Added Person Part Segmentation Using CDCL (Cross Domain Complementary Learning) Added Pose Estimation Using CDCL Minor Bug fixes Updated Documentation More Stable
1.4
Added Support of RetinaNet Training with (resnet50, resnet101, resnet152). Made validation data optional for Training RetinaNet and YOLO Models. Added DeepLabv3+ for Semantic Segmentation Added Dynamic Learning rate Updated Documentation More Efficient and Stable