search

ToughStoneX / Self-Supervised-MVS

Pytorch codes for "Self-supervised Multi-view Stereo via Effective Co-Segmentation and Data-Augmentation"
152 stars 15 forks source link

Model training interruptions #9

Open Run-1123 opened 3 years ago

Run-1123 commented 3 years ago

How to continue training a trained model after it has been interrupted in jdacs-ms.

ToughStoneX commented 2 years ago

The training script has a --resume option. You can enable it in your command and feed the path of saved checkpoint file in option --loadckpt.