https://github.com/minitechy/betavae_variants

beta-VAE variants trained using deep neural networks on the dSprites dataset with PyTorch
https://github.com/minitechy/betavae_variants

beta-vae deep-neural-networks disentanglement dsprites mig mutual-information pytorch sap

Last synced: 3 days ago
JSON representation

beta-VAE variants trained using deep neural networks on the dSprites dataset with PyTorch

• Host: GitHub
• URL: https://github.com/minitechy/betavae_variants
• Owner: Minitechy
• Created: 2025-01-07T14:08:17.000Z (17 days ago)
• Default Branch: main
• Last Pushed: 2025-01-15T01:32:45.000Z (10 days ago)
• Last Synced: 2025-01-22T04:15:50.142Z (3 days ago)
• Topics: beta-vae, deep-neural-networks, disentanglement, dsprites, mig, mutual-information, pytorch, sap
• Language: Jupyter Notebook
• Homepage:
• Size: 9.97 MB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
The results for training

1. $\lambda\beta$-VAE (lambda_beta_VAE.ipynb) can be found at the following link:

    https://drive.google.com/file/d/1r6zRZwzRmI9F6S8PY2LN20GoobCJfJsH/view?usp=sharing

   

2. $\lambda\beta$-VAE with an optional mutual information penalty (lambda_beta_VAE_MI.ipynb) can be found at the following link: 

    https://drive.google.com/file/d/1DzAMU-bb__ycy60_99REopPoRgxWCVuy/view?usp=sharing