automl / automl/ConfigurableOptimizer
Optimize memory usage in weight-entanglement
Open
Low priority
- Dominant language
- Python
- Stars
- 12
- Forks
- 2
- PR merge metrics
- No merged PRs in 30d
Description
In the weight-entanglement case, only the weights of the largest kernel is used but the smaller ones are still loaded onto the GPU. Freeing up the weights of the smaller operations can save a lot of GPU memory.
Contributor guide
No contributing guide indexed for this repository
Assessment
This issue has not been assessed yet.