The most powerful and modular diffusion model GUI, api and backend with a graph/nodes interface. The fastest local inference engine in the world.
Commits on master
fix(userdata): catch invalid workflow filenames (#9434) (#9445)
0737b7e saurabh-pingale committed 1 year ago
Support for Qwen Diffsynth Controlnets canny and depth. (#9465)
0963493 comfyanonymous committed 1 year ago
Add that qwen edit model is supported to readme. (#9463)
e73a9db comfyanonymous committed 1 year ago
LTXV: fix key frame noise mask dimensions for when real noise mask exists (#9425)
fe01885 Harel Cain committed 1 year ago
Fix bf16 precision issue with qwen image embeddings. (#9441)
8d38ea3 comfyanonymous committed 1 year ago
Qwen rotary embeddings should now match reference code. (#9437)
7cd2c4b comfyanonymous committed 1 year ago
Change the TextEncodeQwenImageEdit node to use logic closer to reference. (#9432)
bddd696 comfyanonymous committed 1 year ago
api_nodes: added release version of google's models (#9304)
d844d8b Alexander Piskun committed 1 year ago
fix(WAN-nodes): invalid nodeid for WanTrackToVideo (#9396)
bd2ab73 Alexander Piskun committed 1 year ago
fix(OpenAIGPTImage1): set correct MIME type for multipart uploads to OpenAI edits (#9348)
027c63f Alexander Piskun committed 1 year ago