search

RQ-Wu / LAMP

Official implement code of LAMP: Learn a Motion Pattern by Few-Shot Tuning a Text-to-Image Diffusion Model (Few-shot-based text-to-video diffusion)
https://rq-wu.github.io/projects/LAMP/index.html
Other
236 stars 10 forks source link

does each model only contain an individual motion or can multiple types of motion be trained inside a single model? #2

Closed TruthSearcher closed 8 months ago

TruthSearcher commented 8 months ago

does each model only contain an individual motion or can you train multiple types of motion inside a single model like motiondirector ?

RQ-Wu commented 8 months ago

Each model contains one motion pattern, such as the "horse run" model can generate a horse running with diverse poses, scenes, and styles. Actually, Motiondirector did the same thing as our work (It can not train multiple motions inside a single model, as I know). We are co-current works (around a 5-day gap for paper release). The difference is that Motiondirector tunes a T2V-LDM for a better specific-motion generation, while we tune a T2I-LDM to make it transfer to the video domain with a specific motion pattern.