Comfyui_turbodiffusion

This significantly expanded workflow is based on the standard workflow of the new model https://github.com/thu-ml/TurboDiffusion (the adapted model and node for ComfyUI are here: https://github.com/anveshane/Comfyui_turbodiffusion).

Using the new model significantly speeds up generation (according to the statement on the model's main page, by several times or orders of magnitude, especially on NVIDIA RTX 5090 series graphics cards).

From my experience, on AMD Ryzen 7 2400 with 4060TI 16GB and 64GB of RAM, generating the video shown on standard WAN22 models took up to 6-8 hours.

On TurboDiffusion, the same video was generated in less than 2 hours using the included workflow.

The downside is that the model does not yet support any loras.

You can use this workflow as is, however you like.

If you have any questions, please contact the model's developers on GitHub.