comfyui
https://github.com/comfyanonymous/comfyui
Python
The most powerful and modular diffusion model GUI, api and backend with a graph/nodes interface.
Triage Issues!
When you volunteer to triage issues, you'll receive an email each day with a link to an open issue that needs help in this project. You'll also receive instructions on how to triage issues.
Triage Docs!
Receive a documented method or class from your favorite GitHub repos in your inbox every day. If you're really pro, receive undocumented methods or classes and supercharge your commit history.
Python not yet supported24 Subscribers
Add a CodeTriage badge to comfyui
Help out
- Issues
- uni_pc: phi-coefficient recurrence runs in float32 on the device and breaks Wan 2.2 on MPS (b3 = 55 instead of 0.25); computing it in float64 fixes it
- LTX-2.3 speech is unintelligible with a GGUF Gemma text encoder since #15580: intermediate hidden states alias through `GGMLTensor.clone()`
- Pixal3DMultiViewConditioning keeps ~3 GiB per view of NAF features in system RAM (12 GiB for 4 views); /free can't release them without unloading first
- Add ES module extension as another `endswith` path to fix stale cache for frontend made in JS modules.
- new bug problem
- TextGenerate: use_default_template semantics are counterintuitive and system_prompt handling is confusing
- Fix aliased Llama2_ intermediate_output snapshots when clone() is a no-op
- llama2 module tries to use flash-attention which only works with cuda 130, but not all graphics cards can use it.
- Support comfy-kitchen's AWQ W4A16 layout as the awq_w4a16 quant format
- Fix crash when dynamic VRAM prefetch cannot grow its cast buffer
- Docs
- Python not yet supported