https://github.com/PACerv/ImplicitMotion

https://github.com/PACerv/ImplicitMotion

Last synced: about 1 month ago
JSON representation

• Host: GitHub
• URL: https://github.com/PACerv/ImplicitMotion
• Owner: PACerv
• License: mit
• Created: 2022-07-12T09:18:05.000Z (over 2 years ago)
• Default Branch: main
• Last Pushed: 2022-09-23T01:05:21.000Z (over 2 years ago)
• Last Synced: 2023-02-28T16:09:25.866Z (almost 2 years ago)
• Language: Python
• Size: 9.02 MB
• Stars: 14
• Watchers: 3
• Forks: 1
• Open Issues: 1
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# Official Implementation of the paper "Implicit Neural Representations for Variable Length Human Motion Generation" (ECCV 2022)

![Main figure](etc/main.png)

[[More Visualizations]](etc/Visualizations.md)

## Bibtex

Please consider citing this work, if you find this code useful.

```

@article{cervantes2022implicit,

  title={Implicit Neural Representations for Variable Length Human Motion Generation},

  author={Cervantes, Pablo and Sekikawa, Yusuke and Sato, Ikuro and Shinoda, Koichi},

  journal={arXiv preprint arXiv:2203.13694},

  year={2022}

}

```

---

## Installation

```

pip install -r requirements.txt

pip install torch==1.8.1+cu101 torchvision==0.9.1+cu101 torchaudio==0.8.1 -f https://download.pytorch.org/whl/torch_stable.html

```

To install Pytorch3D follow the instructions [here](https://github.com/facebookresearch/pytorch3d/blob/main/INSTALL.md).

> For CUDA builds with versions earlier than CUDA 11, set `CUB_HOME`

```

pip install "git+https://github.com/facebookresearch/[email protected]"

```

--- 

## Data Preparation

### [Download datasets](etc/Datasets.md)

### [Download evaluation models](etc/EvalModels.md)

### [Download SMPL model](etc/SMPL.md)

---

## Usage

### Training

```

python3 ./ImplicitMotion/main.py 

	--path_config=/path/to/config-file

```

Before training you need to prepare a configuration file.

Configurations for the experiments in the paper are provided [here](configs).

Modifications for the following keyword arguments are necessary:

```

path_dataset: Path to dataset folder

path_results_base: Path to folder for saving checkpoints, etc. (arbitrary).

path_smpl: Path to SMPL file (.pkl)

```

### Evaluation

```

python3 ./ImplicitMotion/test/test_metric.py 

	--path_results /path/to/results

	--path_classifier /path/to/classifier

	--variable_length_testing

	--metrics

```

### Visualization

```

python3 ./ImplicitMotion/test/test_metric.py

	--path_results /path/to/results

	--path_classifier /path/to/classifier

	--variable_length_testing

	--video

	--num_videos 1

	--video_length 60

```

---

## License

This code is distributed under an [MIT LICENSE](LICENSE).