There are a few resrouces that might come in handy in a classroom, particularly for team projects and capstone courses. None of the authors receive money when you use these. We just think they might be helpful.
Onepanel offers free CPU and GPU cloud compute to academic teams and classes upon request. Their platform is built around collaboration and includes TensorFlow, PyTorch, fast.ai, GPU optimized annotation tool, and can run custom code as well. Visit the website for a tour of what they offer. Instructors can contact info@onepanel.io to get a tour.
MatrixDS is a cloud-based workbench for data science. It offers free instances of hosted data science tools like Jupyterlab and RStudio. Also, other tools are available which support the data science project life cycle; including databases, Shiny, and web servers. The workbench supports a broad range of data science projects from introductory topics to advanced deep learning applications. Custom docker containers can also be launched directly from docker hub. All tools are available in a community project setting where students and instructors can share, collaborate, or present their work.