Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/Shmuma/ptan
PyTorch Agent Net: reinforcement learning toolkit for pytorch
https://github.com/Shmuma/ptan
Last synced: 3 months ago
JSON representation
PyTorch Agent Net: reinforcement learning toolkit for pytorch
- Host: GitHub
- URL: https://github.com/Shmuma/ptan
- Owner: Shmuma
- License: mit
- Created: 2017-09-18T13:42:10.000Z (almost 7 years ago)
- Default Branch: master
- Last Pushed: 2024-03-29T18:15:03.000Z (3 months ago)
- Last Synced: 2024-03-30T14:52:55.997Z (3 months ago)
- Language: Python
- Size: 671 KB
- Stars: 520
- Watchers: 23
- Forks: 161
- Open Issues: 21
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Lists
- my-awesome-stars - Shmuma/ptan - PyTorch Agent Net: reinforcement learning toolkit for pytorch (Python)
- awesome-stars - ptan
README
# PTAN
PTAN stands for PyTorch AgentNet -- reimplementation of
[AgentNet](https://github.com/yandexdataschool/AgentNet) library for
[PyTorch](http://pytorch.org/)
This library was used in ["Deep Reinforcement Learning Hands-On"](https://www.packtpub.com/data/deep-reinforcement-learning-hands-on-second-edition) book, here you can find [sample sources](https://github.com/PacktPublishing/Deep-Reinforcement-Learning-Hands-On).
## Code branches
The repository is maintained to keep dependency versions up-to-date.
This requires efforts and time to test all the examples on new versions, so, be patient.
The logic is following: there are several branches of the code, corresponding to
major pytorch version code was tested. Due to incompatibilities in pytorch and other components,
**code in the printed book might differ from the code in the repo**.
At the moment, there are the following branches available:
* `master`: contains the code with the latest pytorch which was tested. At the moment, it is pytorch 1.7.
* `torch-1.3-book-ed2`: code printed in the book (second edition) with minor bug fixes. Uses pytorch=1.3 which
is available only on conda repos.
* `torch-1.7`: pytorch 1.7. Merged with master.
All the branches uses python 3.7, more recent versions weren't tested.
## Installation
From sources:
```bash
python setup.py install
```
From pypi:
```bash
pip install ptan
```
From github:
```bash
pip install pip install git+https://github.com/Shmuma/ptan.git
```
## Requirements
* [PyTorch](http://pytorch.org/): version 1.1.0 is required
* [PyTorch Ignite](https://pytorch.org/ignite/): provides extra bindings for ignite
* [OpenAI Gym](https://gym.openai.com/): ```pip install gym gym[atari]```
* [Python OpenCV](https://pypi.org/project/opencv-python/): ```pip install opencv-python```
* [TensorBoardX](https://github.com/lanpa/tensorboardX): ```pip install tensorboardX```
### Note for [Anaconda Python](https://anaconda.org/anaconda/python) users
To run some of the samples, you will need these modules:
```bash
conda install pytorch torchvision -c pytorch
pip install tensorboard-pytorch
pip install gym
pip install gym[atari]
pip install opencv-python
```
## Documentation
* [Ptan introduction](docs/intro.ipynb)
Random pieces of information
* `ExperienceSource` vs `ExperienceSourceFirstLast`: https://github.com/Shmuma/ptan/issues/17#issuecomment-489584115