DeepView Dataset Library

This project contains python and rust binding for reading and exporting datasets. The library also includes Dataset iterators for both GPU (Tensorflow/Pytorch) and iterators for running validation on embedded devices where ML framewors are not available.

The labrary is distributed by Au-Zone Technologies with the aim of outperforming data processing pipelines while running on embedded devices or even when training a model. In order to generalize as much as possible, several tasks were included on it:

• Object Detection
• Distance Estimation
• Instance Segmentation and Semantic Segmentation
• Radar and Camera Fusion
• PCL classification

Data augmentation is another key feature included into this library. Augmenting data is higly recommended when training ML models. The larger the model, the more data needs.

Python Binding

The python binding can be installed by calling:

pip install --upgrade deepview-datasets

Rust Binding

Coming Soon!