Disclaimer
If you find my models interesting Please Like and Follow on Civitai and other platforms (see link in my bio) it helps, GPU is not cheap nowadays!
Wan2.2 Image to Video sigma face LORA
To use it successfully please use both High and Low noise versions at the same time at inference (might work with only high noise, although I haven't tested it)
Trigger phrase: ... doing sigma face expression
Recommended LORA strength 1-1.4 for both High noise and low noise. I'm generating with 1.0 on both, or 1.2/1.4 on high and 1.0 on low noise.
Lora trained only on 6 real videos (using LORA for different style images like anime/claymation might not yield satisfactory results)
I just started creating for the community, after working for various companies the time has finally come to devote some time to open source. ;)
Have fun folks!
Description
Low noise version
FAQ
Comments (8)
Impressive, very nice. Let's see Paul Allen's LoRA.
If only we could use real people identities on Civitai I would train him just for the meme!
Look at the consistency of the faces, the flexibility of subjects, oh my god... it's even has good tagging.
Nice, btw can you do that jawline thing next haha
wtf i love it 😂😂
what the fuck does sigma mean?
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.