- Notifications
You must be signed in to change notification settings - Fork0
[AAMAS 2024] Guided Exploration in Reinforcement Learning via Monte Carlo Critic Optimization
License
NotificationsYou must be signed in to change notification settings
schatty/MOCCO
Folders and files
| Name | Name | Last commit message | Last commit date | |
|---|---|---|---|---|
Repository files navigation
The source code for the paper Guided Exploration in Reinforcement Learning via Monte Carla Critic OptimizationArxiv, presented at
The experiments were run withpython3.10 andmujoco 2.3.7, install full env via
pip install -r requirements.txtRun single training
python train.py --env point_mass-easy --algo MOCCO --device cuda:0 --seed 0Run on many seeds
For running an algorithm on many many seeds to reproduce paper results, specify needed algorithm as a script file (in/scripts folder) and set needed env as a first argument:
bash scripts/mocco.sh point_mass-easy cuda:0