td3_her_rlbench_reacher

KO | EN

gitlite — search

#python #hacktoberfest #golang #rust #linux #macos #windows #cli #kubernetes #vue #pytorch #mcp #ios #open-source

td3_her_rlbench_reacher

(★ 12)

A implementation for soving reach target task based on TD3 with HER using PaddlePaddle.

td3_her_rlbench_reacher 최신버젼 다운로드

최종 버전 다운로드 (.zip)

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# td3_her_rlbench_reacher
Author: CAO RUI

A implementation for solving reach target task based on Twin Delayed DDPG(TD3) with Hindsight Experience Replay(HER) using PaddlePaddle.

![image](https://github.com/63445538/td3_her_rlbench_reacher/blob/master/RLBench/records/solved/video_7.gif)

# Prerequisites

Python: Python 3.6+

[PaddlePaddle](https://github.com/PaddlePaddle/Paddle) : Deep learning framework

[PARL](https://github.com/PaddlePaddle/PARL) : Reinforcement learning toolbox based on PaddlePaddle

[gym](https://github.com/openai/gym) : Universal environment builder for RL tasks

[RLBench](https://github.com/stepjam/RLBench): RL tasks extension for robotics researches.

# Install
First, create a virtual environment by ```virtualenv```, in it, install PaddlePaddle, gym and PARL by

```pip install requirements.txt```

Then install RLBench via [RLBench](https://github.com/stepjam/RLBench).

# Train

```python rlbench_reach_td3_train.py```

# Evaluate

```python rlbench_reach_td3_eval.py```

# Result
4 stages (initial model, after 40000-episode trained model, after 80000-episode trained model, fianl model) of training model are uploaded and the corresponded render results are recorded in the folder ```records```.

The success rate as shown in the following figure, here every epoch equals to 200 training episodes.
![image](https://github.com/63445538/td3_her_rlbench_reacher/blob/master/RLBench/records/Success%20rate.png)

td3_her_rlbench_reacher

td3_her_rlbench_reacher 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

td3_her_rlbench_reacher

td3_her_rlbench_reacher 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

Download README