<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->
RATS_clean
This model was trained from scratch on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.3775
- Wer: 0.1324
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0001
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 1000
- num_epochs: 20
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss | Wer |
---|---|---|---|---|
0.4025 | 0.42 | 1000 | 0.3019 | 0.1434 |
0.3939 | 0.85 | 2000 | 0.3044 | 0.1377 |
0.3808 | 1.27 | 3000 | 0.3053 | 0.1417 |
0.377 | 1.7 | 4000 | 0.3213 | 0.1414 |
0.3616 | 2.12 | 5000 | 0.2976 | 0.1449 |
0.3519 | 2.55 | 6000 | 0.3266 | 0.1482 |
0.3532 | 2.97 | 7000 | 0.3365 | 0.1455 |
0.3377 | 3.4 | 8000 | 0.3170 | 0.1409 |
0.3321 | 3.82 | 9000 | 0.3070 | 0.1398 |
0.3199 | 4.25 | 10000 | 0.3123 | 0.1366 |
0.322 | 4.67 | 11000 | 0.3166 | 0.1378 |
0.3085 | 5.1 | 12000 | 0.3492 | 0.1448 |
0.2954 | 5.52 | 13000 | 0.3173 | 0.1387 |
0.3003 | 5.95 | 14000 | 0.3341 | 0.1442 |
0.2863 | 6.37 | 15000 | 0.3124 | 0.1382 |
0.2887 | 6.8 | 16000 | 0.3273 | 0.1404 |
0.2777 | 7.22 | 17000 | 0.3291 | 0.1399 |
0.2728 | 7.65 | 18000 | 0.3326 | 0.1352 |
0.2726 | 8.07 | 19000 | 0.3443 | 0.1355 |
0.2519 | 8.5 | 20000 | 0.3448 | 0.1400 |
0.3256 | 8.92 | 21000 | 0.3274 | 0.1396 |
0.3174 | 9.35 | 22000 | 0.3210 | 0.1411 |
0.3075 | 9.77 | 23000 | 0.3319 | 0.1410 |
0.2982 | 10.2 | 24000 | 0.3463 | 0.1424 |
0.2927 | 10.62 | 25000 | 0.3445 | 0.1399 |
0.3027 | 11.05 | 26000 | 0.3488 | 0.1410 |
0.2835 | 11.47 | 27000 | 0.3639 | 0.1371 |
0.2767 | 11.89 | 28000 | 0.3467 | 0.1391 |
0.2792 | 12.32 | 29000 | 0.3459 | 0.1358 |
0.2701 | 12.74 | 30000 | 0.3425 | 0.1351 |
0.269 | 13.17 | 31000 | 0.3790 | 0.1421 |
0.2567 | 13.59 | 32000 | 0.3613 | 0.1352 |
0.2626 | 14.02 | 33000 | 0.3586 | 0.1396 |
0.2436 | 14.44 | 34000 | 0.3694 | 0.1369 |
0.2557 | 14.87 | 35000 | 0.3715 | 0.1321 |
0.2468 | 15.29 | 36000 | 0.3970 | 0.1348 |
0.2463 | 15.72 | 37000 | 0.3675 | 0.1304 |
0.2362 | 16.14 | 38000 | 0.3690 | 0.1377 |
0.2388 | 16.57 | 39000 | 0.3775 | 0.1310 |
0.2314 | 16.99 | 40000 | 0.3601 | 0.1326 |
0.2315 | 17.42 | 41000 | 0.3633 | 0.1322 |
0.2334 | 17.84 | 42000 | 0.3794 | 0.1356 |
0.2255 | 18.27 | 43000 | 0.3670 | 0.1316 |
0.2222 | 18.69 | 44000 | 0.3778 | 0.1341 |
0.225 | 19.12 | 45000 | 0.3708 | 0.1331 |
0.2209 | 19.54 | 46000 | 0.3807 | 0.1332 |
0.2216 | 19.97 | 47000 | 0.3775 | 0.1324 |
Framework versions
- Transformers 4.26.0
- Pytorch 1.12.1
- Datasets 2.9.0
- Tokenizers 0.13.2