TorchCV is a framework built on PyTorch that implements state-of-the-art deep learning methods for common computer vision problems. It provides reimplementations of papers across image classification, semantic segmentation, object detection, and pose estimation.