Flex.1-alpha

Flex.1-alpha : A powerful text-to-image generative model

Flex.1-alpha is a powerful text-to-image generation model based on the 8 billion parameter modified flow transformer architecture. It inherits the features of FLUX.1-schnell and guides the embedder through training to generate images without CFG.

Model properties

High quality image generation

Supports fine-tuning to suit specific tasks

Compatible with inference engines such as Diffusers and ComfyUI

Open source license (Apache 2.0)

8 billion parameters

target users

Creative design content creation artificial intelligence research developer

Usage scenarios

Generate images using ComfyUI

Use Diffusers to fine-tune your model

Participate in open source communities to improve models

Tutorial

1 Download the Flex.1-alpha model file (e.g. Flex.1-alpha .safetensors)

2 Put the file into the checkpoint folder of an inference engine such as ComfyUI

3 Configure the inference engine to generate images

4 Use AI-Toolkit for LoRA training (optional)

5 Generate images using guided embedders or fine-tuned models

Development background

It aims to solve the compression and optimization problems of image generation models and continuously improve model performance.