ContinualAI / ContinualAI/avalanche
Adding CL subnetworking strategies (WSN and RMN)
- Dominant language
- Python
- Stars
- 2.1k
- Forks
- 321
- PR merge metrics
- No merged PRs in 30d
Description
### Discussed in https://github.com/ContinualAI/avalanche/discussions/1641
Originally posted by **niniack** May 8, 2024
Hi,
We typically have a roadmap issue pinned but I didn't see that, so I'm just starting the conversation here.
**I was hoping to contribute two subnetworking strategies (Winning Subnetworks and Relevance Mapping Networks).** I think I have implementations that are architecture agnostic, unlike the ones published by the original authors.
If someone from the core team can just give me a thumbs up that here is interest in this, then I'd like to open up a PR soon! While I have a close to merge-able implementation, I imagine I will still have to do some cleaning up, more testing, and a bit of back and forth in design decisions!
WSN: https://arxiv.org/abs/2303.14962
RMN: https://arxiv.org/abs/2102.11343
Contributor guide
Assessment
This issue has not been assessed yet.