Different Model (RNN + CNN, LSTM, etc.) with Tim

[Ben-Drucker]

Status with Tim as of 4/3/2023 and things to try:

• [ ] Add more layers to the kinase CNN so it is not pooled so drastically
• [ ] Combine CNN with RNN (LSTM)
  • [X] Create "toy" RNN GC to gain familiarity and ensure I am doing RNNs correctly
  • [ ] Clean ^ up
  • [ ] Share "toy" RNN GC with Tim for review
• [ ] Increase the number of attention layers
• [ ] Change the attention activation function to softmax