Description
Computer-vision datasets, transforms, and model utilities can be linked through a ROCm-enabled C++ TorchVision library. ML developers can use vision components on AMD GPU stacks from software that depends on the C++ library. It processes image data that may contain private scenes or people.