kiranpantha/whisper-large-v3-nepali
This model is a fine-tuned version of kiranpantha/whisper-large-v3-nepali on the OpenSLR54 dataset. It achieves the following results on the evaluation set:
- Loss: 0.4135
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: 0.001
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 10
- num_epochs: 10
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| No log | 1.0 | 14 | 0.3783 |
| 0.5414 | 2.0 | 28 | 0.3666 |
| 0.5414 | 3.0 | 42 | 0.3630 |
| 0.1128 | 4.0 | 56 | 0.3761 |
| 0.1128 | 5.0 | 70 | 0.3991 |
| 0.0601 | 6.0 | 84 | 0.4439 |
| 0.0601 | 7.0 | 98 | 0.4124 |
| 0.0442 | 8.0 | 112 | 0.4095 |
| 0.0239 | 9.0 | 126 | 0.4067 |
| 0.0239 | 10.0 | 140 | 0.4135 |
Framework versions
- PEFT 0.14.0
- Transformers 4.47.1
- Pytorch 2.5.1+cxx11.abi
- Datasets 3.2.0
- Tokenizers 0.21.0