NONE GGUF node not existing.

[MrFries1111]

None GGUF node for comfyui is not loading.

[dpsjksjx900]

same here

[dolphin-electric]

Im going crazy I cant install the nodes

[LFd3v]

Make sure you have followed these instructions to install a fork of llama-cpp-python that supports vision:

https://github.com/1038lab/ComfyUI-QwenVL/blob/main/docs/LLAMA_CPP_PYTHON_VISION_INSTALL.md

If you do not do it, or have the stock llama-cpp-python module instead, it will not work. Right now, when ComfyUI was updated to v0.11.0 there was a problem with transformers in my setup (transformers was updated to v5.0.0), so you may need to wait for an update to QwenVL module as well (but GGUF still works for me).

[bandit765]

Make sure you have followed these instructions to install a fork of llama-cpp-python that supports vision:

https://github.com/1038lab/ComfyUI-QwenVL/blob/main/docs/LLAMA_CPP_PYTHON_VISION_INSTALL.md

If you do not do it, or have the stock llama-cpp-python module instead, it will not work. Right now, when ComfyUI was updated to v0.11.0 there was a problem with transformers in my setup (transformers was updated to v5.0.0), so you may need to wait for an update to QwenVL module as well (but GGUF still works for me).

Hi, how do I find the correct wheel if I am using 9070 xt, windows? I tried several wheels with cp312, since it says my python is 3.12 and I looked for versions that did not start with cuXXX in their files like this one llama_cpp_python-0.3.16-cp312-cp312-win_amd64

but its not working and keeps coming back with an error about my installation

[LFd3v]

@bandit765 Hi. I only see prebuilt wheels for CUDA, so unfortunately I think that you would need to build it manually to support or GPU (or just CPU using OpenBLAS btw):

https://github.com/JamePeng/llama-cpp-python#installation

For Win, make sure you check the requirements. I also had problem with some nodes marked as red, that was fixed after I fully closed ComfyUI and started it again to make sure it found the new llama-cpp-python library.

[Sepolian]

Only 2 lines of code need to change to get it working, plz see this pull request I opened.
#130

The problem seems to be new version of transformers(5.0) have different import name.

[LFd3v]

And checking again the screenshot in OP, that cut the top where the node titles appear, what is showing error are the non GGUF nodes, except Prompting Enhancing that actually requires both Transformers and GGUF. 🙄

[github-actions]

This issue has been automatically closed due to 30-day inactivity. Please feel free to reopen it with /reopen if needed.