metadata
base_model: distil-whisper/distil-large-v2
library_name: peft
license: mit
metrics:
- wer
tags:
- generated_from_trainer
model-index:
- name: distil-whisper-large-v2
results: []
distil-whisper-large-v2
This model is a fine-tuned version of distil-whisper/distil-large-v2 on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 7.1657
- Wer Ortho: 100.0
- Wer: 100.0
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: 4
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: constant_with_warmup
- lr_scheduler_warmup_steps: 500
- num_epochs: 10
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer Ortho | Wer |
|---|---|---|---|---|---|
| 8.0272 | 1.0 | 3 | 7.0930 | 100.0 | 100.0 |
| 8.4161 | 2.0 | 6 | 7.1227 | 100.0 | 100.0 |
| 8.1907 | 3.0 | 9 | 7.1657 | 100.0 | 100.0 |
| 8.12 | 4.0 | 12 | 7.1657 | 100.0 | 100.0 |
Framework versions
- PEFT 0.13.2
- Transformers 4.45.2
- Pytorch 2.4.1+cu121
- Datasets 3.0.1
- Tokenizers 0.20.1