Yu-Huan Wu
Yu-Huan Wu
Hi, the request txt file is as below (expected 800 lines): ```` DUT-RGBD/1609.jpg DUT-RGBD/1609_gt.png DUT-RGBD/1565.jpg DUT-RGBD/1565_gt.png DUT-RGBD/0768.jpg DUT-RGBD/0768_gt.png DUT-RGBD/0350.jpg DUT-RGBD/0350_gt.png DUT-RGBD/0680.jpg DUT-RGBD/0680_gt.png DUT-RGBD/0817.jpg DUT-RGBD/0817_gt.png DUT-RGBD/0418.jpg DUT-RGBD/0418_gt.png DUT-RGBD/1056.jpg DUT-RGBD/1056_gt.png DUT-RGBD/0489.jpg DUT-RGBD/0489_gt.png...
Thanks a lot for your enquiry. `train_single.py` is to train the segmentation branch. During the joint training stage, you could load the trained weights of segmentation branch (pretrained on segmentation...
We have no plans to release the entire code of training the classification branch currently. However, you can reimplement the training process as all details (process, hyper-params) have been explained...