homework 7
So far, we were able to train a model to get 99% accuracy using only 0.1% of the dataet. Things to do:
- Train on a GPU somehow
- Finish the experimentation and intuition section, which is just playing around with stuff
- Finish Part 2, which is the variable length section. To do this, We only need to copy everything down from part 1 and make a small change to "convert_sentence_to_fixed_length" function and just.... not convert things to a fixed length.
Here's our current training curve:
