metadata

title: Tamil Audio to Text
emoji: ▶︎ •၊၊||၊|။||||။‌‌‌‌‌၊|• 0:10 ➤ 📄
colorFrom: blue
colorTo: yellow
sdk: gradio
app_file: app.py
pinned: false
license: apache-2.0

Whisper Small Tamil - Hugging Face Demo

This repository hosts a demo for the Whisper Small Tamil model, fine-tuned for Tamil speech recognition. This model is based on OpenAI's Whisper-Small and has been trained to improve Automatic Speech Recognition (ASR) for Tamil language inputs.

🚀 Demo

Try the model directly on 🤗 Hugging Face Spaces.

📝 Model Details

Base Model: OpenAI Whisper-Small
Fine-tuned for: Tamil ASR
Dataset Used: Common Voice Tamil & other curated datasets
Supports: Tamil speech-to-text transcription

🔧 How to Use

You can use this model in Python with the transformers library:

from transformers import pipeline

# Load model from Hugging Face Hub
asr_pipeline = pipeline("automatic-speech-recognition", model="deepakkumar07/whisper-small-tamil")

# Transcribe an audio file
result = asr_pipeline("path/to/audio.wav")
print(result["text"])

📊 Performance

This model is optimized for Tamil speech but may still have minor errors in transcription, especially with noisy audio or mixed-language inputs. Contributions and improvements are welcome!

📌 Training Details

Fine-tuned using the Hugging Face Transformers and datasets libraries.
Trained on GPUs for better performance.
Supports streaming inference for real-time transcription.

💡 Applications

Tamil voice-to-text conversion
Subtitling and transcription services
Voice-controlled Tamil applications

🤝 Contributing

If you find any issues or want to improve the model, feel free to open a PR or reach out!

📜 License

This model is released under an open license. Please refer to OpenAI's original Whisper license for base model terms.

For more details, check out the Hugging Face model page. 🚀