Workflow to do 360P Videos with any photo (Low VRAM Machines)
My Discord https://discord.gg/Axs8MgFRkh
Description
FAQ
Comments (16)
I keep getting this Error, "HunyuanVideo.forward() missing 1 required positional argument: 'y'"
Great work! It works for me. Any possibility to add a LoRa Loader? (IDK how to do it 😅)
i think img2vid you can't add lora idk. i don't tried yet.
why is my UNET Loader (GGUF) architecture not recognized?
you need to install GGUF nodes. go to manager in comfyui and install GGUF nodes.
This is a major breakthrough! Currently i2v on Wan takes ages even on fast GPU. Is there a way to get his working in gradio? Could I just add the model file to the folder and try it?
IDK, i never tried.
@Fabin_2024 I think it's worth it. I see many Comfy users saying they still rather use Gradio. And once it gets put on Gradio, it's only a couple of steps away from the wider community. Keep up the great work!
Just had the same 'architecture not recognized' error about the Unet loader. I already have GGUF nodes installed, so I just placed a GGUF loader with the same Unet specified and swapped their connections. Works 100% now, I was warned of node conflicts when downloading missing nodes. Thanks for the workflow!
PS:- 8GB 3050 gpu, 4790k cpu, 16gb ddr3, 1647'seconds for first run, 1087'seconds for a second run
Not working for me. The anything everywhere node is not sending the data to the nodes.
takes about 2 hours on my GTX1080 8GB, am I doing something wrong or is it really just that big of a difference from the 8GB 3060 you have? (it really only takes you 10 minutes?)
Your lucky it's only taking 8hours.
no tensor cores on 1080