README_INSTRUCTIONS.md

TTS Online Rubric (Gradio)

This is a minimal Gradio-based UI for running TTS rubric evaluations. It's designed to be deployed on Hugging Face Spaces or run locally.

Features:

• Browse audio files from sample-audios/.
• Play reference audio and upload system outputs to compare.
• Rate outputs on three sliders (nativeness, naturalness, overall quality) and add comments.
• Responses are appended to responses.csv.

How to run locally:

1. Create a Python environment and install requirements:

python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

2. Add your reference audio files under sample-audios/.

3. Run the app:

python app.py

4. Open the URL shown in the terminal (default http://127.0.0.1:7860).

Deploying to Hugging Face Spaces:

• Create a new Space using the Gradio SDK (python) and push this repo. Ensure requirements.txt is present. The app will run automatically.

Notes and next steps:

• The current UI accepts system outputs via file upload. You may prefer a fixed set of system files stored in a directory for side-by-side playback.
• Consider adding authentication or locking to avoid duplicate/corrupt CSV writes under heavy load.