kv cache : optimize restoring non-contiguous cells #27991
itsnotoger
changed the title Kv restore run batch kv cache : optimize restoring non-contiguous cells 52 days ago
kv cache : batch state restore scatter reads per contiguous run
ae89c2cc
context : fall back to the byte cursor when read and write chunking d…
a0a8aab5
tests : cover state restore scatter reads on host and on-device paths
e1e96a8b
Merge branch 'ggml-org:master' into kv-restore-run-batch
9e352f51
ggerganov
merged
2d8d612e
into master 50 days ago
itsnotoger
deleted the kv-restore-run-batch branch 44 days ago
Login to write a write a comment.
Login via GitHub