I converted fluxdfp16 to gguf 6k and fluxdp8 to gguf 5ks.and then merged gguf 6k and gguf 5ks plus my lora dreamfusion, badmilk2, generate with fluxgym not available on civitai, text and vae encoder included

https://civitai.com/models/623525/flux1d-fp8

sampler euler simple

steps 15 - 20

ps: for my part the model and above all a base that I use for my loras. and strangely he answers it well.

what is good for me will not necessarily be good for you.

version gguf

https://github.com/city96/ComfyUI-GGUF

https://huggingface.co/city96/t5-v1_1-xxl-encoder-gguf/tree/main