Freaky SVD-Freaky 6+SVD output as alternative to Mochi.  (low vram workflow)

- You can right click the groups to turn them on bypass or set to always.

This workflow is a low (8 gb) vram setup for exporting an image and/or image+related video.

Includes Stable Diffusion 3.5 Medium, Aura Flow, Pixart Sigma, SDXL, and Hunyuan for single image generation.

Includes Moochi and SVD sdxl for exporting video/moving image.

I did try to include Hunyuan video originally but am having issues with my pc recognizing the Hunyuan diffusion model name for some reason.  SVD is an older versiou of video output that does work for low vram machines. However, it does still take a while to produce images if you are using 8gb vram. It probably will work but anything more than a still video will always take longer to produce an image, no matter how much vram you have...  I did not include flux since I am having issues with flux not working in comfy due to low vram causing issue where tensor gets offloaded to cpu causing tensor divided between multiple gpu errors... If I need to generate flux level images in 8gb vram machines than using Forge is a better way, with the beta option and flux realistic option chosen...  I wish there was a way to force newer models or sd3.5 large to work in lower vram environments but they are not built for that and memory will max after so long, even with tiled devae options turned on.  

Have fun. This workflow mainly exists to show folks with low vram that things like video is possible on your machine, you just have to think a little outside of the box on how to do it.