I've been testing between the latest version of InvokeAI (https://github.com/invoke-ai/InvokeAI) and ComfyUI (https://github.com/comfyanonymous/ComfyUI) and interestingly, it appears that ComfyUI is not as hard on the system as InvokeAI.

I tried the same model, LoRA, prompt, and batch size, and while the machine stayed totally fine during the ComfyUI generation (even after multiple batches), with InvokeAI, my MBP fan started going after just one set of images.

I do realize that InvokeAI is in a state of flux as they move to version 3.0, but the difference in usability is great enough that I think I’ll stick with ComfyUI for the time being …

#ImageGeneration #Python #macOS #ComfyUI #InvokeAI