https://github.com/camigord/distributed_ddpg
Parallel implementation of DDPG
https://github.com/camigord/distributed_ddpg
ddpg deep-reinforcement-learning
Last synced: 5 months ago
JSON representation
Parallel implementation of DDPG
- Host: GitHub
- URL: https://github.com/camigord/distributed_ddpg
- Owner: camigord
- Created: 2017-09-05T11:03:55.000Z (about 8 years ago)
- Default Branch: master
- Last Pushed: 2017-09-06T14:58:20.000Z (about 8 years ago)
- Last Synced: 2025-04-05T06:11:28.971Z (6 months ago)
- Topics: ddpg, deep-reinforcement-learning
- Language: Python
- Size: 6.39 MB
- Stars: 14
- Watchers: 3
- Forks: 3
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Distributed-DDPG
## Overview
The purpose of this repository is to implement the Deep Deterministic Policy Gradient algorithm or [DDPG](https://arxiv.org/abs/1509.02971) in a distributed fashion as proposed [here](https://arxiv.org/abs/1704.03073).
I will start by evaluating the performance of DDPG in simple cases and then comparing this performance when distributing the training process among several "workers".
## MountainCarContinuous-v0 (OpenAI)
I evaluated the performance of the standard DDPG approach on the [MountainCarContinuous](https://github.com/openai/gym/wiki/MountainCarContinuous-v0) task. The figure below shows the training curves until the problem is considered solved.
The provided results were obtained by running a single worker. To replicate the results run the following commands in two different consoles:
```
# Parameter server
python ddpg.py --job_name="ps" --task_index=0
```
```
# First worker
python ddpg.py --job_name="worker" --task_index=0
```
To visualize the training process using TensorBoard:
```
# TensorBoard
tensorboard --logdir=results/tboard_ddpg/
```