FLUX.1-DEV-GGUF

Introduction

FLUX.1-DEV-GGUF is a direct GGUF conversion of the black-forest-labs/FLUX.1-dev model. It is designed for text-to-image and image-generation tasks. As a quantized model, it retains the original restrictions and license terms of the base model.

Architecture

The model is a quantized variant of the FLUX.1-dev architecture, which is optimized for running efficiently on compatible hardware without sacrificing performance. It is compatible with the GGUF library, allowing integration with specific tools and workflows.

Training

FLUX.1-DEV-GGUF is not a finetuned model but a quantized conversion, meaning it maintains the training characteristics of its base model, FLUX.1-dev. Users should refer to the original model documentation for detailed training information.

Guide: Running Locally

1. Download the Model: Obtain the model files from the Hugging Face repository.
2. Set Up ComfyUI-GGUF: Use the ComfyUI-GGUF custom node available on GitHub. Follow the installation instructions in the GitHub README.
3. Place Model Files: Copy the model files to the ComfyUI/models/unet directory.
4. Execute Locally: Run the model using compatible hardware. For optimal performance, consider using cloud GPUs that support GGUF models, such as those offered by AWS or Google Cloud.

License

The FLUX.1-DEV-GGUF is distributed under the flux-1-dev-non-commercial-license. Users must adhere to the terms outlined in the LICENSE.md file, as the model retains the original license terms of the base model.