wav2vec-base-CREMA-sentiment-analysis
This model is a fine-tuned version of facebook/wav2vec2-base on an .Supreeta03/CREMA-audioData. It achieves the following results on the evaluation set:
- Loss: 1.3834
- Accuracy: 0.5756
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: 5e-05
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 128
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 50
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 1.7533 | 0.99 | 37 | 1.7438 | 0.3056 |
| 1.5942 | 1.99 | 74 | 1.5781 | 0.3409 |
| 1.4434 | 2.98 | 111 | 1.5167 | 0.3568 |
| 1.4327 | 4.0 | 149 | 1.4241 | 0.4156 |
| 1.3499 | 4.99 | 186 | 1.3148 | 0.4819 |
| 1.2126 | 5.99 | 223 | 1.2915 | 0.4887 |
| 1.1838 | 6.98 | 260 | 1.2672 | 0.4971 |
| 1.1141 | 8.0 | 298 | 1.2433 | 0.5046 |
| 1.1163 | 8.99 | 335 | 1.1968 | 0.5281 |
| 1.0099 | 9.99 | 372 | 1.1610 | 0.5516 |
| 0.9566 | 10.98 | 409 | 1.1547 | 0.5567 |
| 0.906 | 12.0 | 447 | 1.1565 | 0.5584 |
| 0.8275 | 12.99 | 484 | 1.1442 | 0.5718 |
| 0.7813 | 13.99 | 521 | 1.2570 | 0.5550 |
| 0.711 | 14.98 | 558 | 1.1654 | 0.5567 |
| 0.7146 | 16.0 | 596 | 1.4391 | 0.5323 |
| 0.6597 | 16.99 | 633 | 1.2309 | 0.5659 |
| 0.5579 | 17.99 | 670 | 1.2385 | 0.5760 |
| 0.5874 | 18.98 | 707 | 1.2609 | 0.5760 |
| 0.4905 | 20.0 | 745 | 1.3433 | 0.5777 |
| 0.5089 | 20.99 | 782 | 1.3727 | 0.5584 |
| 0.4414 | 21.99 | 819 | 1.3488 | 0.5676 |
| 0.3837 | 22.98 | 856 | 1.3572 | 0.5819 |
| 0.4419 | 24.0 | 894 | 1.5063 | 0.5651 |
| 0.387 | 24.99 | 931 | 1.4656 | 0.5659 |
| 0.4068 | 25.99 | 968 | 1.5354 | 0.5701 |
| 0.3496 | 26.98 | 1005 | 1.4607 | 0.5684 |
| 0.3579 | 28.0 | 1043 | 1.5049 | 0.5651 |
| 0.3135 | 28.99 | 1080 | 1.4441 | 0.5743 |
| 0.3612 | 29.99 | 1117 | 1.5329 | 0.5701 |
| 0.2599 | 30.98 | 1154 | 1.5920 | 0.5668 |
| 0.2517 | 32.0 | 1192 | 1.5633 | 0.5626 |
| 0.2439 | 32.99 | 1229 | 1.5979 | 0.5718 |
| 0.2891 | 33.99 | 1266 | 1.5590 | 0.5785 |
| 0.2564 | 34.98 | 1303 | 1.5978 | 0.5751 |
| 0.2132 | 36.0 | 1341 | 1.6400 | 0.5634 |
| 0.1882 | 36.99 | 1378 | 1.6309 | 0.5718 |
| 0.2027 | 37.99 | 1415 | 1.6376 | 0.5743 |
| 0.2555 | 38.98 | 1452 | 1.7064 | 0.5693 |
| 0.1872 | 40.0 | 1490 | 1.6575 | 0.5819 |
| 0.1891 | 40.99 | 1527 | 1.6606 | 0.5735 |
| 0.1795 | 41.99 | 1564 | 1.6507 | 0.5735 |
| 0.1931 | 42.98 | 1601 | 1.6627 | 0.5760 |
| 0.1574 | 44.0 | 1639 | 1.6944 | 0.5802 |
| 0.1842 | 44.99 | 1676 | 1.7082 | 0.5768 |
| 0.1859 | 45.99 | 1713 | 1.7004 | 0.5768 |
| 0.2088 | 46.98 | 1750 | 1.7002 | 0.5802 |
| 0.197 | 48.0 | 1788 | 1.6969 | 0.5751 |
| 0.1902 | 48.99 | 1825 | 1.6996 | 0.5743 |
| 0.1771 | 49.66 | 1850 | 1.6998 | 0.5743 |
Framework versions
- Transformers 4.38.2
- Pytorch 2.2.1+cu121
- Datasets 2.18.0
- Tokenizers 0.15.2
- Downloads last month
- 34