GGUF quants for https://huggingface.co/ostris/Flex.2-preview using sd.cpp.
Some examples using q4_k with cfg set to 1 and guidance to 5 with dpm++2m sampling, differents schedulers and easycache set to 0.05,0.15,0.95.
$ sd --diffusion-model models/Flex.2-preview-q4_k.gguf --t5xxl models/flux-extra/t5xxl_q8_0.gguf --clip_l models/flux-extra/clip_l-f16.gguf -t 8 --vae models/flux-extra/ae-f16.gguf --sampling-method dpm++2m --scheduler simple --steps 40 --cfg-scale 1 --guidance 5 -p 'Photograph of the Alps. Vivid spring afternoon. Everything is perfectly in focus.' -v -W 1280 -H 768 --diffusion-fa --offload-to-cpu --vae-tiling --vae-tile-size 64 --easycache 0.05,0.15,0.95
Use t5xxl, clip_l and ae that are compatible with Flux.1.
Some image input experiments (I dont know what I am doing):
- Downloads last month
- 77
6-bit
8-bit
Model tree for Green-Sky/Flex.2-preview-GGUF
Base model
ostris/Flex.2-preview