Tamil Whisper ASR

Tamil Whisper ASR

Turns Tamil speech into text in a Gradio demo, using Whisper small adapted to Tamil with LoRA. For developers, researchers and learners working with Tamil voice recognition.

Made by skpranav22 on Hugging Face. These notes are Mindket's, written from its Space page and README; the app's own files and license are on Hugging Face.

About

What it is: This Space is a live speech recognition demo for Tamil. It runs Whisper small that was adapted for Tamil with Low-Rank Adaptation, known as LoRA, and it runs on ZeroGPU at Hugging Face. You give it Tamil speech and it writes the text.

Who it suits: Developers, researchers and learners who want to try Tamil speech recognition, and anyone studying how a small Whisper model is adapted to one language.

Industry | Who uses it | How they use it | Example
Speech technology | Speech developers | Test a Tamil recognition model | Speak a short Tamil sentence and read the text
Research | Language researchers | Check how a LoRA adapted Whisper handles Tamil | Compare its text with what was said
Education | Tamil learners | Hear and read the same sentence | Say a phrase and check the written result
Media | Content teams | Try transcription for Tamil audio | Run a short Tamil clip through the page

If you need a transcription service for long recordings, a short demo is only a first check.

Where to use it: It runs in two places.

• A web browser, on the Space's page at Hugging Face with nothing to install
• A local copy on Windows, macOS or Linux, which needs Python

Use case | Who | What they get back
Transcribe Tamil speech | Developers and learners | Tamil text from spoken audio
Judge the model | Researchers | A quick feel for the quality of the adapted model
Run your own copy | Builders | The same page on your own computer

How to set it up: Using it online needs nothing. A local copy needs a few things.

• Python and pip
• The packages listed in requirements.txt

1. Install the Hugging Face tools.
2. Download the Space into a folder.
3. Move into that folder.
4. Install Gradio and the packages from requirements.txt.
5. Start the app.

pip install -U huggingface_hub
hf download skpranav22/tamil-whisper-demo --repo-type space --local-dir tamil-whisper-demo
cd tamil-whisper-demo
pip install gradio==5.20.0
pip install -r requirements.txt
python app.py

How to use it: Open the Space page and give it Tamil audio. The page returns the recognized text. In a local copy, run python app.py to open the same page on your own computer.

Example: You give the page a short clip of someone speaking Tamil. A moment later, the recognized Tamil text appears on the page for you to read.

For newcomers and experts: Speech recognition turns spoken words into written text. Start with a clear, short clip and a quiet room, then check the text against what you said.

LoRA trains a small set of extra weights instead of the whole model, which keeps adaptation cheap. The Space names a merged model, so the LoRA weights are already folded into Whisper small. The demo runs on zero-a10g hardware.

You provide Tamil audio
The Gradio app passes it to the model
Whisper small with LoRA weights decodes it
The Tamil text appears on the page

Good to know: The README is brief, so the training data, accuracy and audio limits are not described.

• License: Apache 2.0
• Cost to run: none named
• Data it reaches: the audio you provide to the page

On Hugging Face, 13 November 2026

• Likes: 1
• Built with: gradio
• License: apache-2.0

Links

• Space: https://huggingface.co/spaces/skpranav22/tamil-whisper-demo
• Files: https://huggingface.co/spaces/skpranav22/tamil-whisper-demo/tree/main

Try it

Open the Space in your browser, or copy it to your own Hugging Face account from its page. Nothing in this kit runs by itself.