5. Run on TPU !

Github Link : https://github.com/satyajitghana/ProjektDepth/blob/master/notebooks/09_DepthModel_Experiments_TPU.ipynb Colab Link : https://colab.research.google.com/github/satyajitghana/ProjektDepth/blob/master/notebooks/09_DepthModel_Experiments_TPU.ipynb

Refer to this article: https://medium.com/@satyajitghana7/speed-up-your-model-training-w-tpu-on-google-colab-c55ac0f634d9

Running on TPU almost halfed the training time, but at the time of writing this documentation, the code doesn’t work, or is buggy since the PyTorch-XLA package is being updated frequently and there has been a lot of changes that’s being introduced, which is the reason i decided to stay away from it when actually training my model. But its good for small trains.