Rose: Range-Of-Slice Equilibration optimizer
Fuente:
Zenodo
Gespeichert in:
| 1. Verfasser: | |
|---|---|
| Format: | Recurso digital |
| Sprache: | Englisch |
| Veröffentlicht: |
Zenodo
2026
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866902125036961792 |
|---|---|
| author | Kieren, Matthew Everet |
| author_facet | Kieren, Matthew Everet |
| contents | <p>The Rose optimizer rescales gradients using a per-slice `max - min` range computed by reducing all dimensions beyond the leading axis. Unlike Adam and other stateful optimizers, Rose maintains no per-parameter state between steps: no momentum buffers, variance estimates, or step counters.</p> |
| format | Recurso digital |
| id | zenodo_https___doi_org_10_5281_zenodo_19589765 |
| institution | Zenodo |
| language | eng |
| publishDate | 2026 |
| publisher | Zenodo |
| record_format | zenodo |
| spellingShingle | Rose: Range-Of-Slice Equilibration optimizer Kieren, Matthew Everet Artificial intelligence optimizer pytorch Machine learning Deep learning <p>The Rose optimizer rescales gradients using a per-slice `max - min` range computed by reducing all dimensions beyond the leading axis. Unlike Adam and other stateful optimizers, Rose maintains no per-parameter state between steps: no momentum buffers, variance estimates, or step counters.</p> |
| title | Rose: Range-Of-Slice Equilibration optimizer |
| topic | Artificial intelligence optimizer pytorch Machine learning Deep learning |
| url | https://doi.org/10.5281/zenodo.19589765 |