# ResNet-VAE

**Repository Path**: teslatasy/ResNet-VAE

## Basic Information

- **Project Name**: ResNet-VAE
- **Description**: No description available
- **Primary Language**: Unknown
- **License**: Not specified
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 1
- **Forks**: 0
- **Created**: 2021-06-17
- **Last Updated**: 2022-05-26

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# ResNet-VAE
Can't find a satisfying VAE implementation on Github, so I decide to write one myself.


## Usage

`python train.py`


## Features

1. Closed form update for the variance of p(x | z)
2. Network structure borrowed from BigGAN
3. Parameterize the variance of q(z | x) by directly squaring the output of the network. (surprisingly stable compared with exp or softplus)