gguf-node 0.1.2

gguf node for comfyui

gguf node for comfyui

install it via pip/pip3

pip install gguf-node

enter the user menu by (if no py command; use python/python3 instead)

py -m gguf_node

Please select:

1. download the full pack
2. clone the node only

Enter your choice (1 to 2): _

for new/all user(s)

opt 1 to download the compressed comfy pack (7z), decompress it, and run the .bat file striaght (idiot option)

for existing user/developer(s)

opt 2 to clone the gguf repo to the current directory (either navigate to ./ComfyUI/custom_nodes first or drag and drop there after the clone)

alternatively, you could execute the git clone command to perform that task (see below):

• navigate to ./ComfyUI/custom_nodes
• clone the gguf repo to that folder by

git clone https://github.com/calcuis/gguf

same operation for the standalone pack; then you should be able to see it under Add Node > check the dropdown menu for gguf

🐷🐷📄 for the latest update, gguf-connector deployment copy is now attached to the node itself; don't need to clone it to site-packages; and, as the default setting in comfyui is sufficient; no dependencies needed right away 🙌 no extra step anymore

other(s): get it somewhere else trustworthy/reliable

you are also welcome to get the node through other available channels, i.e., comfy-cli, comfyui-manager (search gguf from the bar; and opt to install it there should be fine; see picture below), etc. gguf node is no conflict with the popular comfyui-gguf node (can coexist; then you could test them both; or mixing up use, switch in between freely, all for your own purpose); and is more lightweight (no dependencies needed), more function (i.e., built-in gguf convertor), compatible with the latest version numpy and other updated libraries come with comfyui

for the demo workflow (picture) above, you could get the test model gguf here, test it rather you can get the similar outcome or not

setup (in general)

• drag gguf file(s) to diffusion_models folder (./ComfyUI/models/diffusion_models)
• drag clip or encoder(s) to text_encoders folder (./ComfyUI/models/text_encoders)
• drag controlnet adapter(s), if any, to controlnet folder (./ComfyUI/models/controlnet)
• drag lora adapter(s), if any, to loras folder (./ComfyUI/models/loras)
• drag vae decoder(s) to vae folder (./ComfyUI/models/vae)

workflow

• drag the workflow json file to the activated browser; or
• drag any generated output file (i.e., picture, video, etc.; which contains the workflow metadata) to the activated browser

simulator

• design your own prompt; or
• generate a random prompt/descriptor by the simulator (though it might not be applicable for all)

convertor (new feature: convert safetensors to gguf)

• drag safetensors file(s) to diffusion_models folder (./ComfyUI/models/diffusion_models)
• choose the last option from the gguf menu: GGUF Convertor (Alpha)
• select your safetensors model inside the box; don't need to connect anything; it works independently
• click Queue (run); then you can simply check the processing progress from console
• when it was done; the converted gguf file will be saved to the output folder (./ComfyUI/output)

reference

comfyui confyui_vlm_nodes comfyui-gguf gguf-comfy gguf-connector testkit

parent

node is a member of family gguf