README.md

metadata

title: ModiTrans
emoji: 📜
colorFrom: yellow
colorTo: red
sdk: gradio
sdk_version: 5.0.0
app_file: app.py
pinned: false
license: mit
short_description: Transliterate historic Modi script images to Devanagari

ModiTrans — Modi Script to Devanagari

Gradio demo for historyHulk/ModiTrans-12B-Gemma-Teacher, a LoRA adapter on google/gemma-3-12b-it that transliterates scanned images of historic Modi script into modern Devanagari text.

Setup notes before deploying

1. Accept the gated model license. Visit the model page while logged in and accept the access conditions. You'll also need to accept the license for the base model google/gemma-3-12b-it.
2. Add an HF_TOKEN secret to this Space (Settings → Variables and secrets) using a token from an account that has been granted access to both gated repos.
3. Hardware: this Space is configured for ZeroGPU (@spaces.GPU in app.py). ZeroGPU requires a PRO or verified Hugging Face account. If you don't have ZeroGPU access, remove the spaces import/decorator in app.py and instead select a dedicated GPU hardware tier (T4 or A10G) in the Space settings.

Files

• app.py — Gradio app + inference logic
• requirements.txt — Python dependencies