👗 Outfit Anyone - Virtual Trial Room with ComfyUI Workflow

Welcome to Outfit Anyone, a virtual trial room that allows users to try on different outfits using a unique LoRA model trained with the "Outfit Anyone" ComfyUI workflow. This project leverages deep learning to provide a futuristic shopping experience, allowing users to visualize themselves in various clothing styles from the comfort of their homes.

🌟 Features

• Virtual Try-On: Easily visualize various outfits with a simple UI powered by the "Outfit Anyone" workflow.
• Custom LoRA Model: Trained a LoRA model using over 100 images, offering a wide range of outfit simulations, from casual wear to formal attire.
• Easy Integration: Utilizes AWS SageMaker, Automatic 1111, Kohya SS, python, Jupyter Notebooks for model training and deployment.
• Realistic Simulations: Generated outputs showcase the LoRA model's capability to simulate outfits in diverse scenarios, bringing fashion retail to a whole new digital level!

🚀 Tech Stack

• ComfyUI: Streamlined the user interface for virtual trial room experience.
• LoRA Model: Trained on a diverse dataset for high-quality outfit simulations.
• AWS SageMaker: Utilized for scalable model training.
• Python & Jupyter Notebooks: For model training, data preprocessing, and analysis.
• Terminal Commands: To manage workflow and troubleshoot during development.

📸 Example Outputs

Here’s a sneak peek of some outfits generated by the virtual trial room:

Note: More examples can be found in the results/ folder!

🛠️ Getting Started

Follow these instructions to set up the project on your local machine.

Prerequisites

• Python 3.8+
• ComfyUI installed
• AWS SageMaker account (optional for model training)
• Jupyter Notebooks

Installation

1. Clone this repository:

   git clone https://github.com/your-username/outfit-anyone-comfyui-workflow.git
   cd outfit-anyone-comfyui-workflow

2. Install the required Python packages:

   pip install -r requirements.txt

3. Set up ComfyUI: Follow the ComfyUI setup guide to integrate it with this project.

4. Prepare the Dataset (Optional):

   • Add your dataset to the data/ folder.
   • Run the data preprocessing script in notebooks/data_preprocessing.ipynb.

Usage

1. Train the LoRA Model (Optional):

   • Open notebooks/train_lora_model.ipynb.
   • Follow the instructions to train your model using your dataset.

2. Run the Virtual Trial Room:

   python src/main.py

3. Use the ComfyUI Interface:

   • Open your browser and navigate to localhost:8000 to access the virtual trial room.
   • Upload an image and explore different outfit options!

🧠 How It Works

1. Data Collection: The LoRA model is trained using a diverse set of images, enabling it to generalize across different body shapes and outfits.
2. Model Training: The model is fine-tuned using a combination of AWS SageMaker, Jupyter Notebooks, and terminal commands, optimizing for accuracy and performance.
3. Outfit Simulation: Through the ComfyUI workflow, users can select an outfit, and the model generates a realistic trial room image, presenting how the outfit might look on them.

📝 Project Structure

Here's a breakdown of the key components of this project:

• src/: Contains the main code files and UI scripts.
• assets/: Stores images and icons used in the UI.
• models/: Pre-trained LoRA models (or instructions on how to download them if they're large).
• results/: Example outputs from the virtual trial room.
• notebooks/: Jupyter Notebooks for data preprocessing, model training, and analysis.
• requirements.txt: Lists the Python dependencies for this project.

🎯 What's Next?

• Adding support for multiple outfit categories.
• Integrating more advanced deep learning techniques for even more realistic simulations.
• Enhancing the user interface for a seamless virtual shopping experience.