Nunchaku still fails after installation: align the node, wheel and model

Treat these as three components

The Nunchaku installation guide distinguishes the library wheel, the ComfyUI extension and the bundled Python environment. It requires a wheel that matches the actual Python, Torch and platform combination.1

python -m pip show nunchaku torch
python -c "import sys, platform; print(sys.version); print(platform.system(), platform.machine())"

Compare those results with the wheel's Python, OS, architecture and Torch tags. The second command reads standard-library information; it does not test a GPU. Do not rename an incompatible wheel to make its filename look acceptable.

Our suggested branches

For a missing module, verify the interpreter. For an unsupported-wheel error, verify the tags. For a native-library failure, retain the complete loading error. For a model-initialization failure, check the quantized weights expected by the node.

Avoid upgrading the entire Torch stack merely to fit an arbitrary wheel. Validate changes in a copy, then load an upstream-supported model through the intended node. Record import, model loading and generation separately. Neither a successful install nor a publisher benchmark establishes compatibility or speed on your hardware.

Related checks

• llama_cpp failures: distinguish installation, native builds and model loading
• Keep a reproducible ComfyUI repair record without exposing secrets

Research draft. No local/GPU reproduction or independent editorial sign-off. Upstream statements and our recommendations are distinguished.