Using the Hugging Face Demo Version

1. Step 1: Access the Demo Version - Click on the Hugging Face link provided in the description. This will take you to the MagicFace demo page.
2. Step 2: Upload a Face Set - Upload one or multiple pictures of the same face. This helps the model maintain consistency in facial features.
3. Step 3: Select Gender - Choose the correct gender. This step is crucial for achieving accurate results.
4. Step 4: Enter a Prompt - Type a detailed prompt describing how you want the portrait to look, or select a character from the provided list.
5. Step 5: Generate the Image - Click the "Generate Magic" button. MagicFace will process your request and create a stunning portrait. If you want to refine your results, check the tool's guide for additional tips.

Running MagicFace on Google Colab

1. Step 1: Open Google Colab Link - Click on the Google Colab link provided in the description. This will open the MagicFace notebook.
2. Step 2: Set Up the Environment - Click on "Runtime" and select "Change Runtime Type." Choose "GPU" if it is not already selected. Click "Save."
3. Step 3: Connect to Runtime - Click on the "Connect" button to establish a connection.
4. Step 4: Install Dependencies - Run the first installation cell. This will install all necessary dependencies.
5. Step 5: Launch MagicFace - Run the main cell. The first-time execution will download the required models. Once the installation is complete, you will see a "Gradio URL." Click on this URL to open MagicFace in your browser.
6. Step 6: Use MagicFace - Upload a face set or an image. Enter a detailed prompt. Click "Generate" to create your portrait.

The results are truly remarkable, making MagicFace a fantastic tool for character transformation.

Installing MagicFace Locally

1. Step 1: Install Dependencies - Go to the Hugging Face demo page. You need the following software installed on your system:
   Python 3.10.11
   FFmpeg
   and an Nvidia graphics card.
2. Step 2: Clone the Repository - Click on the three dots in the Hugging Face repository and select "Clone." Copy the Git clone link. Open the folder where you want to install MagicFace. In the folder path, type
   cmd
   and press Enter to open the command prompt. Paste the Git clone command and press Enter to clone the repository.
3. Step 3: Navigate to the Folder - Use the
   cd
   command to navigate inside the cloned folder.
4. Step 4: Set Up a Virtual Environment - Run the following command to create a virtual environment:
   python -m venv venv
   Activate the virtual environment using the appropriate command for your OS.
5. Step 5: Install PyTorch - Go to the official PyTorch website. Copy the installation command based on your CUDA version. Paste the command in the terminal and run it.
6. Step 6: Install Requirements - Open the
   requirements.txt
   file. Add
   gradio
   and
   spaces
   to the list. Remove
   XFormers
   from the file. Save the file. Run the following command to install the dependencies:
   pip install -r requirements.txt
7. Step 7: Run MagicFace - Execute the main script:
   python main.py
   The first-time execution will download the required models. Once completed, you will see a "Local URL." Copy and paste this URL into your web browser to launch MagicFace.
8. Step 8: Create a Batch File for Easy Access - To simplify launching MagicFace in the future: Open Notepad. Type the following command:
   @echo off cd /d path_to_magicface venv\Scripts\activate python main.py
   Save the file as
   run.bat
   Now, you can launch MagicFace with just one click.