dependencies.txt list the default settings of Deep Learning AMI (Ubuntu 18.04) Version 44.0, pytorch_latest_p37 env
which are compatible with the project (for all experiments except BookEmbedding which requires a different environment configuration, provided in BookEmbeddings/book_embedding_requirements.txt). 

Additionally, Weights and Biases (wandb==0.9.4) is required.

For instructions to run the LSTM experiment, see the README in the experiments/lstm_recall folder.

For (pathfinder/world of numbers/game of life) experiments:
Change directory to experiments and run (pathfinder/mnist/conway)_exp_runner.py to train the model, then change the load_path in (pathfinder/mnist/conway)_exploration.py to point to the relevant saved model checkpoint and run to test metrics on the result. Results will be logged to Weights and Biases. Training parameters can be changed by modifying the (pathfinder/mnist/conway)_exp_runner.py files.

