Description
Computer-vision operations can be used inside PyTorch models with differentiable image-processing primitives. ML engineers and researchers use this library for training, augmentation, geometry, and vision experiments. Datasets, model outputs, and image content can contain private or biometric information.