ContinualAI / ContinualAI/continual-learning-baselines
Two Different Buffers Are Used On LaMAML Strategy
Open
- Dominant language
- Python
- Stars
- 358
- Forks
- 48
- PR merge metrics
- No merged PRs in 30d
Description
I think that there is a mistake in the lamaml.py files.
The LaMAML Strategy is still using a ReplayPlugin, but after switching from lamaml to lamaml_v2, the buffer should already be handled internally by the LaMAML strategy, and the plugin is adding an additional buffer that is not supposed to exist.
I think the ReplayPlugin should be removed.
If someone can confirm this is the case, I can quickly fix the problem.
Contributor guide
No contributing guide indexed for this repository
Assessment
This issue has not been assessed yet.