search

facebookresearch / AGRoL

Code release for "Avatars Grow Legs Generating Smooth Human Motion from Sparse Tracking Inputs with Diffusion Model", CVPR 2023
Other
238 stars 24 forks source link

Question about retrain model #17

Closed Feiyu-Yao closed 1 year ago

Feiyu-Yao commented 1 year ago

Hi,

I have a question about the training process. I have retrained the DiffMLP model about 54 epochs. However, the performance is far below that of the pretrained model. Are there some strategies in the training process? Also, I found the mean and std achieved from AMASS are different from those provided. Would you mind giving some hints about this?