This project is meant to explore the possibility of training and RNN to classify trajectories generated by the Standard map dynamical system.
-
requirements:
- python==3.10
- pytorch lightning
- pyyaml
- tensorboard
- pandas
- matplotlib
-
activate conda environment with source ~/.bashrc conda activate rnn_env
-
install packages in venv environment with conda install package_name
-
main functionalities:
- to sample parameters from grid: python gridsearch.py
- to run a single training session: python trainer.py
- to run a single parameter update: python update.py
- to run hyperparameter optimizaton: bash main_c.slurm