Type | Workflows |
Stats | 3,265 |
Reviews | (211) |
Published | Jan 25, 2025 |
Base Model | |
Hash | AutoV2 68AE9119AE |
HUNYUAN | Img 2 Vid LeapFusion
Requirements: 1) Kijai's nodes 2) LeapFusion Lora v2 (544p) or v1 (320p)
YOU MUST UPDATE KIJAI NODES IF YOU HAVE THEM ALREADY INSTALLED
In short: it uses a special LORA to do the trick.
Actually a hit or miss, but when it works is the best I2V you can find around.
Workflow originally provided by Kijai (who implemented support for LeapFusion)
I modified a bit. You can compare to original settings here
It works combined with avaible loras around. Prompting helps a lot but works even without.
Raise resolution for more consistence and similarity with input image.
*you may want to change steps on your needs. I used few steps for testing.
Bonus TIPS:
Here an article with all tips and trick i'm writing as i test this model since December:
https://civitai.com/articles/9584
you will get a lot of precious quality of life tips to build and improving your hunyuan experience.
no need to buzz me, ty💗 ..feedbacks are much more appreciated.
Here a Discord link if you want to chat with me or need support
⚠️choose ATTENTION_MODE that works for you⚠️
NOTES:
The current situation is that the Fast model, when used alone, may cause artifacts. However, when used in combination with LoRAs, it works really well.
So my advice is:
Use the Fast model (which allows for a low number of steps, as low as 6 to 9) only if you are also using LoRAs together.
In all cases where the LoRA is not compatible or the Fast model causes artifacts, it's recommended to switch to the standard model and increase the number of steps to at least 15 or more. (25 to 30 steps is a good choice)
That said, not all LoRAs are compatible with Kijai's nodes, so it's up to you to test which ones work.
There's also a FAST LoRA, which can be loaded onto the standard model. Unfortunately, that one isn’t compatible with these nodes, and no one has bothered to convert it yet to make it work.
I don’t know how to convert it.. otherwise, I would have done it myself.
A lot of things would be much simpler by now 😒.