← all repositories

TianhongDai/hindsight-experience-replay

PyTorch implementation of the Hindsight Experience Replay reinforcement learning algorithm for training agents on robotic manipulation tasks.

451 stars Python Domain AppsML Frameworks
hindsight-experience-replay
Velocity · 7d
+0.2
★ / day
Trend
steady
star history

This repository implements the HER algorithm in PyTorch to train DDPG agents on OpenAI Gym’s Fetch robotic manipulation environments. The implementation supports multi-process training via MPI and includes pre-trained models for FetchReach, FetchPush, FetchPickAndPlace, and FetchSlide tasks. It builds on the OpenAI Baselines codebase and targets robotics research applications.

heatdrop uses Google Analytics to see which pages get read — nothing else. Your call. How we handle data.