fid is bigger than 80, unable to get the result you provided using checkpoints/pretrained_models/humanml3d/vqvae/net_last.pth

ZcyMonkey / AttT2M

Code of ICCV 2023 paper: "AttT2M: Text-Driven Human Motion Generation with Multi-Perspective Attention Mechanism"

https://arxiv.org/abs/2309.00796

Apache License 2.0

42 stars 3 forks source link

fid is bigger than 80, unable to get the result you provided using checkpoints/pretrained_models/humanml3d/vqvae/net_last.pth #7

Closed ShungJhon closed 3 months ago

ShungJhon commented 3 months ago

Unable to reproduce the results obtained in the paper, is the model you provided correct? https://drive.usercontent.google.com/download?id=1UguYU5YSmhOanzPp79AsLxTeyIjPHEZA&export=download&authuser=0