llama.cpp
graph : avoid huge warm-up graphs for MoE models
#14753

Merged

graph : avoid huge warm-up graphs for MoE models #14753

ggerganov merged 2 commits into master from gg/context-reduce-min-nodes

ggerganov force pushed 207 days ago

ggerganov requested a review from

slaren 207 days ago

graph : avoid huge warm-up graphs for MoE models

033b3066

ggerganov force pushed to 033b3066 207 days ago

ggerganov commented on 2025-07-18

cont : bump max nodes to 8x model tensors

5883f014

slaren approved these changes on 2025-07-18

ggerganov merged d498af3d into master 207 days ago

ggerganov deleted the gg/context-reduce-min-nodes branch 207 days ago

Reviewers

slaren

Assignees

No one assigned

Labels

None yet

Milestone

No milestone