DNN for Depth Completion in an IoT environment using Github actions to deploy to Azure ML.
- Folders starting with . (.cloud, .config, .github) -> Either the code of this folders is autogenerated or they contain some configurations (for example for the github action that deploys the training to Azure ML).
- Azure_scripts -> Contains all the logic related to Azure, with scripts to create/connect to computing devices/environments on Azure ML and run training experiments.
- Depth_inpainting -> Contains all the logic implementation for the DNN model responsible of complete the depth scene with help of LIDAR's incomplete depth map
- Edge -> Holds the logic involved in the Edge devices (connect to Azure, decompress the model and run inference).