openai/whisper-large-v2
This model is a fine-tuned version of openai/whisper-large-v2 on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.2637
- Wer: 10.1437
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: 1e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 4000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.2612 | 0.12 | 500 | 0.2687 | 12.4717 |
| 0.5072 | 0.25 | 1000 | 0.2606 | 12.2762 |
| 0.1023 | 1.05 | 1500 | 0.2436 | 10.0626 |
| 0.1379 | 1.18 | 2000 | 0.2447 | 11.1944 |
| 0.1237 | 1.3 | 2500 | 0.2412 | 11.0989 |
| 0.0684 | 2.11 | 3000 | 0.2715 | 10.2703 |
| 0.0925 | 2.23 | 3500 | 0.2553 | 10.2648 |
| 0.1484 | 3.03 | 4000 | 0.2637 | 10.1437 |
Framework versions
- Transformers 4.29.0
- Pytorch 1.14.0a0+44dac51
- Datasets 2.12.0
- Tokenizers 0.13.3
- Downloads last month
- 3
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social
visibility and check back later, or deploy to Inference Endpoints (dedicated)
instead.
Evaluation results
- WER on rishabhjain16/infer_so_chinesetest set self-reported13.830
- WER on rishabhjain16/infer_pf_germantest set self-reported31.460
- WER on rishabhjain16/infer_pf_italiantest set self-reported3.980
- WER on rishabhjain16/infer_pf_swedishtest set self-reported7.240
- WER on rishabhjain16/libritts_dev_cleantest set self-reported4.470
- WER on rishabhjain16/infer_mysttest set self-reported11.600
- WER on rishabhjain16/infer_cmutest set self-reported9.220
- WER on rishabhjain16/infer_pfstest set self-reported3.090