https://github.com/adijo/svhn-conv-vae
An implementation of a Convolutional VAE on the SVHN dataset.
https://github.com/adijo/svhn-conv-vae
deep-learning machine-learning pytorch svhn-dataset vae
Last synced: 12 months ago
JSON representation
An implementation of a Convolutional VAE on the SVHN dataset.
- Host: GitHub
- URL: https://github.com/adijo/svhn-conv-vae
- Owner: adijo
- License: apache-2.0
- Created: 2019-04-16T19:05:37.000Z (almost 7 years ago)
- Default Branch: master
- Last Pushed: 2019-04-25T21:59:46.000Z (almost 7 years ago)
- Last Synced: 2025-02-01T08:14:22.444Z (about 1 year ago)
- Topics: deep-learning, machine-learning, pytorch, svhn-dataset, vae
- Language: Python
- Size: 4.57 MB
- Stars: 1
- Watchers: 2
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Convolutional Variational Autoencoder on the SVHN Dataset
## Data Description
* 10 classes, `73257` train images
* `3 x 32 x 32` tensors.
## Install
```
python3 -m venv venv
source venv/bin/activate
pip install requirements.txt
```
## Train
```
python train.py --gen_images_dir images --num_epochs=100 --batch_size=64
```
Images will be samples and generated at the end of each epoch in the `--gen_images-dir` directory.
## Sample Generated Images (103 epochs)
