whisper-large-v2-Irish
This model is a fine-tuned version of kpriyanshu256/whisper-large-v2-br-1000-32-1e-05 on the Common Voice 11.0 dataset. It achieves the following results on the evaluation set:
- Loss: 1.2167
- Wer: 42.3635
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: 4
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 8
- total_train_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 50
- training_steps: 800
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.4381 | 3.0 | 100 | 0.9324 | 47.2416 |
| 0.0465 | 6.01 | 200 | 1.0565 | 45.6156 |
| 0.0169 | 9.02 | 300 | 1.0763 | 43.2636 |
| 0.0063 | 12.02 | 400 | 1.1362 | 44.0476 |
| 0.0024 | 15.03 | 500 | 1.1534 | 42.7410 |
| 0.0011 | 18.03 | 600 | 1.1959 | 42.3926 |
| 0.0009 | 21.04 | 700 | 1.2123 | 42.1893 |
| 0.0008 | 24.04 | 800 | 1.2167 | 42.3635 |
Framework versions
- Transformers 4.26.0.dev0
- Pytorch 1.13.0+cu117
- Datasets 2.7.1.dev0
- Tokenizers 0.13.2
- Downloads last month
- 10
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.