/3D-GAE

A generalized autoencoder for volumetric shape generation

Primary LanguagePythonMIT LicenseMIT

3D-GAE

This is the tf.keras implementation of the volumetric generalized autoencoder (GAE) described in the paper "Generalized Autoencoder for Volumetric Shape Generation".

Preparing the Data

Some experimental shapes from the COSEG and auto-aligned ModelNet40 datasets are saved in the datasets folder. The model consumes volumetric shapes compressed in the TAR file format. For details about the structure and preparation of the TAR files, please refer to voxnet.

Training

python train.py

Testing

python test.py