Description
Computer-vision datasets, transforms, and pretrained model utilities can be used from PyTorch with ROCm support. Researchers and ML engineers can train or evaluate vision pipelines on AMD GPU stacks. It processes image data that may contain private scenes or people.