Reparameterized Policy Learning

Project Page | Paper

This is the official implementation in PyTorch of paper:

Reparameterized Policy Learning for Multimodal Trajectory Optimization

Zhiao Huang, Litian Liang, Zhan Ling, Xuanlin Li, Chuang Gan, Hao Su

ICML 2023 (Oral Presentation)

use the below command for running sparse and dense reward experiments

cd run
python3 mbrpg.py --env_name EEArm --exp rpgcv2 --seed 0
python3 mbrpg.py --env_name AntPushDense --exp dense --seed 0