How to finetune flownet while training DFF model?

open-mmlab / mmtracking

OpenMMLab Video Perception Toolbox. It supports Video Object Detection (VID), Multiple Object Tracking (MOT), Single Object Tracking (SOT), Video Instance Segmentation (VIS) with a unified framework.

https://mmtracking.readthedocs.io/en/latest/

Apache License 2.0

3.53k stars 591 forks source link

How to finetune flownet while training DFF model? #659

Closed yan811 closed 2 years ago

yan811 commented 2 years ago

I have set 'init_cfg' to your pretrained flownet_simple.pth file. What should I do to upgrade the weights of flownet while training DFF model?

dyhBUPT commented 2 years ago

Hi, the flownet is part of DFF model, so when you train DFF, the weights of flownet is also trained.

yan811 commented 2 years ago

I got it. Thanks.