https://github.com/shuaibinli/rl_carla
Train auto_car in CARLA simulator with RL algorithms(SAC).
https://github.com/shuaibinli/rl_carla
carla rl
Last synced: 4 months ago
JSON representation
Train auto_car in CARLA simulator with RL algorithms(SAC).
- Host: GitHub
- URL: https://github.com/shuaibinli/rl_carla
- Owner: ShuaibinLi
- License: apache-2.0
- Created: 2021-02-01T08:09:54.000Z (over 4 years ago)
- Default Branch: main
- Last Pushed: 2023-04-26T04:00:03.000Z (over 2 years ago)
- Last Synced: 2025-04-13T03:39:49.125Z (6 months ago)
- Topics: carla, rl
- Language: Python
- Homepage:
- Size: 6.8 MB
- Stars: 104
- Watchers: 2
- Forks: 12
- Open Issues: 6
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# RL_CARLA
## SAC in Carla simulator
Based on [PARL](https://github.com/PaddlePaddle/PARL) and Torch/Paddle(Baidu deep learning framework),
a parallel version of SAC was implemented and achieved high performance in the CARLA environment.
> Paper: SAC in [Soft Actor-Critic: Off-Policy Maximum Entropy Deep Reinforcement Learning with a Stochastic Actor](https://arxiv.org/abs/1801.01290)
### Carla simulator introduction
Please see [Carla simulator](https://github.com/carla-simulator/carla/releases/tag/0.9.6) to know more about Carla simulator.
### Benchmark result
+ Result was evaluated with mode `Lane`
## How to use
+ System: Ubuntu 16.04
### Dependencies:
+ Simulator: [CARLA](https://github.com/carla-simulator/carla/releases/tag/0.9.6)
+ RL env: gym_carla
### Installation
1. Create conda environment
```env
$ conda create -n rl_carla python=3.6
$ conda activate rl_carla
```
2. Download [CARLA_0.9.6](https://github.com/carla-simulator/carla/releases/tag/0.9.6),
extract it to some folder, and add CARLA to `PYTHONPATH` environment variable
```
$ export PYTHONPATH="SOMEFOLDER/CARLA_0.9.6/PythonAPI/carla/dist/carla-0.9.6-py3.5-linux-x86_64.egg:$PYTHONPATH"
```
3. Clone repository from [this repo](https://github.com/ShuaibinLi/RL_CARLA.git)
```clone
$ git clone https://github.com/ShuaibinLi/RL_CARLA.git
```
4. Install the packages
```
## install requirements，
## Install paddle or torch wrt which base you are using(paddle or torch)
## Make sure only one deep-learning framework exists in your conda_env during traing
$ pip install -r requirements.txt
## install gym env of carla
$ cd gym_carla
$ pip install -e .
```
or you can install the package that you need by `pip/conda install [package name]`
#### Start Training
1. Open another(new) terminal, enter the CARLA root folder and launch CARLA service.
There are two modes to start the CARLA server:
(1) non-display mode
```start env
$ DISPLAY= ./CarlaUE4.sh -opengl -carla-port=2021
```
(2) display mode
```start_env
$ ./CarlaUE4.sh -windowed -carla-port=2021
```
+ Start three CARLA services (ports: 2021,2023,2025) for data collecting and training,
one service (port: 2027) for evaluating.
2. For parallel training, we can execute the following [xparl](https://parl.readthedocs.io/en/stable/parallel_training/setup.html) command to start a PARL cluster：
```Parallelization
$ xparl start --port 8080
```
check xparl cluster status by `xparl status`
3. Start training
```train
$ python train.py --xparl_addr localhost:8080
```
#### Evaluate trained agent
Open another(new) terminal, enter the CARLA root folder and launch CARLA service with display mode.
```start_test
$ ./CarlaUE4.sh -windowed -carla-port=2029
```
Restore saved model to see performance.
```
$ python evaluate.py --restore_model model.ckpt