Qwen Image Edit crashes in ComfyUI when blur nodes are added first
UnorthodoxyMedia · reddit · 2026-08-04
A user reports that Qwen Image Edit breaks when an edge-preserving blur or regular blur node is inserted before the editor in ComfyUI.
The workflow works when the image goes directly from input to Qwen nodes, but adding blur causes TextEncodeQwenImageEditPlus to fail with a shape mismatch error: shape '[-1, 3, 2, 14, 14]' is invalid for input of size 1229312.
More from Multimodal
- ChatGPT Work turns a Backrooms prompt into a seamless 5-second looping video — goodside · 2026-08-04
- Ostris AI Toolkit now supports MiniMax H3 video-model training — RayHell666 · 2026-08-04
- MiniMax video model users ask how to match input resolution while praising its output quality — Careless-Constant-33 · 2026-08-04
- Minimax H3’s John Wick-style demo turns into a viral AI movie-moment — Parogarr · 2026-08-04
- A ComfyUI workflow pushes Wan 2.2 image-to-video out to roughly 45 seconds — embryo10 · 2026-08-04
- A tiny H3 demo detail: the weights shift during the conversation — Oatilis · 2026-08-04