PyTorch sometimes works quirkily which makes it intuitively a bit different from the theory one learns. This is my attempt to tackle that, in seemingly different topics. However, I hope that, as one reads more, they will appreciate the nuances of the topics and choice of the torch functionality alongside it.
One can also think of this as my database for Human RAG during conversations with Tom, as a means to match his infinite test-time knowledge base.
Lastly, as always, Talk is cheap, show me the code.