Download the .pth models from https://github.com/madebyollin/taesd and stick them in ComfyUI\ComfyUI\models\taesd
Edit ComfyUI\run_nvidia_gpu.bat, add --preview-method auto
Example:
.\python_embeded\python.exe -s ComfyUI\main.py --windows-standalone-build --use-pytorch-cross-attention --preview-method auto
Slightly slower image generation, went from 1.79 it/s to 1.53 it/s (1280x768 RTX3070)
Download the .pth models from https://github.com/madebyollin/taesd and stick them in ComfyUI\ComfyUI\models\taesd Edit ComfyUI\run_nvidia_gpu.bat, add --preview-method auto Example: .\python_embeded\python.exe -s ComfyUI\main.py --windows-standalone-build --use-pytorch-cross-attention --preview-method auto
Slightly slower image generation, went from 1.79 it/s to 1.53 it/s (1280x768 RTX3070)