Fixing Weight Decay Regularization in Adam
We note that common implementations of adaptive gradient algorithms, such as Adam, limit the potential benefit of weight decay regularization, because the weights do not decay m...
Publications
1 shownFrequent Co-Authors
Researcher Info
- h-index
- 1
- Publications
- 1
- Citations
- 1,137
External Links
Identifiers
- ORCID
- 0000-0001-9007-2064
Impact Metrics
h-index
1
h-index: Number of publications with at least h citations each.