maybe one last question... i have the gguf loaders, and models.. but i cant pick the models in the loader. I assume i copied them to wrong place but couldnt find the information where to put it.
Is Kontext node "Load Image (from Outputs)" works only with node "save image", or I do something wrong and it can work with node "preview image " as well?
On MacOS I get the dreaded "Trying to convert Float8_e4m3fn to the MPS backend" Error. UGH!!
GGUF go into the unet folder
same issue here
The trick is to use the template workflow (the Basic one, not the grouped one) and swap out the load diffusion model node for a GGUF loader and model.
thanks so much.. i couldnt find the proper node or model for.. can you point me to the proper node/download to search for?
Models at https://huggingface.co/bullerwins/FLUX.1-Kontext-dev-GGUF/tree/main. In Manager, install gguf custom node from gguf to get the loader node. There’s been a lot of discussion about this today in the Comfy Org discord #comfy-on-mac channel.
maybe one last question... i have the gguf loaders, and models.. but i cant pick the models in the loader. I assume i copied them to wrong place but couldnt find the information where to put it.
Can you give me a hint?
ComfyUI/models/unet
Kudos to Black Forest Labs for this release as well as ComfyUI for native support! Great for the community!
Is it just me or is the workflow not in the templates? I've updated my Comfy, and am running in browser.
I had to:
pip install -r requirements.txt
to update the workflows library package.
That did it - thanks so much again!
Awesome, ty!
o ~
Is Kontext node "Load Image (from Outputs)" works only with node "save image", or I do something wrong and it can work with node "preview image " as well?
updated comfyui portable to latest version, no flux 1.1 template
Thank you for the very well designed and detailed blog post and documentation.
You are doing it great guys! Please keep go on 👍