Antipodal Robotic Grasping using GR-ConvNet. IROS 2020.
Fixed bug introduced by adding quality in Grasp definition Configurable IoU threshold Configurable network input size Added trained models for Jacquard dataset
Bugfix in ResidualBlock Updated default params
Refactored models to reduce duplicate code Added trained models
Here are key updates in this release: Configurable GR-ConvNets Configurable IOU threshold in evaluation Updated train/val split method to RandomSampler based method Support for configurable optimizer Updated logging to single directory Save training logs and args Updated evaluate.py to support multiple networks
Implementation of the Generative Residual Convolutional Neural Network (GR-ConvNet) from the paper: Antipodal Robotic Grasping using Generative Residual Convolutional Neural Network
Supports: Model Training Model Evaluation Cornell and Jacquard Datasets Calibration Task Grasp Generator Task