EfficientZero

https://github.com/yewr/efficientzero

Admin

EfficientZero is a sample-efficient model-based reinforcement learning algorithm that masters Atari...

Tokens:1,095

Snippets:14

Trust Score:6.9

License:GPL-3.0

Update:1 month ago

Show doc for...

Tokens:

Raw

### Install dependencies

Source: https://github.com/yewr/efficientzero/blob/main/README.md

Install the required Python packages for the codebase.

```bash
pip install -r requirements.txt
```

--------------------------------

### Train and test via CLI

Source: https://github.com/yewr/efficientzero/blob/main/README.md

Commands to start training or testing the model using the main.py entry point.

```bash
python main.py --env BreakoutNoFrameskip-v4 --case atari --opr train --amp_type torch_amp --num_gpus 1 --num_cpus 10 --cpu_actor 1 --gpu_actor 1 --force
```

```bash
python main.py --env BreakoutNoFrameskip-v4 --case atari --opr test --amp_type torch_amp --num_gpus 1 --load_model --model_path model.p \
```

--------------------------------

### Build external packages

Source: https://github.com/yewr/efficientzero/blob/main/README.md

Build the C++/Cython external packages required before starting training. GCC version 7.5+ is required.

```bash
cd core/ctree
bash make.sh
```

--------------------------------

### Train and test via shell scripts

Source: https://github.com/yewr/efficientzero/blob/main/README.md

Use the provided shell scripts for training and evaluation.

```bash
bash train.sh
```

```bash
bash test.sh
```

--------------------------------

### Citation

Source: https://github.com/yewr/efficientzero/blob/main/README.md

BibTeX citation for the EfficientZero paper.

```bibtex
@inproceedings{ye2021mastering,
  title={Mastering Atari Games with Limited Data},
  author={Weirui Ye, and Shaohuai Liu, and Thanard Kurutach, and Pieter Abbeel, and Yang Gao},
  booktitle={NeurIPS},
  year={2021}
}
```

Tokens:

Raw