official implementation of Fully Spiking Variational Autoencoder
Accepted to AAAI2022!!
arxiv: https://arxiv.org/abs/2110.00375
- install dependencies
pip install -r requirements.txt
- initialize the fid stats
python init_fid_stats.py
The following command calculates the Inception score & FID of FSVAE trained on CelebA. After that, it outputs demo_input.png, demo_recons.png, and demo_sample.png.
python demo.py
python main_fsvae.py exp_name -config NetworkConfigs/dataset_name.yaml
Training settings are defined in NetworkConfigs/*.yaml.
args:
- name: [required] experiment name
- config: [required] config file path
- checkpoint: checkpoint path (if use pretrained model)
- device: device id of gpu, default 0
You can watch the logs with below command and access http://localhost:8009/
tensorboard --logdir checkpoint --bind_all --port 8009
As a comparison method, we prepared vanilla VAEs of the same network architecture built with ANN, and trained on the same settings.
python main_ann_vae.py exp_name -dataset dataset_name
args:
- name: [required] experiment name
- dataset:[required] dataset name [mnist, fashion, celeba, cifar10]
- batch_size: default 250
- latent_dim: default 128
- checkpoint: checkpoint path (if use pretrained model)
- device: device id of gpu, default 0
