Search results

Node working well, just a comment re llama-cpp-python using CPU instead of GPU, and a Q on expected run duration.

Status: Open.

#16 In 1038lab/ComfyUI-MiniCPM;

· 808charlie opened on Dec 10, 2025

This plugin is garbage; it increased my VAE decoding time by several times.

Status: Open.

#15 In 1038lab/ComfyUI-MiniCPM;

· ouyangixaobo opened on Dec 7, 2025

The ComfyUI log did not display any errors, which were treated as results.

Status: Open.

#12 In 1038lab/ComfyUI-MiniCPM;

· bookong opened on Oct 21, 2025

Global PyTorch settings cause massive memory usage and compilation delays across all ComfyUI workflows

Status: Open.

#10 In 1038lab/ComfyUI-MiniCPM;

· Shoe-Lac3 opened on Sep 25, 2025

Model loading error: Model initialization failed: No package metadata was found for bitsandbytes

Status: Open.

#9 In 1038lab/ComfyUI-MiniCPM;

· shimmery1210 opened on Sep 19, 2025

Model loading error: Model initialization failed: cannot import name 'Qwen3Config' from 'transformers'

Status: Open.

#8 In 1038lab/ComfyUI-MiniCPM;

· amirkerr opened on Sep 14, 2025

Error: Model loading error: Model initialization failed: exception: access violation writing 0x0000000000000108

Status: Open.

#7 In 1038lab/ComfyUI-MiniCPM;

· mkat2 opened on Sep 12, 2025

Error: Model loading error: Model initialization failed: 'Resampler' object has no attribute '_initialize_weights'

Status: Open.

#6 In 1038lab/ComfyUI-MiniCPM;

· kajfblsdkgnlsndg opened on Sep 3, 2025

Can you add the ability to unload the GPU after use? Otherwise, subsequent processes will not be able to execute.

Status: Open.

#5 In 1038lab/ComfyUI-MiniCPM;

· w2253543w opened on Aug 31, 2025