search

rocketappslab / BDMM

Official PyTorch implementation of BDMM: Bidirectionally Deformable Motion Modulation For Video-based Human Pose Transfer [ICCV 2023]
Apache License 2.0
55 stars 3 forks source link

about myself image test and run train.py have a error？ #1

Open zhanghongyong123456 opened 1 year ago

zhanghongyong123456 commented 1 year ago

when i run train.py ,i have a error: image i print self.pre_warp_feat_iter ,i find this image i find define : nxt_warp_feat_iter = [None, feat_warp64_state, feat_warp128_state],but None have no detach
image

zhanghongyong123456 commented 1 year ago

First of all thanks a lot for the code you posted, a few questions

  1. How to drive my own images instead of the Fashion Video Dataset dataset,
  2. about iper datasets trained model ？
  3. whether i need to retrain the model to drive my own images

i want to test my source image and driving video（reference video ）,like this demo ,how to do this on our project image

My test results: image image

my test video，The driving action is fine, but the generated portrait seems to be problematic

https://github.com/rocketappslab/BDMM/assets/48466610/c7704496-4a0a-48d9-8209-a472c89fcfee

9B8DY6 commented 7 months ago

Me, either! I also have the same problem!! self.pre_warp_feat_iter = [each.detach() for each in self.pre_warp_feat_iter] AttributeError: 'NoneType' object has no attribute 'detach'