Skip to content

drdh/Synergy-RL

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

1 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Low-Rank Modular Reinforcement Learning via Muscle Synergy

This repository is the official implementation of Low-Rank Modular Reinforcement Learning via Muscle Synergy.

Setup

pip install -r requirements.txt

Training

python main.py 
--custom_xml environments/humanoids 
--actor_type transformer 
--critic_type transformer 
--grad_clipping_value 0.1 
--attention_layers 3 
--attention_heads 2 
--lr 0.0001 
--transformer_norm 1 
--attention_hidden_size 256 
--condition_decoder 1
--label SOLAR
--save_buffer 1
--save_freq 100000
--start_timesteps 2000
--enable_synergy_obs 1\
--enable_synergy_act 1\
--synergy_action_dim 1\
--d_model 1\

We were using Sacred and WandB for experiment management. You can disable WandB by setting --enable_wandb 0.

To train other environments, change the --custom_xml to hoppers, walkers, walker_humanoids_hopper.

Acknowledgement

About

This repository is the official implementation of Low-Rank Modular Reinforcement Learning via Muscle Synergy.

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages