wassname/kair_algorithms_draft
1
0
Fork 0
mirror of https://github.com/wassname/kair_algorithms_draft.git synced 2026-09-09 11:25:10 +08:00
Code Issues Packages Projects Releases Wiki Activity

Add installation, train documentation (#55)

Browse Source 
* Add docker/local train guidelines

* Add wandb log info

* Add environment

* Remove duplicated commands

* Add environment info

* Add period
This commit is contained in:
Whi Kwon
2019-05-21 12:00:29 +09:00
committed by Seungjae Ryan Lee
parent c1f348775e
commit 5ff59a280e
1 changed files with 65 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files
README.md
+65
View File
@@ -47,6 +47,71 @@ cd scripts
pip install -r requirements.txt
```
To train [LunarLanderContinuous-v2](https://gym.openai.com/envs/LunarLanderContinuous-v2/), install [OpenAI Gym](https://github.com/openai/gym) environment.
To train [Reacher-v1](https://gym.openai.com/evaluations/eval_6kY02G5DSpekMBmAeRLVyA/), install [MuJoCo](https://github.com/kairproject/kair_algorithms_draft/wiki/MuJoCo-Setup) environment.
## Environment
The code is developed using python 2.7, ROS kinetic on Ubuntu 16.04. NVIDIA GPU is needed. The code is developed and tested using 1 NVIDIA GeForce GTX 1080 Ti GPU card. Other platforms or GPU cards are not fully tested.
## How to Train
### Docker
To use docker, check the [installation guide](https://github.com/kairproject/kair_algorithms_draft/wiki/Docker).
#### Build Image
```
docker build -t kairproject/open_manipulator:0.1
```
or
```
docker pull kairproject/open_manipulator:0.1
```
#### OpenManipulator
```
docker run -v [path_to_kair_algorithms_draft]/save:/root/catkin_ws/src/kair_algorithms_draft/scripts/save --runtime=nvidia [image_id] openmanipulator [algo]
```
#### LunarLanderContinuous-v2
```
docker run -v [path_to_kair_algorithms_draft]/save:/root/catkin_ws/src/kair_algorithms_draft/scripts/save --runtime=nvidia [image_id] lunarlander [algo]
```
#### Reacher-v1
```
docker run -v [path_to_kair_algorithms_draft]/save:/root/catkin_ws/src/kair_algorithms_draft/scripts/save --runtime=nvidia [image_id] reacher [algo]
```
### Local
Our training [wandb](https://www.wandb.com/) log can be found in https://app.wandb.ai/kairproject/kair_algorithms_draft-scripts.
```
cd scripts
wandb login
```
#### OpenManipulator
```
python run_open_manipulator_reacher_v0.py --algo [algo] --off-render --log
```
#### LunarLanderContinuous-v2
```
python run_lunarlander_continuous.py --algo [algo] --off-render --log
```
#### Reacher-v1
```
python run_reacher_v1.py --algo [algo] --off-render --log
```
## How to Cite