File size: 1,217 Bytes
842bbce
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
# 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:

```bash
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:

```bash
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.