TransDynaMo

A project to model dynamical systems using Transformers

Data Generation and Download

Data for the exepeirments can be generated by running the data_generation files. Alternatively, data can be downloaded from the following link - https://uofi.box.com/s/zq82d749jx4qrsr2042h3cv966hkyruh

The main results for each environment can be found in their respective notebooks. The respective plots are saved in the /figures folder.

Name		Name	Last commit message	Last commit date
Latest commit History 22 Commits
.ipynb_checkpoints		.ipynb_checkpoints
data		data
figures		figures
.DS_Store		.DS_Store
.gitmodules		.gitmodules
README.md		README.md
TranDynaMo-Acrobot.ipynb		TranDynaMo-Acrobot.ipynb
TranDynaMo-HalfCheetah.ipynb		TranDynaMo-HalfCheetah.ipynb
TranDynaMo-Pendulum-2.ipynb		TranDynaMo-Pendulum-2.ipynb
TranDynaMo-Pendulum-Bert.ipynb		TranDynaMo-Pendulum-Bert.ipynb
TranDynaMo-Pendulum-Normalized-RNN.ipynb		TranDynaMo-Pendulum-Normalized-RNN.ipynb
TranDynaMo-Pendulum-Normalized.ipynb		TranDynaMo-Pendulum-Normalized.ipynb
TranDynaMo-Pendulum-diff-state.ipynb		TranDynaMo-Pendulum-diff-state.ipynb
TranDynaMo-Pendulum.ipynb		TranDynaMo-Pendulum.ipynb
TransDynaMo_report.pdf		TransDynaMo_report.pdf
acrobot_data_genration.py		acrobot_data_genration.py
configuration_bert.py		configuration_bert.py
decision_transformer.py		decision_transformer.py
half_cheetah_data_generation.py		half_cheetah_data_generation.py
mlp_bc.py		mlp_bc.py
model.py		model.py
modeling_bert.py		modeling_bert.py
pendulum_data_genration.py		pendulum_data_genration.py
trajectory_gpt2.py		trajectory_gpt2.py