README.md

---
library_name: transformers
language:
- si
license: apache-2.0
base_model: openai/whisper-small
tags:
- generated_from_trainer
datasets:
- Lingalingeswaran/asr-sinhala-dataset_json_v1
metrics:
- wer
model-index:
- name: Whisper Small sinhala v3 - Lingalingeswaran
  results:
  - task:
      name: Automatic Speech Recognition
      type: automatic-speech-recognition
    dataset:
      name: Lingalingeswaran/asr-sinhala-dataset_json_v1
      type: Lingalingeswaran/asr-sinhala-dataset_json_v1
      args: 'config: si, split: test'
    metrics:
    - name: Wer
      type: wer
      value: 46.457654723127035
---

<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->

# Whisper Small sinhala v3 - Lingalingeswaran

This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the Lingalingeswaran/asr-sinhala-dataset_json_v1 dataset.
It achieves the following results on the evaluation set:
- Loss: 0.2086
- Wer: 46.4577

## 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: 16
- 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: 500
- training_steps: 3000
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch  | Step | Validation Loss | Wer     |
|:-------------:|:------:|:----:|:---------------:|:-------:|
| 0.1852        | 1.7606 | 1000 | 0.1875          | 50.9772 |
| 0.0602        | 3.5211 | 2000 | 0.1886          | 47.5774 |
| 0.0238        | 5.2817 | 3000 | 0.2086          | 46.4577 |


### Framework versions

- Transformers 4.48.1
- Pytorch 2.5.1+cu121
- Datasets 3.2.0
- Tokenizers 0.21.0


## Example Usage

Here is an example of how to use the model for Sinhala speech recognition with Gradio:

```python
import gradio as gr
from transformers import pipeline

# Initialize the pipeline with the specified model
pipe = pipeline(model="Lingalingeswaran/whisper-small-sinhala_v3")

def transcribe(audio):
    # Transcribe the audio file to text
    text = pipe(audio)["text"]
    return text

# Create the Gradio interface

iface = gr.Interface(
    fn=transcribe,
    inputs=gr.Audio(sources=["microphone", "upload"], type="filepath"),
    outputs="text",
    title="Whisper Small Sinhala",
    description="Realtime demo for Sinhala speech recognition using a fine-tuned Whisper small model.",
)

# Launch the interface
if __name__ == "__main__":
    iface.launch()