Difference between revisions of "LightR"
Jump to navigation
Jump to search
(Initial commit) |
m (Design principle & Plans) |
||
| Line 6: | Line 6: | ||
This page is under construction. For recent activities, see [[/Version History|Version History]]. | This page is under construction. For recent activities, see [[/Version History|Version History]]. | ||
| + | |||
| + | |||
| + | ; Design principle | ||
| + | : Strategy light, machine learning heavy. | ||
| + | |||
| + | ; Planned experiments | ||
| + | : Multiple hand-tuned surfing models -> Expert network & gate network | ||
| + | : Handed-crafted features with naive KNN -> Search-based sequence model | ||
| + | : Slow networks -> Knowedge Distill & Quantization aware training | ||
Revision as of 09:13, 11 February 2022
- LightR Sub-pages:
- Version History
This page is under construction. For recent activities, see Version History.
- Design principle
- Strategy light, machine learning heavy.
- Planned experiments
- Multiple hand-tuned surfing models -> Expert network & gate network
- Handed-crafted features with naive KNN -> Search-based sequence model
- Slow networks -> Knowedge Distill & Quantization aware training
