---
title: Atalink-TTS-Talker
emoji: 😭
colorFrom: purple
colorTo: blue
sdk: gradio
sdk_version: 5.45.0
python_version: 3.10.18
app_file: app.py
pinned: false
license: mit
---

# 😭 Atalink-TTS-Talker

A Hugging Face Space powered by **Gradio**.  
This project demonstrates **SadTalker** with local environment setup, backup/restore guides, and Docker deployment.

## Reference: [Spaces Config Docs](https://huggingface.co/docs/hub/spaces-config-reference)

---

### Use local:

## 🖥️ Local Setup

### 1. Python environment

- Use **Python 3.10**

```bash
python -m venv .venv
source .venv/bin/activate
```

### 2. Install PyTorch with CUDA 12.4

```bash
pip install torch==2.4.0+cu124 torchaudio==2.4.0+cu124 torchvision==0.19.0   --extra-index-url https://download.pytorch.org/whl/cu124
```

### 3. Install dependencies

```bash
sudo apt-get update
sudo apt-get install sox ffmpeg
```

---

## 💾 Backup & Restore Volumes

> ⚠️ Make sure you **mount the backup folder** into the container before running these commands.

### 🔹 Backup

```bash
# Cache
tar -czvf /backup/data_cache_backup.tar.gz /home/user/.cache
tar -czvf backup\data_backup.tar.gz "C:\Users\<user>\.cache\huggingface" (windows)

# GFPGAN data
tar -czvf /backup/data_gfpgan_backup.tar.gz /home/user/app/gfpgan
tar -czvf backup\data_gfpgan_backup.tar.gz gfpan/weights (windows)
```

### 🔹 Restore

```bash
# Restore cache
mkdir -p /home/user/.cache
cd /home/user/.cache
tar -xzvf /backup/data_cache_backup.tar.gz --strip 1

# Restore GFPGAN
mkdir -p /home/user/app/gfpgan
cd /home/user/app/gfpgan
tar -xzvf /backup/data_gfpgan_backup.tar.gz --strip 1
```

---

## 🚀 Running the App

1. Create and activate Python 3.10 environment
2. Prepare environment:

   ```bash
   python utils/prepare_environment.py
   ```

3. Start with Docker Compose:

   ```bash
   docker compose up -d
   ```

4. If you change code, rebuild:
   ```bash
   docker compose up -d --build
   ```

---

✨ Done! You can now run **SadTalker** locally or deploy with Docker.