<!-- 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. -->
whisper-tiny-v2-ta_e6_bp
This model is a fine-tuned version of openai/whisper-tiny on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.3422
- Wer: 42.5666
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: 1.5e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 8
- total_train_batch_size: 64
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 3000
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.491 | 0.21 | 300 | 0.4824 | 58.7409 |
| 0.4042 | 0.42 | 600 | 0.4242 | 56.5617 |
| 0.3571 | 0.63 | 900 | 0.3949 | 51.5254 |
| 0.3132 | 0.84 | 1200 | 0.3754 | 45.9080 |
| 0.2878 | 1.05 | 1500 | 0.3621 | 44.8910 |
| 0.2801 | 1.26 | 1800 | 0.3545 | 43.6320 |
| 0.2708 | 1.47 | 2100 | 0.3496 | 43.0508 |
| 0.2718 | 1.68 | 2400 | 0.3442 | 43.0024 |
| 0.261 | 1.9 | 2700 | 0.3422 | 42.5666 |
Framework versions
- Transformers 4.28.0.dev0
- Pytorch 1.12.1
- Datasets 2.11.0
- Tokenizers 0.13.3
