Concept detail
kernel count and memory roundtrip reduction
0 tagged videos · 1 connection videos · degree 0
Publication timeline
- 2026-091 video
Source videos
Observed neighbors
Weight folding (offline gain/weight merge)
Folding gain into the next layer's weights eliminates per-pass gain computation and memory trips.
1 source videos
Weight Folding, CUDA Streams, and the Bug That Made My Model Speak Backwards — Filip MakraduliRelated tagged insights
No core insights are tagged with this concept.