llama.cpp
Hybrid model cache: add `--checkpoint-every-nb`
#20087

Merged

Hybrid model cache: add `--checkpoint-every-nb` #20087

pwilkin merged 1 commit into ggml-org:master from pwilkin:batch-checkpoints

pwilkin requested a review from

ngxson 10 days ago

pwilkin requested a review from

ggerganov 10 days ago

pwilkin changed the title ~~Add `--checkpoint-every-nb`~~ Hybrid model cache: add `--checkpoint-every-nb` 10 days ago

github-actions added examples

github-actions added server

ggerganov commented on 2026-03-04

ggerganov commented on 2026-03-04

pwilkin force pushed from 68b87a89 to 6874373b 8 days ago

pwilkin requested a review from

aldehir 8 days ago

pwilkin requested a review from

CISC 8 days ago

pwilkin force pushed from 6874373b to c9d8cdc2 8 days ago

Checkpoint every n tokens: squash

516c5d60

pwilkin force pushed from c9d8cdc2 to 516c5d60 8 days ago

ggerganov approved these changes on 2026-03-06

pwilkin merged f5ddcd16 into master 7 days ago

Reviewers

ggerganov

ngxson

aldehir

CISC

Assignees

No one assigned

Labels

examples server

Milestone

No milestone