GitHub

Reinforcement Learning from Human Feedback for the Construction of Spanning Structures

Code from the paper "Reinforcement learning for scaffold free construction of spanning structures" adapted for the Semester Project "Reinforcement Learning from Human Feedback for the Construction of Spanning Structures," taking inspiration from the imitation library.

The top-level file that runs the RLHF algorithm is called rlhf_main.py. The hyperparameters of the run are all set at the top of this file.

Name		Name	Last commit message	Last commit date
Latest commit History 53 Commits
wandb		wandb
.gitignore		.gitignore
Agents.py		Agents.py
LICENSE		LICENSE
README.md		README.md
discrete_blocks.py		discrete_blocks.py
discrete_graphics.py		discrete_graphics.py
discrete_simulator.py		discrete_simulator.py
gen_plot_results.py		gen_plot_results.py
generate_plots.py		generate_plots.py
generate_structures_from_file.py		generate_structures_from_file.py
geometric_internal_model.py		geometric_internal_model.py
internal_models.py		internal_models.py
physics_scipy.py		physics_scipy.py
pyg_graphics.py		pyg_graphics.py
pyg_single_agent.py		pyg_single_agent.py
relative_single_agent.py		relative_single_agent.py
requirements.txt		requirements.txt
rlhf_fragmenter.py		rlhf_fragmenter.py
rlhf_main.ipynb		rlhf_main.ipynb
rlhf_main.py		rlhf_main.py
rlhf_pair_generator.py		rlhf_pair_generator.py
rlhf_preference_comparisons.py		rlhf_preference_comparisons.py
rlhf_preference_dataset.py		rlhf_preference_dataset.py
rlhf_preference_gatherer.py		rlhf_preference_gatherer.py
rlhf_preference_model.py		rlhf_preference_model.py
rlhf_reward_comparison.py		rlhf_reward_comparison.py
rlhf_reward_fn.py		rlhf_reward_fn.py
rlhf_reward_model.py		rlhf_reward_model.py
rlhf_reward_net.py		rlhf_reward_net.py
rlhf_reward_trainer.py		rlhf_reward_trainer.py
rlhf_trajectory_generator.py		rlhf_trajectory_generator.py
single_agent_gym.py		single_agent_gym.py