Skip to content

Latest commit

 

History

History
27 lines (13 loc) · 784 Bytes

README.md

File metadata and controls

27 lines (13 loc) · 784 Bytes

ECE566 Final Project Variational Auto-Encoders (vanilla, IWAE, MIWAE)

Yite Wang, Yulun Wu

This is the Pytorch implementation of VAE, IWAE and MIWAE.

To run our code:

vanilla VAE

python main.py --stamp VAE_final --vae_type vanilla --batch_size 128

IWAE

python main.py --stamp iwae_final --vae_type IWAE --batch_size 128 --num_particles 5

MIWAE

python main.py --stamp MIWAE_t1 --vae_type MIWAE --batch_size 128 --num_samples 3 --num_particles 5

Acknowledgement

  1. We adopt and modify the neural network architecture and some hyper-parameters from this repo.
  2. Original paper: VAE, IWAE, MIWAE.