Introduction
Flux1-dev is a streamlined model checkpoint designed for ComfyUI users with limited VRAM, specifically under 24GB. It includes two text encoders within one safetensor, optimizing performance for users with constrained resources.

Architecture
The model architecture integrates two text encoders into a single safetensor, allowing seamless operation within the ComfyUI environment. This design ensures efficient processing and compatibility with lower VRAM capacities.

Training
Details on the training process are not specified in the README. The focus is on optimizing the model for efficient use in environments with restricted GPU capacity.

Guide: Running Locally

  1. Ensure your system meets the requirements, ideally having less than 24GB of VRAM.
  2. Download the flux1-dev checkpoint from the specified source.
  3. Use the Load Checkpoint node in ComfyUI to integrate the model.
  4. For enhanced performance, consider using cloud GPUs from providers like AWS, Google Cloud, or Azure, which offer scalable GPU resources.

License
The model is distributed under the flux-1-dev-non-commercial-license. For more details, refer to the license document.

More Related APIs