Reconsider TorchData, DataLoader2, etc

[albertz]

There were multiple reasons why we adopted TorchData:

• It seemed to be the future for datasets in PyTorch, thus deprecating the old way.
• The design looked overall better, specifically better suited for big datasets, iterable datasets.

However, it also had some downsides:

• It was still in beta. I did not really encounter any errors, but docs etc could all be improved more.
• The internal design was a bit strange to me. Specifically that it heavily relied on deepcopy.

And now, it seems as if they are also not happy with the overall design, and development has been halted.

⚠️ As of July 2023, we have paused active development on TorchData and have paused new releases. We have learnt a lot from building it and hearing from users, but also believe we need to re-evaluate the technical design and approach given how much the industry has changed since we began the project. During the rest of 2023 we will be re-evaluating our plans in this space. Please reach out if you suggestions or comments (please use pytorch/data#1196 for feedback).