Do we need src/cudnn/activations.jl?
Open
CUDA
- Dominant language
- Julia
- Stars
- 258
- Forks
- 138
- Avg merge
- 2d 1h
- Merged PRs (30d)
- 3
Description
I noticed there was a thread on Discourse recently where this came up as the culprit: https://github.com/FluxML/NNlibCUDA.jl/blob/master/src/cudnn/activations.jl. Are these overloads still useful now (for performance) or can they be nixed?
Contributor guide
No contributing guide indexed for this repository
Research direction
Start with src/cudnn/activations.jl in NNlibCUDA.jl and review the linked Discourse thread for the reported concern. Check whether these overloads are still used and whether they provide a performance benefit. Done means documenting a supported conclusion about retaining or removing them.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- julia
- Domain
- machine-learning
- Issue type
- Refactor
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100