svip-lab / impersonator

PyTorch implementation of our ICCV 2019 paper: Liquid Warping GAN: A Unified Framework for Human Motion Imitation, Appearance Transfer and Novel View Synthesis
https://svip-lab.github.io/project/impersonator
Other
1.73k stars 317 forks source link

Bad results with my own data #56

Closed syp1997 closed 4 years ago

syp1997 commented 4 years ago

I used my own source image and dance videos, but get bad results, how can i solve it?

StevenLiuWen commented 4 years ago

@syp1997 , Hi, since the model has been pre-trained on the iPER dataset, and the iPER dataset has the shortcomings in the diversity, especially in the background, you might need to firstly train the model on you own dataset, and see https://github.com/svip-lab/impersonator/issues/28 more details.