Should I include some kind of GGUF text encoder in the workflow, or is it okay to use the default umt5_xxl_fp8_e4m3fn_scaled.safetensors?
Thanks for your work!
· Sign up or log in to comment