metadata
library_name: peft
license: apache-2.0
base_model: openai/whisper-large-v3
tags:
- generated_from_trainer
datasets:
- common_voice_11_0
metrics:
- wer
model-index:
- name: whisper-hindi-peft
results: []
language:
- hi
whisper-hindi-peft
This model is a fine-tuned version of openai/whisper-large-v3 on the common_voice_11_0 dataset. It achieves the following results on the evaluation set:
- Loss: 0.1044
- Wer: 0.2829
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: 0.0001
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 32
- 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: 100
- training_steps: 1000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.4265 | 1.4689 | 200 | 0.1282 | 0.3696 |
| 0.2581 | 2.9377 | 400 | 0.0941 | 0.2797 |
| 0.1579 | 4.4103 | 600 | 0.0984 | 0.2850 |
| 0.1285 | 5.8791 | 800 | 0.0999 | 0.2795 |
| 0.0862 | 7.3516 | 1000 | 0.1044 | 0.2829 |
Framework versions
- PEFT 0.14.0
- Transformers 4.48.3
- Pytorch 2.5.1+cu124
- Datasets 3.3.2
- Tokenizers 0.21.0