| 1251 |
-
"eval_steps_per_second": 0.
|
| 1252 |
-
"eval_wer": 18.
|
| 1253 |
"step": 5000
|
| 1254 |
},
|
| 1255 |
{
|
| 1256 |
"epoch": 12.01,
|
| 1257 |
"step": 5000,
|
| 1258 |
-
"total_flos": 4.
|
| 1259 |
-
"train_loss": 0.
|
| 1260 |
-
"train_runtime":
|
| 1261 |
-
"train_samples_per_second": 5.
|
| 1262 |
-
"train_steps_per_second": 0.
|
| 1263 |
}
|
| 1264 |
],
|
| 1265 |
"max_steps": 5000,
|
| 1266 |
"num_train_epochs": 9223372036854775807,
|
| 1267 |
-
"total_flos": 4.
|
| 1268 |
"trial_name": null,
|
| 1269 |
"trial_params": null
|
| 1270 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 18.775568066750374,
|
| 3 |
+
"best_model_checkpoint": "./checkpoint-4000",
|
| 4 |
+
"epoch": 12.0112,
|
| 5 |
"global_step": 5000,
|
| 6 |
"is_hyper_param_search": false,
|
| 7 |
"is_local_process_zero": true,
|
|
|
|
| 10 |
{
|
| 11 |
"epoch": 0.01,
|
| 12 |
"learning_rate": 4.4e-07,
|
| 13 |
+
"loss": 1.6639,
|
| 14 |
"step": 25
|
| 15 |
},
|
| 16 |
{
|
| 17 |
"epoch": 0.01,
|
| 18 |
"learning_rate": 9.400000000000001e-07,
|
| 19 |
+
"loss": 1.4728,
|
| 20 |
"step": 50
|
| 21 |
},
|
| 22 |
{
|
| 23 |
"epoch": 0.01,
|
| 24 |
"learning_rate": 1.44e-06,
|
| 25 |
+
"loss": 1.1563,
|
| 26 |
"step": 75
|
| 27 |
},
|
| 28 |
{
|
| 29 |
"epoch": 0.02,
|
| 30 |
"learning_rate": 1.94e-06,
|
| 31 |
+
"loss": 0.9368,
|
| 32 |
"step": 100
|
| 33 |
},
|
| 34 |
{
|
| 35 |
"epoch": 0.03,
|
| 36 |
"learning_rate": 2.4400000000000004e-06,
|
| 37 |
+
"loss": 0.8889,
|
| 38 |
"step": 125
|
| 39 |
},
|
| 40 |
{
|
| 41 |
"epoch": 0.03,
|
| 42 |
"learning_rate": 2.9400000000000002e-06,
|
| 43 |
+
"loss": 0.8114,
|
| 44 |
"step": 150
|
| 45 |
},
|
| 46 |
{
|
| 47 |
"epoch": 0.04,
|
| 48 |
"learning_rate": 3.44e-06,
|
| 49 |
+
"loss": 0.6924,
|
| 50 |
"step": 175
|
| 51 |
},
|
| 52 |
{
|
| 53 |
"epoch": 0.04,
|
| 54 |
"learning_rate": 3.94e-06,
|
| 55 |
+
"loss": 0.661,
|
| 56 |
"step": 200
|
| 57 |
},
|
| 58 |
{
|
| 59 |
"epoch": 0.04,
|
| 60 |
"learning_rate": 4.440000000000001e-06,
|
| 61 |
+
"loss": 0.6427,
|
| 62 |
"step": 225
|
| 63 |
},
|
| 64 |
{
|
| 65 |
"epoch": 0.05,
|
| 66 |
"learning_rate": 4.94e-06,
|
| 67 |
+
"loss": 0.5217,
|
| 68 |
"step": 250
|
| 69 |
},
|
| 70 |
{
|
| 71 |
"epoch": 0.06,
|
| 72 |
"learning_rate": 5.4400000000000004e-06,
|
| 73 |
+
"loss": 0.4991,
|
| 74 |
"step": 275
|
| 75 |
},
|
| 76 |
{
|
| 77 |
"epoch": 0.06,
|
| 78 |
"learning_rate": 5.94e-06,
|
| 79 |
+
"loss": 0.4768,
|
| 80 |
"step": 300
|
| 81 |
},
|
| 82 |
{
|
| 83 |
"epoch": 0.07,
|
| 84 |
"learning_rate": 6.440000000000001e-06,
|
| 85 |
+
"loss": 0.4632,
|
| 86 |
"step": 325
|
| 87 |
},
|
| 88 |
{
|
| 89 |
"epoch": 0.07,
|
| 90 |
"learning_rate": 6.9400000000000005e-06,
|
| 91 |
+
"loss": 0.4289,
|
| 92 |
"step": 350
|
| 93 |
},
|
| 94 |
{
|
| 95 |
"epoch": 0.07,
|
| 96 |
"learning_rate": 7.440000000000001e-06,
|
| 97 |
+
"loss": 0.404,
|
| 98 |
"step": 375
|
| 99 |
},
|
| 100 |
{
|
| 101 |
"epoch": 0.08,
|
| 102 |
"learning_rate": 7.94e-06,
|
| 103 |
+
"loss": 0.3833,
|
| 104 |
"step": 400
|
| 105 |
},
|
| 106 |
{
|
| 107 |
"epoch": 1.0,
|
| 108 |
"learning_rate": 8.44e-06,
|
| 109 |
+
"loss": 0.4201,
|
| 110 |
"step": 425
|
| 111 |
},
|
| 112 |
{
|
| 113 |
"epoch": 1.01,
|
| 114 |
"learning_rate": 8.94e-06,
|
| 115 |
+
"loss": 0.3989,
|
| 116 |
"step": 450
|
| 117 |
},
|
| 118 |
{
|
| 119 |
"epoch": 1.01,
|
| 120 |
"learning_rate": 9.440000000000001e-06,
|
| 121 |
+
"loss": 0.3453,
|
| 122 |
"step": 475
|
| 123 |
},
|
| 124 |
{
|
| 125 |
"epoch": 1.02,
|
| 126 |
"learning_rate": 9.940000000000001e-06,
|
| 127 |
+
"loss": 0.3559,
|
| 128 |
"step": 500
|
| 129 |
},
|
| 130 |
{
|
| 131 |
"epoch": 1.02,
|
| 132 |
"learning_rate": 9.951111111111111e-06,
|
| 133 |
+
"loss": 0.3429,
|
| 134 |
"step": 525
|
| 135 |
},
|
| 136 |
{
|
| 137 |
"epoch": 1.03,
|
| 138 |
"learning_rate": 9.895555555555557e-06,
|
| 139 |
+
"loss": 0.3701,
|
| 140 |
"step": 550
|
| 141 |
},
|
| 142 |
{
|
| 143 |
"epoch": 1.03,
|
| 144 |
"learning_rate": 9.84e-06,
|
| 145 |
+
"loss": 0.3084,
|
| 146 |
"step": 575
|
| 147 |
},
|
| 148 |
{
|
| 149 |
"epoch": 1.04,
|
| 150 |
"learning_rate": 9.784444444444445e-06,
|
| 151 |
+
"loss": 0.2956,
|
| 152 |
"step": 600
|
| 153 |
},
|
| 154 |
{
|
| 155 |
"epoch": 1.04,
|
| 156 |
"learning_rate": 9.72888888888889e-06,
|
| 157 |
+
"loss": 0.2923,
|
| 158 |
"step": 625
|
| 159 |
},
|
| 160 |
{
|
| 161 |
"epoch": 1.05,
|
| 162 |
"learning_rate": 9.673333333333334e-06,
|
| 163 |
+
"loss": 0.2786,
|
| 164 |
"step": 650
|
| 165 |
},
|
| 166 |
{
|
| 167 |
"epoch": 1.05,
|
| 168 |
"learning_rate": 9.617777777777778e-06,
|
| 169 |
+
"loss": 0.2427,
|
| 170 |
"step": 675
|
| 171 |
},
|
| 172 |
{
|
| 173 |
"epoch": 1.06,
|
| 174 |
"learning_rate": 9.562222222222223e-06,
|
| 175 |
+
"loss": 0.2387,
|
| 176 |
"step": 700
|
| 177 |
},
|
| 178 |
{
|
| 179 |
"epoch": 1.06,
|
| 180 |
"learning_rate": 9.506666666666667e-06,
|
| 181 |
+
"loss": 0.2371,
|
| 182 |
"step": 725
|
| 183 |
},
|
| 184 |
{
|
| 185 |
"epoch": 1.07,
|
| 186 |
"learning_rate": 9.451111111111112e-06,
|
| 187 |
+
"loss": 0.2114,
|
| 188 |
"step": 750
|
| 189 |
},
|
| 190 |
{
|
| 191 |
"epoch": 1.07,
|
| 192 |
"learning_rate": 9.395555555555556e-06,
|
| 193 |
+
"loss": 0.1894,
|
| 194 |
"step": 775
|
| 195 |
},
|
| 196 |
{
|
| 197 |
"epoch": 1.08,
|
| 198 |
"learning_rate": 9.340000000000002e-06,
|
| 199 |
+
"loss": 0.1867,
|
| 200 |
"step": 800
|
| 201 |
},
|
| 202 |
{
|
| 203 |
"epoch": 2.0,
|
| 204 |
"learning_rate": 9.284444444444444e-06,
|
| 205 |
+
"loss": 0.1927,
|
| 206 |
"step": 825
|
| 207 |
},
|
| 208 |
{
|
| 209 |
"epoch": 2.01,
|
| 210 |
"learning_rate": 9.22888888888889e-06,
|
| 211 |
+
"loss": 0.1951,
|
| 212 |
"step": 850
|
| 213 |
},
|
| 214 |
{
|
| 215 |
"epoch": 2.01,
|
| 216 |
"learning_rate": 9.173333333333334e-06,
|
| 217 |
+
"loss": 0.1775,
|
| 218 |
"step": 875
|
| 219 |
},
|
| 220 |
{
|
| 221 |
"epoch": 2.02,
|
| 222 |
"learning_rate": 9.117777777777778e-06,
|
| 223 |
+
"loss": 0.1672,
|
| 224 |
"step": 900
|
| 225 |
},
|
| 226 |
{
|
| 227 |
"epoch": 2.02,
|
| 228 |
"learning_rate": 9.062222222222224e-06,
|
| 229 |
+
"loss": 0.1591,
|
| 230 |
"step": 925
|
| 231 |
},
|
| 232 |
{
|
| 233 |
"epoch": 2.03,
|
| 234 |
"learning_rate": 9.006666666666666e-06,
|
| 235 |
+
"loss": 0.1703,
|
| 236 |
"step": 950
|
| 237 |
},
|
| 238 |
{
|
| 239 |
"epoch": 2.03,
|
| 240 |
"learning_rate": 8.951111111111112e-06,
|
| 241 |
+
"loss": 0.1636,
|
| 242 |
"step": 975
|
| 243 |
},
|
| 244 |
{
|
| 245 |
"epoch": 2.04,
|
| 246 |
"learning_rate": 8.895555555555556e-06,
|
| 247 |
+
"loss": 0.1413,
|
| 248 |
"step": 1000
|
| 249 |
},
|
| 250 |
{
|
| 251 |
"epoch": 2.04,
|
| 252 |
+
"eval_loss": 0.31775403022766113,
|
| 253 |
+
"eval_runtime": 963.5028,
|
| 254 |
+
"eval_samples_per_second": 6.841,
|
| 255 |
+
"eval_steps_per_second": 0.428,
|
| 256 |
+
"eval_wer": 22.013852322896838,
|
| 257 |
"step": 1000
|
| 258 |
},
|
| 259 |
{
|
| 260 |
"epoch": 2.04,
|
| 261 |
"learning_rate": 8.84e-06,
|
| 262 |
+
"loss": 0.1371,
|
| 263 |
"step": 1025
|
| 264 |
},
|
| 265 |
{
|
| 266 |
"epoch": 2.05,
|
| 267 |
"learning_rate": 8.784444444444446e-06,
|
| 268 |
+
"loss": 0.1494,
|
| 269 |
"step": 1050
|
| 270 |
},
|
| 271 |
{
|
| 272 |
"epoch": 2.05,
|
| 273 |
"learning_rate": 8.72888888888889e-06,
|
| 274 |
+
"loss": 0.1196,
|
| 275 |
"step": 1075
|
| 276 |
},
|
| 277 |
{
|
| 278 |
"epoch": 2.06,
|
| 279 |
"learning_rate": 8.673333333333334e-06,
|
| 280 |
+
"loss": 0.1152,
|
| 281 |
"step": 1100
|
| 282 |
},
|
| 283 |
{
|
| 284 |
"epoch": 2.06,
|
| 285 |
"learning_rate": 8.617777777777778e-06,
|
| 286 |
+
"loss": 0.1158,
|
| 287 |
"step": 1125
|
| 288 |
},
|
| 289 |
{
|
| 290 |
"epoch": 2.07,
|
| 291 |
"learning_rate": 8.562222222222224e-06,
|
| 292 |
+
"loss": 0.1054,
|
| 293 |
"step": 1150
|
| 294 |
},
|
| 295 |
{
|
| 296 |
"epoch": 2.07,
|
| 297 |
"learning_rate": 8.506666666666668e-06,
|
| 298 |
+
"loss": 0.0912,
|
| 299 |
"step": 1175
|
| 300 |
},
|
| 301 |
{
|
| 302 |
"epoch": 2.08,
|
| 303 |
"learning_rate": 8.451111111111112e-06,
|
| 304 |
+
"loss": 0.0915,
|
| 305 |
"step": 1200
|
| 306 |
},
|
| 307 |
{
|
| 308 |
"epoch": 2.08,
|
| 309 |
"learning_rate": 8.395555555555557e-06,
|
| 310 |
+
"loss": 0.0901,
|
| 311 |
"step": 1225
|
| 312 |
},
|
| 313 |
{
|
| 314 |
"epoch": 3.0,
|
| 315 |
"learning_rate": 8.34e-06,
|
| 316 |
+
"loss": 0.0953,
|
| 317 |
"step": 1250
|
| 318 |
},
|
| 319 |
{
|
| 320 |
"epoch": 3.01,
|
| 321 |
"learning_rate": 8.284444444444446e-06,
|
| 322 |
+
"loss": 0.0923,
|
| 323 |
"step": 1275
|
| 324 |
},
|
| 325 |
{
|
| 326 |
"epoch": 3.01,
|
| 327 |
"learning_rate": 8.22888888888889e-06,
|
| 328 |
+
"loss": 0.0749,
|
| 329 |
"step": 1300
|
| 330 |
},
|
| 331 |
{
|
| 332 |
"epoch": 3.02,
|
| 333 |
"learning_rate": 8.173333333333334e-06,
|
| 334 |
+
"loss": 0.0796,
|
| 335 |
"step": 1325
|
| 336 |
},
|
| 337 |
{
|
| 338 |
"epoch": 3.02,
|
| 339 |
"learning_rate": 8.11777777777778e-06,
|
| 340 |
+
"loss": 0.0786,
|
| 341 |
"step": 1350
|
| 342 |
},
|
| 343 |
{
|
| 344 |
"epoch": 3.03,
|
| 345 |
"learning_rate": 8.062222222222222e-06,
|
| 346 |
+
"loss": 0.0819,
|
| 347 |
"step": 1375
|
| 348 |
},
|
| 349 |
{
|
| 350 |
"epoch": 3.03,
|
| 351 |
"learning_rate": 8.006666666666667e-06,
|
| 352 |
+
"loss": 0.0747,
|
| 353 |
"step": 1400
|
| 354 |
},
|
| 355 |
{
|
| 356 |
"epoch": 3.04,
|
| 357 |
"learning_rate": 7.951111111111111e-06,
|
| 358 |
+
"loss": 0.0646,
|
| 359 |
"step": 1425
|
| 360 |
},
|
| 361 |
{
|
| 362 |
"epoch": 3.04,
|
| 363 |
"learning_rate": 7.895555555555557e-06,
|
| 364 |
+
"loss": 0.0707,
|
| 365 |
"step": 1450
|
| 366 |
},
|
| 367 |
{
|
| 368 |
"epoch": 3.05,
|
| 369 |
"learning_rate": 7.840000000000001e-06,
|
| 370 |
+
"loss": 0.0639,
|
| 371 |
"step": 1475
|
| 372 |
},
|
| 373 |
{
|
| 374 |
"epoch": 3.05,
|
| 375 |
"learning_rate": 7.784444444444445e-06,
|
| 376 |
+
"loss": 0.0544,
|
| 377 |
"step": 1500
|
| 378 |
},
|
| 379 |
{
|
| 380 |
"epoch": 3.06,
|
| 381 |
"learning_rate": 7.72888888888889e-06,
|
| 382 |
+
"loss": 0.0531,
|
| 383 |
"step": 1525
|
| 384 |
},
|
| 385 |
{
|
| 386 |
"epoch": 3.06,
|
| 387 |
"learning_rate": 7.673333333333333e-06,
|
| 388 |
+
"loss": 0.0518,
|
| 389 |
"step": 1550
|
| 390 |
},
|
| 391 |
{
|
| 392 |
"epoch": 3.07,
|
| 393 |
"learning_rate": 7.617777777777778e-06,
|
| 394 |
+
"loss": 0.0461,
|
| 395 |
"step": 1575
|
| 396 |
},
|
| 397 |
{
|
| 398 |
"epoch": 3.07,
|
| 399 |
"learning_rate": 7.562222222222223e-06,
|
| 400 |
+
"loss": 0.0424,
|
| 401 |
"step": 1600
|
| 402 |
},
|
| 403 |
{
|
| 404 |
"epoch": 3.08,
|
| 405 |
"learning_rate": 7.506666666666668e-06,
|
| 406 |
+
"loss": 0.0392,
|
| 407 |
"step": 1625
|
| 408 |
},
|
| 409 |
{
|
| 410 |
"epoch": 4.0,
|
| 411 |
"learning_rate": 7.451111111111111e-06,
|
| 412 |
+
"loss": 0.0376,
|
| 413 |
"step": 1650
|
| 414 |
},
|
| 415 |
{
|
| 416 |
"epoch": 4.01,
|
| 417 |
"learning_rate": 7.395555555555556e-06,
|
| 418 |
+
"loss": 0.0447,
|
| 419 |
"step": 1675
|
| 420 |
},
|
| 421 |
{
|
| 422 |
"epoch": 4.01,
|
| 423 |
"learning_rate": 7.340000000000001e-06,
|
| 424 |
+
"loss": 0.0362,
|
| 425 |
"step": 1700
|
| 426 |
},
|
| 427 |
{
|
| 428 |
"epoch": 4.02,
|
| 429 |
"learning_rate": 7.284444444444445e-06,
|
| 430 |
+
"loss": 0.0368,
|
| 431 |
"step": 1725
|
| 432 |
},
|
| 433 |
{
|
| 434 |
"epoch": 4.02,
|
| 435 |
"learning_rate": 7.22888888888889e-06,
|
| 436 |
+
"loss": 0.0367,
|
| 437 |
"step": 1750
|
| 438 |
},
|
| 439 |
{
|
| 440 |
"epoch": 4.03,
|
| 441 |
"learning_rate": 7.173333333333335e-06,
|
| 442 |
+
"loss": 0.0391,
|
| 443 |
"step": 1775
|
| 444 |
},
|
| 445 |
{
|
| 446 |
"epoch": 4.03,
|
| 447 |
"learning_rate": 7.117777777777778e-06,
|
| 448 |
+
"loss": 0.0331,
|
| 449 |
"step": 1800
|
| 450 |
},
|
| 451 |
{
|
| 452 |
"epoch": 4.04,
|
| 453 |
"learning_rate": 7.062222222222223e-06,
|
| 454 |
+
"loss": 0.028,
|
| 455 |
"step": 1825
|
| 456 |
},
|
| 457 |
{
|
| 458 |
"epoch": 4.04,
|
| 459 |
"learning_rate": 7.006666666666667e-06,
|
| 460 |
+
"loss": 0.0315,
|
| 461 |
"step": 1850
|
| 462 |
},
|
| 463 |
{
|
| 464 |
"epoch": 4.05,
|
| 465 |
"learning_rate": 6.951111111111112e-06,
|
| 466 |
+
"loss": 0.0308,
|
| 467 |
"step": 1875
|
| 468 |
},
|
| 469 |
{
|
| 470 |
"epoch": 4.05,
|
| 471 |
"learning_rate": 6.8955555555555565e-06,
|
| 472 |
+
"loss": 0.0248,
|
| 473 |
"step": 1900
|
| 474 |
},
|
| 475 |
{
|
| 476 |
"epoch": 4.06,
|
| 477 |
"learning_rate": 6.8400000000000014e-06,
|
| 478 |
+
"loss": 0.023,
|
| 479 |
"step": 1925
|
| 480 |
},
|
| 481 |
{
|
| 482 |
"epoch": 4.06,
|
| 483 |
"learning_rate": 6.784444444444445e-06,
|
| 484 |
+
"loss": 0.0233,
|
| 485 |
"step": 1950
|
| 486 |
},
|
| 487 |
{
|
| 488 |
"epoch": 4.07,
|
| 489 |
"learning_rate": 6.7288888888888895e-06,
|
| 490 |
+
"loss": 0.0236,
|
| 491 |
"step": 1975
|
| 492 |
},
|
| 493 |
{
|
| 494 |
"epoch": 4.07,
|
| 495 |
"learning_rate": 6.6733333333333335e-06,
|
| 496 |
+
"loss": 0.0181,
|
| 497 |
"step": 2000
|
| 498 |
},
|
| 499 |
{
|
| 500 |
"epoch": 4.07,
|
| 501 |
+
"eval_loss": 0.33755382895469666,
|
| 502 |
+
"eval_runtime": 952.5181,
|
| 503 |
+
"eval_samples_per_second": 6.92,
|
| 504 |
+
"eval_steps_per_second": 0.433,
|
| 505 |
+
"eval_wer": 20.286362347604197,
|
| 506 |
"step": 2000
|
| 507 |
},
|
| 508 |
{
|
| 509 |
"epoch": 4.08,
|
| 510 |
"learning_rate": 6.617777777777778e-06,
|
| 511 |
+
"loss": 0.0176,
|
| 512 |
"step": 2025
|
| 513 |
},
|
| 514 |
{
|
| 515 |
"epoch": 4.08,
|
| 516 |
"learning_rate": 6.562222222222223e-06,
|
| 517 |
+
"loss": 0.0177,
|
| 518 |
"step": 2050
|
| 519 |
},
|
| 520 |
{
|
| 521 |
"epoch": 5.0,
|
| 522 |
"learning_rate": 6.5066666666666665e-06,
|
| 523 |
+
"loss": 0.0198,
|
| 524 |
"step": 2075
|
| 525 |
},
|
| 526 |
{
|
| 527 |
"epoch": 5.01,
|
| 528 |
"learning_rate": 6.451111111111111e-06,
|
| 529 |
+
"loss": 0.0188,
|
| 530 |
"step": 2100
|
| 531 |
},
|
| 532 |
{
|
| 533 |
"epoch": 5.01,
|
| 534 |
"learning_rate": 6.395555555555556e-06,
|
| 535 |
+
"loss": 0.0153,
|
| 536 |
"step": 2125
|
| 537 |
},
|
| 538 |
{
|
| 539 |
"epoch": 5.02,
|
| 540 |
"learning_rate": 6.34e-06,
|
| 541 |
+
"loss": 0.0157,
|
| 542 |
"step": 2150
|
| 543 |
},
|
| 544 |
{
|
| 545 |
"epoch": 5.02,
|
| 546 |
"learning_rate": 6.284444444444445e-06,
|
| 547 |
+
"loss": 0.0196,
|
| 548 |
"step": 2175
|
| 549 |
},
|
| 550 |
{
|
| 551 |
"epoch": 5.03,
|
| 552 |
"learning_rate": 6.22888888888889e-06,
|
| 553 |
+
"loss": 0.0153,
|
| 554 |
"step": 2200
|
| 555 |
},
|
| 556 |
{
|
| 557 |
"epoch": 5.03,
|
| 558 |
"learning_rate": 6.173333333333333e-06,
|
| 559 |
+
"loss": 0.0132,
|
| 560 |
"step": 2225
|
| 561 |
},
|
| 562 |
{
|
| 563 |
"epoch": 5.04,
|
| 564 |
"learning_rate": 6.117777777777778e-06,
|
| 565 |
+
"loss": 0.0144,
|
| 566 |
"step": 2250
|
| 567 |
},
|
| 568 |
{
|
| 569 |
"epoch": 5.04,
|
| 570 |
"learning_rate": 6.062222222222223e-06,
|
| 571 |
+
"loss": 0.0145,
|
| 572 |
"step": 2275
|
| 573 |
},
|
| 574 |
{
|
| 575 |
"epoch": 5.05,
|
| 576 |
"learning_rate": 6.006666666666667e-06,
|
| 577 |
+
"loss": 0.0131,
|
| 578 |
"step": 2300
|
| 579 |
},
|
| 580 |
{
|
| 581 |
"epoch": 5.05,
|
| 582 |
"learning_rate": 5.951111111111112e-06,
|
| 583 |
+
"loss": 0.0107,
|
| 584 |
"step": 2325
|
| 585 |
},
|
| 586 |
{
|
| 587 |
"epoch": 5.06,
|
| 588 |
"learning_rate": 5.895555555555557e-06,
|
| 589 |
+
"loss": 0.0122,
|
| 590 |
"step": 2350
|
| 591 |
},
|
| 592 |
{
|
| 593 |
"epoch": 5.06,
|
| 594 |
"learning_rate": 5.84e-06,
|
| 595 |
+
"loss": 0.0121,
|
| 596 |
"step": 2375
|
| 597 |
},
|
| 598 |
{
|
| 599 |
"epoch": 5.07,
|
| 600 |
"learning_rate": 5.784444444444445e-06,
|
| 601 |
+
"loss": 0.0103,
|
| 602 |
"step": 2400
|
| 603 |
},
|
| 604 |
{
|
| 605 |
"epoch": 5.07,
|
| 606 |
"learning_rate": 5.72888888888889e-06,
|
| 607 |
+
"loss": 0.0089,
|
| 608 |
"step": 2425
|
| 609 |
},
|
| 610 |
{
|
| 611 |
"epoch": 5.08,
|
| 612 |
"learning_rate": 5.673333333333334e-06,
|
| 613 |
+
"loss": 0.0087,
|
| 614 |
"step": 2450
|
| 615 |
},
|
| 616 |
{
|
| 617 |
"epoch": 6.0,
|
| 618 |
"learning_rate": 5.617777777777779e-06,
|
| 619 |
+
"loss": 0.0087,
|
| 620 |
"step": 2475
|
| 621 |
},
|
| 622 |
{
|
| 623 |
"epoch": 6.01,
|
| 624 |
"learning_rate": 5.562222222222222e-06,
|
| 625 |
+
"loss": 0.0106,
|
| 626 |
"step": 2500
|
| 627 |
},
|
| 628 |
{
|
| 629 |
"epoch": 6.01,
|
| 630 |
"learning_rate": 5.506666666666667e-06,
|
| 631 |
+
"loss": 0.0094,
|
| 632 |
"step": 2525
|
| 633 |
},
|
| 634 |
{
|
| 635 |
"epoch": 6.02,
|
| 636 |
"learning_rate": 5.451111111111112e-06,
|
| 637 |
+
"loss": 0.0082,
|
| 638 |
"step": 2550
|
| 639 |
},
|
| 640 |
{
|
| 641 |
"epoch": 6.02,
|
| 642 |
"learning_rate": 5.3955555555555565e-06,
|
| 643 |
+
"loss": 0.0079,
|
| 644 |
"step": 2575
|
| 645 |
},
|
| 646 |
{
|
| 647 |
"epoch": 6.03,
|
| 648 |
"learning_rate": 5.3400000000000005e-06,
|
| 649 |
+
"loss": 0.0084,
|
| 650 |
"step": 2600
|
| 651 |
},
|
| 652 |
{
|
| 653 |
"epoch": 6.03,
|
| 654 |
"learning_rate": 5.2844444444444454e-06,
|
| 655 |
+
"loss": 0.0076,
|
| 656 |
"step": 2625
|
| 657 |
},
|
| 658 |
{
|
| 659 |
"epoch": 6.04,
|
| 660 |
"learning_rate": 5.228888888888889e-06,
|
| 661 |
+
"loss": 0.0069,
|
| 662 |
"step": 2650
|
| 663 |
},
|
| 664 |
{
|
| 665 |
"epoch": 6.04,
|
| 666 |
"learning_rate": 5.1733333333333335e-06,
|
| 667 |
+
"loss": 0.0067,
|
| 668 |
"step": 2675
|
| 669 |
},
|
| 670 |
{
|
| 671 |
"epoch": 6.05,
|
| 672 |
"learning_rate": 5.117777777777778e-06,
|
| 673 |
+
"loss": 0.0073,
|
| 674 |
"step": 2700
|
| 675 |
},
|
| 676 |
{
|
| 677 |
"epoch": 6.05,
|
| 678 |
"learning_rate": 5.062222222222222e-06,
|
| 679 |
+
"loss": 0.0062,
|
| 680 |
"step": 2725
|
| 681 |
},
|
| 682 |
{
|
| 683 |
"epoch": 6.06,
|
| 684 |
"learning_rate": 5.006666666666667e-06,
|
| 685 |
+
"loss": 0.0057,
|
| 686 |
"step": 2750
|
| 687 |
},
|
| 688 |
{
|
| 689 |
"epoch": 6.06,
|
| 690 |
"learning_rate": 4.951111111111111e-06,
|
| 691 |
+
"loss": 0.0064,
|
| 692 |
"step": 2775
|
| 693 |
},
|
| 694 |
{
|
|
|
|
| 700 |
{
|
| 701 |
"epoch": 6.07,
|
| 702 |
"learning_rate": 4.84e-06,
|
| 703 |
+
"loss": 0.0047,
|
| 704 |
"step": 2825
|
| 705 |
},
|
| 706 |
{
|
| 707 |
"epoch": 6.08,
|
| 708 |
"learning_rate": 4.784444444444445e-06,
|
| 709 |
+
"loss": 0.005,
|
| 710 |
"step": 2850
|
| 711 |
},
|
| 712 |
{
|
| 713 |
"epoch": 6.08,
|
| 714 |
"learning_rate": 4.728888888888889e-06,
|
| 715 |
+
"loss": 0.0047,
|
| 716 |
"step": 2875
|
| 717 |
},
|
| 718 |
{
|
| 719 |
"epoch": 7.0,
|
| 720 |
"learning_rate": 4.673333333333333e-06,
|
| 721 |
+
"loss": 0.0052,
|
| 722 |
"step": 2900
|
| 723 |
},
|
| 724 |
{
|
| 725 |
"epoch": 7.01,
|
| 726 |
"learning_rate": 4.617777777777778e-06,
|
| 727 |
+
"loss": 0.005,
|
| 728 |
"step": 2925
|
| 729 |
},
|
| 730 |
{
|
| 731 |
"epoch": 7.01,
|
| 732 |
"learning_rate": 4.562222222222222e-06,
|
| 733 |
+
"loss": 0.0041,
|
| 734 |
"step": 2950
|
| 735 |
},
|
| 736 |
{
|
| 737 |
"epoch": 7.02,
|
| 738 |
"learning_rate": 4.506666666666667e-06,
|
| 739 |
+
"loss": 0.0045,
|
| 740 |
"step": 2975
|
| 741 |
},
|
| 742 |
{
|
| 743 |
"epoch": 7.02,
|
| 744 |
"learning_rate": 4.451111111111112e-06,
|
| 745 |
+
"loss": 0.0044,
|
| 746 |
"step": 3000
|
| 747 |
},
|
| 748 |
{
|
| 749 |
"epoch": 7.02,
|
| 750 |
+
"eval_loss": 0.3602650463581085,
|
| 751 |
+
"eval_runtime": 960.4855,
|
| 752 |
+
"eval_samples_per_second": 6.862,
|
| 753 |
+
"eval_steps_per_second": 0.429,
|
| 754 |
+
"eval_wer": 18.876827737048888,
|
| 755 |
"step": 3000
|
| 756 |
},
|
| 757 |
{
|
| 758 |
"epoch": 7.03,
|
| 759 |
"learning_rate": 4.395555555555556e-06,
|
| 760 |
+
"loss": 0.0043,
|
| 761 |
"step": 3025
|
| 762 |
},
|
| 763 |
{
|
| 764 |
"epoch": 7.03,
|
| 765 |
"learning_rate": 4.34e-06,
|
| 766 |
+
"loss": 0.0041,
|
| 767 |
"step": 3050
|
| 768 |
},
|
| 769 |
{
|
| 770 |
"epoch": 7.04,
|
| 771 |
"learning_rate": 4.284444444444445e-06,
|
| 772 |
+
"loss": 0.0035,
|
| 773 |
"step": 3075
|
| 774 |
},
|
| 775 |
{
|
| 776 |
"epoch": 7.04,
|
| 777 |
"learning_rate": 4.228888888888889e-06,
|
| 778 |
+
"loss": 0.0044,
|
| 779 |
"step": 3100
|
| 780 |
},
|
| 781 |
{
|
| 782 |
"epoch": 7.05,
|
| 783 |
"learning_rate": 4.173333333333334e-06,
|
| 784 |
+
"loss": 0.0039,
|
| 785 |
"step": 3125
|
| 786 |
},
|
| 787 |
{
|
| 788 |
"epoch": 7.05,
|
| 789 |
"learning_rate": 4.117777777777779e-06,
|
| 790 |
+
"loss": 0.0031,
|
| 791 |
"step": 3150
|
| 792 |
},
|
| 793 |
{
|
|
|
|
| 805 |
{
|
| 806 |
"epoch": 7.07,
|
| 807 |
"learning_rate": 3.951111111111112e-06,
|
| 808 |
+
"loss": 0.0028,
|
| 809 |
"step": 3225
|
| 810 |
},
|
| 811 |
{
|
| 812 |
"epoch": 7.07,
|
| 813 |
"learning_rate": 3.895555555555556e-06,
|
| 814 |
+
"loss": 0.0032,
|
| 815 |
"step": 3250
|
| 816 |
},
|
| 817 |
{
|
| 818 |
"epoch": 7.08,
|
| 819 |
"learning_rate": 3.8400000000000005e-06,
|
| 820 |
+
"loss": 0.003,
|
| 821 |
"step": 3275
|
| 822 |
},
|
| 823 |
{
|
| 824 |
"epoch": 8.0,
|
| 825 |
"learning_rate": 3.784444444444445e-06,
|
| 826 |
+
"loss": 0.0034,
|
| 827 |
"step": 3300
|
| 828 |
},
|
| 829 |
{
|
| 830 |
"epoch": 8.01,
|
| 831 |
"learning_rate": 3.728888888888889e-06,
|
| 832 |
+
"loss": 0.003,
|
| 833 |
"step": 3325
|
| 834 |
},
|
| 835 |
{
|
| 836 |
"epoch": 8.01,
|
| 837 |
"learning_rate": 3.673333333333334e-06,
|
| 838 |
+
"loss": 0.0027,
|
| 839 |
"step": 3350
|
| 840 |
},
|
| 841 |
{
|
|
|
|
| 847 |
{
|
| 848 |
"epoch": 8.02,
|
| 849 |
"learning_rate": 3.5622222222222224e-06,
|
| 850 |
+
"loss": 0.0028,
|
| 851 |
"step": 3400
|
| 852 |
},
|
| 853 |
{
|
| 854 |
"epoch": 8.03,
|
| 855 |
"learning_rate": 3.5066666666666673e-06,
|
| 856 |
+
"loss": 0.0029,
|
| 857 |
"step": 3425
|
| 858 |
},
|
| 859 |
{
|
|
|
|
| 865 |
{
|
| 866 |
"epoch": 8.04,
|
| 867 |
"learning_rate": 3.3955555555555558e-06,
|
| 868 |
+
"loss": 0.0022,
|
| 869 |
"step": 3475
|
| 870 |
},
|
| 871 |
{
|
|
|
|
| 877 |
{
|
| 878 |
"epoch": 8.05,
|
| 879 |
"learning_rate": 3.2844444444444447e-06,
|
| 880 |
+
"loss": 0.0028,
|
| 881 |
"step": 3525
|
| 882 |
},
|
| 883 |
{
|
| 884 |
"epoch": 8.05,
|
| 885 |
"learning_rate": 3.228888888888889e-06,
|
| 886 |
+
"loss": 0.0025,
|
| 887 |
"step": 3550
|
| 888 |
},
|
| 889 |
{
|
| 890 |
"epoch": 8.06,
|
| 891 |
"learning_rate": 3.173333333333334e-06,
|
| 892 |
+
"loss": 0.0022,
|
| 893 |
"step": 3575
|
| 894 |
},
|
| 895 |
{
|
| 896 |
"epoch": 8.06,
|
| 897 |
"learning_rate": 3.117777777777778e-06,
|
| 898 |
+
"loss": 0.0021,
|
| 899 |
"step": 3600
|
| 900 |
},
|
| 901 |
{
|
| 902 |
"epoch": 8.07,
|
| 903 |
"learning_rate": 3.0622222222222225e-06,
|
| 904 |
+
"loss": 0.002,
|
| 905 |
"step": 3625
|
| 906 |
},
|
| 907 |
{
|
| 908 |
"epoch": 8.07,
|
| 909 |
"learning_rate": 3.0066666666666674e-06,
|
| 910 |
+
"loss": 0.0019,
|
| 911 |
"step": 3650
|
| 912 |
},
|
| 913 |
{
|
| 914 |
"epoch": 8.08,
|
| 915 |
"learning_rate": 2.9511111111111114e-06,
|
| 916 |
+
"loss": 0.0019,
|
| 917 |
"step": 3675
|
| 918 |
},
|
| 919 |
{
|
| 920 |
+
"epoch": 8.08,
|
| 921 |
"learning_rate": 2.895555555555556e-06,
|
| 922 |
"loss": 0.0019,
|
| 923 |
"step": 3700
|
| 924 |
},
|
| 925 |
{
|
| 926 |
+
"epoch": 9.0,
|
| 927 |
"learning_rate": 2.84e-06,
|
| 928 |
+
"loss": 0.0021,
|
| 929 |
"step": 3725
|
| 930 |
},
|
| 931 |
{
|
| 932 |
"epoch": 9.01,
|
| 933 |
"learning_rate": 2.784444444444445e-06,
|
| 934 |
+
"loss": 0.0021,
|
| 935 |
"step": 3750
|
| 936 |
},
|
| 937 |
{
|
| 938 |
+
"epoch": 9.01,
|
| 939 |
"learning_rate": 2.7288888888888893e-06,
|
| 940 |
+
"loss": 0.0018,
|
| 941 |
"step": 3775
|
| 942 |
},
|
| 943 |
{
|
| 944 |
"epoch": 9.02,
|
| 945 |
"learning_rate": 2.6733333333333333e-06,
|
| 946 |
+
"loss": 0.0019,
|
| 947 |
"step": 3800
|
| 948 |
},
|
| 949 |
{
|
| 950 |
+
"epoch": 9.02,
|
| 951 |
"learning_rate": 2.617777777777778e-06,
|
| 952 |
+
"loss": 0.0018,
|
| 953 |
"step": 3825
|
| 954 |
},
|
| 955 |
{
|
| 956 |
"epoch": 9.03,
|
| 957 |
"learning_rate": 2.5622222222222226e-06,
|
| 958 |
+
"loss": 0.0021,
|
| 959 |
"step": 3850
|
| 960 |
},
|
| 961 |
{
|
| 962 |
+
"epoch": 9.03,
|
| 963 |
"learning_rate": 2.5066666666666667e-06,
|
| 964 |
+
"loss": 0.0017,
|
| 965 |
"step": 3875
|
| 966 |
},
|
| 967 |
{
|
| 968 |
"epoch": 9.04,
|
| 969 |
"learning_rate": 2.451111111111111e-06,
|
| 970 |
+
"loss": 0.0018,
|
| 971 |
"step": 3900
|
| 972 |
},
|
| 973 |
{
|
| 974 |
+
"epoch": 9.04,
|
| 975 |
"learning_rate": 2.3955555555555556e-06,
|
| 976 |
"loss": 0.0017,
|
| 977 |
"step": 3925
|
|
|
|
| 979 |
{
|
| 980 |
"epoch": 9.05,
|
| 981 |
"learning_rate": 2.3400000000000005e-06,
|
| 982 |
+
"loss": 0.002,
|
| 983 |
"step": 3950
|
| 984 |
},
|
| 985 |
{
|
| 986 |
+
"epoch": 9.05,
|
| 987 |
"learning_rate": 2.2844444444444445e-06,
|
| 988 |
"loss": 0.0017,
|
| 989 |
"step": 3975
|
|
|
|
| 991 |
{
|
| 992 |
"epoch": 9.06,
|
| 993 |
"learning_rate": 2.228888888888889e-06,
|
| 994 |
+
"loss": 0.0016,
|
| 995 |
"step": 4000
|
| 996 |
},
|
| 997 |
{
|
| 998 |
"epoch": 9.06,
|
| 999 |
+
"eval_loss": 0.3812163174152374,
|
| 1000 |
+
"eval_runtime": 968.0397,
|
| 1001 |
+
"eval_samples_per_second": 6.809,
|
| 1002 |
+
"eval_steps_per_second": 0.426,
|
| 1003 |
+
"eval_wer": 18.775568066750374,
|
| 1004 |
"step": 4000
|
| 1005 |
},
|
| 1006 |
{
|
| 1007 |
+
"epoch": 9.06,
|
| 1008 |
"learning_rate": 2.1733333333333334e-06,
|
| 1009 |
"loss": 0.0016,
|
| 1010 |
"step": 4025
|
|
|
|
| 1016 |
"step": 4050
|
| 1017 |
},
|
| 1018 |
{
|
| 1019 |
+
"epoch": 9.07,
|
| 1020 |
"learning_rate": 2.0622222222222223e-06,
|
| 1021 |
"loss": 0.0015,
|
| 1022 |
"step": 4075
|
|
|
|
| 1030 |
{
|
| 1031 |
"epoch": 10.0,
|
| 1032 |
"learning_rate": 1.9511111111111113e-06,
|
| 1033 |
+
"loss": 0.0015,
|
| 1034 |
"step": 4125
|
| 1035 |
},
|
| 1036 |
{
|
|
|
|
| 1054 |
{
|
| 1055 |
"epoch": 10.02,
|
| 1056 |
"learning_rate": 1.728888888888889e-06,
|
| 1057 |
+
"loss": 0.0014,
|
| 1058 |
"step": 4225
|
| 1059 |
},
|
| 1060 |
{
|
|
|
|
| 1066 |
{
|
| 1067 |
"epoch": 10.03,
|
| 1068 |
"learning_rate": 1.6177777777777778e-06,
|
| 1069 |
+
"loss": 0.0015,
|
| 1070 |
"step": 4275
|
| 1071 |
},
|
| 1072 |
{
|
|
|
|
| 1078 |
{
|
| 1079 |
"epoch": 10.04,
|
| 1080 |
"learning_rate": 1.506666666666667e-06,
|
| 1081 |
+
"loss": 0.0016,
|
| 1082 |
"step": 4325
|
| 1083 |
},
|
| 1084 |
{
|
|
|
|
| 1124 |
"step": 4500
|
| 1125 |
},
|
| 1126 |
{
|
| 1127 |
+
"epoch": 10.08,
|
| 1128 |
"learning_rate": 1.0622222222222222e-06,
|
| 1129 |
+
"loss": 0.0013,
|
| 1130 |
"step": 4525
|
| 1131 |
},
|
| 1132 |
{
|
| 1133 |
+
"epoch": 11.0,
|
| 1134 |
"learning_rate": 1.0066666666666668e-06,
|
| 1135 |
"loss": 0.0014,
|
| 1136 |
"step": 4550
|
|
|
|
| 1138 |
{
|
| 1139 |
"epoch": 11.01,
|
| 1140 |
"learning_rate": 9.511111111111111e-07,
|
| 1141 |
+
"loss": 0.0014,
|
| 1142 |
"step": 4575
|
| 1143 |
},
|
| 1144 |
{
|
| 1145 |
+
"epoch": 11.01,
|
| 1146 |
"learning_rate": 8.955555555555557e-07,
|
| 1147 |
"loss": 0.0013,
|
| 1148 |
"step": 4600
|
|
|
|
| 1154 |
"step": 4625
|
| 1155 |
},
|
| 1156 |
{
|
| 1157 |
+
"epoch": 11.02,
|
| 1158 |
"learning_rate": 7.844444444444445e-07,
|
| 1159 |
+
"loss": 0.0014,
|
| 1160 |
"step": 4650
|
| 1161 |
},
|
| 1162 |
{
|
| 1163 |
"epoch": 11.03,
|
| 1164 |
"learning_rate": 7.28888888888889e-07,
|
| 1165 |
+
"loss": 0.0015,
|
| 1166 |
"step": 4675
|
| 1167 |
},
|
| 1168 |
{
|
| 1169 |
+
"epoch": 11.03,
|
| 1170 |
"learning_rate": 6.733333333333334e-07,
|
| 1171 |
"loss": 0.0013,
|
| 1172 |
"step": 4700
|
|
|
|
| 1178 |
"step": 4725
|
| 1179 |
},
|
| 1180 |
{
|
| 1181 |
+
"epoch": 11.04,
|
| 1182 |
"learning_rate": 5.622222222222223e-07,
|
| 1183 |
"loss": 0.0013,
|
| 1184 |
"step": 4750
|
|
|
|
| 1190 |
"step": 4775
|
| 1191 |
},
|
| 1192 |
{
|
| 1193 |
+
"epoch": 11.05,
|
| 1194 |
"learning_rate": 4.511111111111111e-07,
|
| 1195 |
"loss": 0.0013,
|
| 1196 |
"step": 4800
|
|
|
|
| 1202 |
"step": 4825
|
| 1203 |
},
|
| 1204 |
{
|
| 1205 |
+
"epoch": 11.06,
|
| 1206 |
"learning_rate": 3.4000000000000003e-07,
|
| 1207 |
+
"loss": 0.0013,
|
| 1208 |
"step": 4850
|
| 1209 |
},
|
| 1210 |
{
|
| 1211 |
"epoch": 11.07,
|
| 1212 |
"learning_rate": 2.844444444444445e-07,
|
| 1213 |
+
"loss": 0.0013,
|
| 1214 |
"step": 4875
|
| 1215 |
},
|
| 1216 |
{
|
| 1217 |
+
"epoch": 11.07,
|
| 1218 |
"learning_rate": 2.2888888888888892e-07,
|
| 1219 |
"loss": 0.0012,
|
| 1220 |
"step": 4900
|
|
|
|
| 1222 |
{
|
| 1223 |
"epoch": 11.08,
|
| 1224 |
"learning_rate": 1.7333333333333335e-07,
|
| 1225 |
+
"loss": 0.0012,
|
| 1226 |
"step": 4925
|
| 1227 |
},
|
| 1228 |
{
|
|
|
|
| 1240 |
{
|
| 1241 |
"epoch": 12.01,
|
| 1242 |
"learning_rate": 6.666666666666667e-09,
|
| 1243 |
+
"loss": 0.0012,
|
| 1244 |
"step": 5000
|
| 1245 |
},
|
| 1246 |
{
|
| 1247 |
"epoch": 12.01,
|
| 1248 |
+
"eval_loss": 0.3913772702217102,
|
| 1249 |
+
"eval_runtime": 973.2149,
|
| 1250 |
+
"eval_samples_per_second": 6.772,
|
| 1251 |
+
"eval_steps_per_second": 0.423,
|
| 1252 |
+
"eval_wer": 18.83024828871157,
|
| 1253 |
"step": 5000
|
| 1254 |
},
|
| 1255 |
{
|
| 1256 |
"epoch": 12.01,
|
| 1257 |
"step": 5000,
|
| 1258 |
+
"total_flos": 4.616673800159232e+19,
|
| 1259 |
+
"train_loss": 0.106446673027426,
|
| 1260 |
+
"train_runtime": 27653.1068,
|
| 1261 |
+
"train_samples_per_second": 5.786,
|
| 1262 |
+
"train_steps_per_second": 0.181
|
| 1263 |
}
|
| 1264 |
],
|
| 1265 |
"max_steps": 5000,
|
| 1266 |
"num_train_epochs": 9223372036854775807,
|
| 1267 |
+
"total_flos": 4.616673800159232e+19,
|
| 1268 |
"trial_name": null,
|
| 1269 |
"trial_params": null
|
| 1270 |
}
|