metadata
library_name: peft
license: apache-2.0
base_model: openai/whisper-medium
tags:
- generated_from_trainer
metrics:
- wer
model-index:
- name: whisper-medium-cdsd1h-lora
results: []
whisper-medium-cdsd1h-lora
This model is a fine-tuned version of openai/whisper-medium on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.7490
- Wer: 69.0689
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: 16
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 100
- num_epochs: 5
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.8319 | 1.0 | 559 | 0.8337 | 78.2453 |
| 0.5523 | 2.0 | 1118 | 0.7616 | 73.5900 |
| 0.3329 | 3.0 | 1677 | 0.7141 | 70.2328 |
| 0.1729 | 4.0 | 2236 | 0.7248 | 70.4566 |
| 0.0681 | 5.0 | 2795 | 0.7490 | 69.0689 |
Framework versions
- PEFT 0.13.2
- Transformers 4.44.2
- Pytorch 2.0.0+cu118
- Datasets 2.19.1
- Tokenizers 0.19.1