Flux LORA Training CLI

This is a command-line interface (CLI) tool for creating a Flux LORA (LoRA: Low-Rank Adaptation) using replicate.com. Flux LORA is a technique for fine-tuning a pre-trained model to learn a new concept.

Prerequisites

• Python 3.6 or higher
• A replicate.com account and API token
• A GitHub account

Installation

1. Clone this repository:

git clone https://github.com/your-username/flux-lora-train-cli.git
cd flux-lora-train-cli

2. Install the required Python packages:

pip install -r requirements.txt

3. Set your replicate.com API token as an environment variable:

export REPLICATE_API_TOKEN=your-api-token

Usage

python main.py owner model_name image_path token

• owner: Your GitHub username.
• model_name: The name for your LORA model.
• image_path: The path to the zip file containing training images.
• token: The token for the concept to be learned.

Example

python main.py my-github-username my-lora-model path/to/images.zip my-token

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgments

• replicate.com for providing the infrastructure for training the LORA model.
• flux-dev-lora-trainer for the LORA training util.