search

ContinualAI / avalanche

Avalanche: an End-to-End Library for Continual Learning based on PyTorch.
http://avalanche.continualai.org
MIT License
1.79k stars 292 forks source link

Support for Latent Replay #1651

Closed ZexinLi0w0 closed 3 months ago

ZexinLi0w0 commented 6 months ago

I see currently in training strategies, AR1 is implemented with latent replay, but it seems that this strategy is still in beta testing, and there still exists tight coupling among AR1, latent replay, and model=MobilenetV1 is hard coded.

Is it possible to support the latent replay method in a separate way?

🐛 Describe the bug A clear and concise description of what the bug is.

🐜 To Reproduce A minimal working example to reproduce the issue. The code should be executable without modifications.

🐝 Expected behavior A clear and concise description of what you expected to happen.

🐞 Screenshots If applicable, add screenshots to help explain your problem.

🦋 Additional context Add any other context about the problem here like your python setup.

AntonioCarta commented 6 months ago

As far as I know, each latent replay method is quite different from the others, so it may be difficult to create a general component. I think AR1 solution is quite good, so if you want to implement your own method you can easily start from that and modify it for your needs.