https://github.com/le-big-mac/vcl_diffusion

Variational continual learning of a conditional diffusion model to generate MNIST. Based on 'Conditional Diffusion MNIST'.
https://github.com/le-big-mac/vcl_diffusion

bayesian-deep-learning continual-learning deep-learning diffusion-models neural-network

Last synced: 7 months ago
JSON representation

Variational continual learning of a conditional diffusion model to generate MNIST. Based on 'Conditional Diffusion MNIST'.

• Host: GitHub
• URL: https://github.com/le-big-mac/vcl_diffusion
• Owner: le-big-mac
• License: mit
• Created: 2024-04-02T01:18:19.000Z (about 2 years ago)
• Default Branch: main
• Last Pushed: 2024-05-22T14:20:04.000Z (almost 2 years ago)
• Last Synced: 2025-07-23T10:52:24.130Z (9 months ago)
• Topics: bayesian-deep-learning, continual-learning, deep-learning, diffusion-models, neural-network
• Language: Python
• Homepage:
• Size: 46.5 MB
• Stars: 0
• Watchers: 0
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

          
# Variational continual learning for diffusion models.

A codespace for training diffusion models with [variational continual learning](https://arxiv.org/abs/1710.10628), to see if it can mitigate catastrophic forgetting.



  

    

      

        

        


        Standard training.

      

      

        

        


        VCL training.

      

    

    

      

        Images generated during continual learning of MNIST digits, with no data replay. Each row i shows generations from when the model had been trained on

            the first i tasks (digits), each column j shows generations for digit j.

      

    

  



Note that the variational inference for VCL is very expensive, so this code will take several hours to run on a powerful GPU.

This diffusion model in this project is based on original work by Tim Pearce, released under the MIT License in 2022. The original code can be found at [Conditional Diffusion MNIST](https://github.com/TeaPearce/Conditional_Diffusion_MNIST). The modifications made by me in 2024 are also released under the MIT License.