August 3, 2021
Like many image processing problems, deep learning has brought many effective solutions to the task of object detection. The TorchVision library is part of the PyTorch project, and it offers well-established and successful methods for object detection (as well as many other problems). This workshop will demonstrate the process of preparing your own image dataset and training it using TorchVision with one of Google Colab’s free-to-use GPUs. The workshop will be done online via Zoom. Some experience with Python is helpful, but no previous experience with PyTorch is needed. Google Colab requires a Google account (e.g. your umich account).