|
|
---
|
|
|
title: SIMBOTI Live
|
|
|
emoji: 🎙️
|
|
|
colorFrom: green
|
|
|
colorTo: blue
|
|
|
sdk: gradio
|
|
|
sdk_version: 5.12.0
|
|
|
app_file: live_app.py
|
|
|
pinned: false
|
|
|
---
|
|
|
|
|
|
# 🎙️ SIMBOTI Live (Powered by TranslateGemma 4B)
|
|
|
|
|
|
**SIMBOTI** = **S**peech **I**ntelligent **M**ultimodal **B**ot for **O**utreach **T**ranslation **I**mplementation.
|
|
|
|
|
|
A privacy-first, multimodal translation agent for clinical care.
|
|
|
|
|
|
## 🚀 Features
|
|
|
- 💬 **Text-to-Text** translation
|
|
|
- 🎙️ **Live Speech** translation (microphone input)
|
|
|
- 📄 **Document Scan** (image/photo translation)
|
|
|
- 🎥 **Video OCR** (translate text in videos)
|
|
|
- 🔊 **Text-to-Speech** audio output for patients
|
|
|
|
|
|
## 🌍 Supported Languages
|
|
|
Polish, Romanian, Punjabi, Urdu, Portuguese, Spanish, Arabic, Bengali, Gujarati, Italian.
|
|
|
|
|
|
## 🚀 Deployment (ZeroGPU)
|
|
|
1. Create a [Hugging Face Space](https://huggingface.co/spaces/new) (SDK: Docker, Hardware: ZeroGPU).
|
|
|
2. Upload all files from this folder.
|
|
|
3. Run! (First start downloads ~8GB model)
|
|
|
|
|
|
## 🔒 Privacy
|
|
|
All inference runs within the ephemeral ZeroGPU container. No data is logged or stored.
|
|
|
|
