metadata
license: apache-2.0
base_model: facebook/wav2vec2-xls-r-300m
tags:
- generated_from_trainer
metrics:
- wer
model-index:
- name: wav2vec2_xls_r_300m_FLEURS_Shona_1hr_v2
results: []
wav2vec2_xls_r_300m_FLEURS_Shona_1hr_v2
This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the None dataset. It achieves the following results on the evaluation set:
- Loss: 14.3078
- Wer: 1.0
- Cer: 0.9752
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
- gradient_accumulation_steps: 2
- total_train_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 500
- num_epochs: 100
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
|---|---|---|---|---|---|
| No log | 1.0 | 8 | 14.7146 | 1.0 | 0.9727 |
| No log | 2.0 | 16 | 14.6551 | 1.0 | 0.9785 |
| No log | 3.0 | 24 | 14.4901 | 1.0 | 0.9773 |
| No log | 4.0 | 32 | 14.2300 | 1.0 | 0.9908 |
| No log | 5.0 | 40 | 13.8814 | 1.0 | 0.9943 |
| No log | 6.0 | 48 | 13.1475 | 1.0 | 1.0 |
| No log | 7.0 | 56 | 11.5715 | 1.0 | 1.0 |
| No log | 8.0 | 64 | 9.0914 | 1.0 | 1.0 |
| No log | 9.0 | 72 | 7.1873 | 1.0 | 1.0 |
| No log | 10.0 | 80 | 5.8653 | 1.0 | 1.0 |
| No log | 11.0 | 88 | 5.1712 | 1.0 | 1.0 |
| No log | 12.0 | 96 | 4.7087 | 1.0 | 1.0 |
| 10.9024 | 13.0 | 104 | 4.3815 | 1.0 | 1.0 |
| 10.9024 | 14.0 | 112 | 4.1675 | 1.0 | 1.0 |
| 10.9024 | 15.0 | 120 | 3.9986 | 1.0 | 1.0 |
| 10.9024 | 16.0 | 128 | 3.8652 | 1.0 | 1.0 |
| 10.9024 | 17.0 | 136 | 3.7428 | 1.0 | 1.0 |
| 10.9024 | 18.0 | 144 | 3.6411 | 1.0 | 1.0 |
| 10.9024 | 19.0 | 152 | 3.5597 | 1.0 | 1.0 |
| 10.9024 | 20.0 | 160 | 3.4685 | 1.0 | 1.0 |
| 10.9024 | 21.0 | 168 | 3.3991 | 1.0 | 1.0 |
| 10.9024 | 22.0 | 176 | 3.3369 | 1.0 | 1.0 |
| 10.9024 | 23.0 | 184 | 3.2783 | 1.0 | 1.0 |
| 10.9024 | 24.0 | 192 | 3.2254 | 1.0 | 1.0 |
| 3.5833 | 25.0 | 200 | 3.1803 | 1.0 | 1.0 |
| 3.5833 | 26.0 | 208 | 3.1373 | 1.0 | 1.0 |
| 3.5833 | 27.0 | 216 | 3.1228 | 1.0 | 1.0 |
| 3.5833 | 28.0 | 224 | 3.0768 | 1.0 | 1.0 |
| 3.5833 | 29.0 | 232 | 3.0552 | 1.0 | 1.0 |
| 3.5833 | 30.0 | 240 | 3.0262 | 1.0 | 1.0 |
| 3.5833 | 31.0 | 248 | 3.0120 | 1.0 | 1.0 |
| 3.5833 | 32.0 | 256 | 2.9929 | 1.0 | 1.0 |
| 3.5833 | 33.0 | 264 | 2.9810 | 1.0 | 1.0 |
| 3.5833 | 34.0 | 272 | 2.9802 | 1.0 | 1.0 |
| 3.5833 | 35.0 | 280 | 2.9659 | 1.0 | 1.0 |
| 3.5833 | 36.0 | 288 | 2.9622 | 1.0 | 1.0 |
| 3.5833 | 37.0 | 296 | 2.9551 | 1.0 | 1.0 |
| 3.0223 | 38.0 | 304 | 2.9496 | 1.0 | 1.0 |
| 3.0223 | 39.0 | 312 | 2.9459 | 1.0 | 1.0 |
| 3.0223 | 40.0 | 320 | 2.9398 | 1.0 | 1.0 |
| 3.0223 | 41.0 | 328 | 2.9399 | 1.0 | 1.0 |
| 3.0223 | 42.0 | 336 | 2.9365 | 1.0 | 1.0 |
| 3.0223 | 43.0 | 344 | 2.9239 | 1.0 | 1.0 |
| 3.0223 | 44.0 | 352 | 2.9338 | 1.0 | 1.0 |
| 3.0223 | 45.0 | 360 | 2.8903 | 1.0 | 1.0 |
| 3.0223 | 46.0 | 368 | 2.8530 | 1.0 | 1.0 |
| 3.0223 | 47.0 | 376 | 2.7876 | 1.0 | 0.9951 |
| 3.0223 | 48.0 | 384 | 2.7081 | 1.0 | 0.9034 |
| 3.0223 | 49.0 | 392 | 2.6055 | 1.0 | 0.9060 |
| 2.8295 | 50.0 | 400 | 2.4135 | 1.0 | 0.8673 |
| 2.8295 | 51.0 | 408 | 2.1838 | 1.0 | 0.7115 |
| 2.8295 | 52.0 | 416 | 1.8849 | 1.0 | 0.5230 |
| 2.8295 | 53.0 | 424 | 1.5853 | 1.0 | 0.3966 |
| 2.8295 | 54.0 | 432 | 1.3813 | 1.0 | 0.3381 |
| 2.8295 | 55.0 | 440 | 1.2596 | 1.0 | 0.3092 |
| 2.8295 | 56.0 | 448 | 1.1627 | 1.0 | 0.2872 |
| 2.8295 | 57.0 | 456 | 1.0824 | 1.0 | 0.2847 |
| 2.8295 | 58.0 | 464 | 1.0633 | 1.0 | 0.2739 |
| 2.8295 | 59.0 | 472 | 0.9979 | 1.0 | 0.2628 |
| 2.8295 | 60.0 | 480 | 1.0218 | 1.0 | 0.2730 |
| 2.8295 | 61.0 | 488 | 1.0080 | 1.0 | 0.2629 |
| 2.8295 | 62.0 | 496 | 0.9897 | 1.0 | 0.2627 |
| 0.9322 | 63.0 | 504 | 0.9869 | 1.0 | 0.2639 |
| 0.9322 | 64.0 | 512 | 0.9348 | 1.0 | 0.2448 |
| 0.9322 | 65.0 | 520 | 0.9493 | 1.0 | 0.2459 |
| 0.9322 | 66.0 | 528 | 0.9553 | 1.0 | 0.2400 |
| 0.9322 | 67.0 | 536 | 0.9731 | 1.0 | 0.2468 |
| 0.9322 | 68.0 | 544 | 0.9631 | 1.0 | 0.2450 |
| 0.9322 | 69.0 | 552 | 0.9685 | 1.0 | 0.2521 |
| 0.9322 | 70.0 | 560 | 0.9508 | 1.0 | 0.2409 |
| 0.9322 | 71.0 | 568 | 1.0106 | 1.0 | 0.2484 |
| 0.9322 | 72.0 | 576 | 0.9573 | 1.0 | 0.2449 |
| 0.9322 | 73.0 | 584 | 0.9931 | 1.0 | 0.2482 |
| 0.9322 | 74.0 | 592 | 0.9571 | 1.0 | 0.2447 |
| 0.2354 | 75.0 | 600 | 0.9946 | 1.0 | 0.2439 |
| 0.2354 | 76.0 | 608 | 0.9376 | 1.0 | 0.2328 |
| 0.2354 | 77.0 | 616 | 0.9882 | 1.0 | 0.2415 |
| 0.2354 | 78.0 | 624 | 0.9730 | 1.0 | 0.2343 |
| 0.2354 | 79.0 | 632 | 0.9602 | 1.0 | 0.2355 |
| 0.2354 | 80.0 | 640 | 0.9791 | 1.0 | 0.2376 |
| 0.2354 | 81.0 | 648 | 0.9717 | 1.0 | 0.2327 |
| 0.2354 | 82.0 | 656 | 0.9760 | 1.0 | 0.2388 |
| 0.2354 | 83.0 | 664 | 0.9586 | 1.0 | 0.2359 |
| 0.2354 | 84.0 | 672 | 0.9799 | 1.0 | 0.2304 |
| 0.2354 | 85.0 | 680 | 0.9717 | 1.0 | 0.2299 |
| 0.2354 | 86.0 | 688 | 0.9727 | 1.0 | 0.2338 |
| 0.2354 | 87.0 | 696 | 0.9768 | 1.0 | 0.2331 |
| 0.1354 | 88.0 | 704 | 0.9938 | 1.0 | 0.2351 |
| 0.1354 | 89.0 | 712 | 0.9861 | 1.0 | 0.2304 |
| 0.1354 | 90.0 | 720 | 0.9772 | 1.0 | 0.2330 |
| 0.1354 | 91.0 | 728 | 0.9781 | 1.0 | 0.2298 |
| 0.1354 | 92.0 | 736 | 0.9765 | 1.0 | 0.2289 |
| 0.1354 | 93.0 | 744 | 0.9721 | 1.0 | 0.2275 |
| 0.1354 | 94.0 | 752 | 0.9741 | 1.0 | 0.2282 |
| 0.1354 | 95.0 | 760 | 0.9737 | 1.0 | 0.2291 |
| 0.1354 | 96.0 | 768 | 0.9715 | 1.0 | 0.2278 |
| 0.1354 | 97.0 | 776 | 0.9715 | 1.0 | 0.2282 |
| 0.1354 | 98.0 | 784 | 0.9728 | 1.0 | 0.2282 |
| 0.1354 | 99.0 | 792 | 0.9735 | 1.0 | 0.2287 |
| 0.1046 | 100.0 | 800 | 0.9737 | 1.0 | 0.2288 |
Framework versions
- Transformers 4.42.3
- Pytorch 2.1.0+cu118
- Datasets 2.20.0
- Tokenizers 0.19.1