Flux "micro-micro-finetune" experimental product from the Colorful Realistic series, based on Flux.1 Dev, using some materials from Colorful Realistic XL and some new materials. The current few rounds of fine-tuning have not achieved the expected effect. The limbs have deteriorated, and although it's less oily now, the materials have also introduced some blurriness. Therefore, I only uploaded the Fp8 version. The Fp16 version is too large and not worth wasting that storage space, and most people don't have enough VRAM. I mainly use it to generate images with my own trained LoRA. If you're interested, you can download it and try it out.

For image generation, please use the ComfyUI Flux standard text-to-image workflow, with parameters CFG 1, Simple Euler sampling, Steps 20 or above. The workflows you can download from my github repo. Of course, the original images are also included in the example images.

Comply with relevant laws and regulations. When using this model, do not infringe upon the legitimate rights and interests of others. Based on Flux.1 Dev training, it's completely free to use and not for commercial use.