TPU support

Hi, first of all thanks for this amazing work 👍.
Is it possible to run the `train_network.py` script on a TPU?

I actually tried but it's not working, I even remove the `.to("cuda")` now I'm not seeing any error but the training stuck after it finishes caching the latens.

How can I make this work on TPU?

PS: I'm a software engineer, but I'm new to the machine learning world.

Start with train_network.py and reproduce the reported TPU run, checking the point where training stalls after latent caching. Trace the existing device handling around the reported .to("cuda") calls and determine the TPU execution requirements. Done means train_network.py completes training on a TPU with a clear invocation and no post-caching stall.

Written by the indexing model from the issue text.