https://github.com/gaborvecsei/variational-autoencoder
Variational-Autoencoder w/ Convolutional Layers
https://github.com/gaborvecsei/variational-autoencoder
cnn convolutional-networks convolutional-neural-networks deep-learning image-reconstruction keras machine-learning python research tensorflow vae variational-autoencoder
Last synced: 11 months ago
JSON representation
Variational-Autoencoder w/ Convolutional Layers
- Host: GitHub
- URL: https://github.com/gaborvecsei/variational-autoencoder
- Owner: gaborvecsei
- Created: 2019-03-20T22:12:03.000Z (almost 7 years ago)
- Default Branch: master
- Last Pushed: 2019-03-24T17:47:23.000Z (almost 7 years ago)
- Last Synced: 2025-03-27T06:44:46.746Z (12 months ago)
- Topics: cnn, convolutional-networks, convolutional-neural-networks, deep-learning, image-reconstruction, keras, machine-learning, python, research, tensorflow, vae, variational-autoencoder
- Language: Jupyter Notebook
- Homepage:
- Size: 731 KB
- Stars: 3
- Watchers: 2
- Forks: 2
- Open Issues: 1
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Variational Autoencoder - VAE
In this repository you will find a *Variational Autoencoder* implementation which uses
*Convolutional* layers to encode the input images to a latent vector, and *Traansposed Convolutional* layers to
reconstruct the encoded vectors into images.
## Latent Space Playground
Check out the [notebook](latent_playground.ipynb) which contains the code for the experiments
### Feature Importance
We can measure that in the latent space which feature is the most important for the reconstruction.
By "most important" I mean the ones which contribute to the bigger changes on the decoded images.
### Interactive Image Reconstruction
In this part of the [notebook](latent_playground.ipynb), you can play with the latent space to generate your own
images based on the latent vector values.
