search

yuantn / MI-AOD

Code for Multiple Instance Active Learning for Object Detection, CVPR 2021
https://openaccess.thecvf.com/content/CVPR2021/papers/Yuan_Multiple_Instance_Active_Learning_for_Object_Detection_CVPR_2021_paper.pdf
Apache License 2.0
333 stars 43 forks source link

How to resume from file #65

Closed yozoral closed 2 years ago

yozoral commented 2 years ago

Hi thank you for this great project. Can you give some instructions on how to resume the training process at specific cycle from model file if it paused in the middle? I changed the config 'resume_from' and it got optimizer loading issue, ValueError: loaded state dict contains a parameter group that doesn't match the size of optimizer's group Even if I commented the optimizer loading, it will stuck and do no training. Thank you so much if you can give some instructions on this.

yuantn commented 2 years ago

Considering that we have modified the detector architecture and training pipeline, we suggest you resume from checkpoints by modifing Line 107-114 and Line 133-135 in tools/train.py, as mentioned in part pth file in section Results of README.md.