llama.cpp
kv-cache : rework kv_cell
#13706
Merged

kv-cache : rework kv_cell #13706

ggerganov merged 7 commits into master from gg/kv-cache-simplify-part2
ggerganov
ggerganov
ggerganov ggerganov requested a review from slaren slaren 141 days ago
slaren
slaren commented on 2025-05-22
ggerganov ggerganov force pushed from 0a8cdc3a to eda2e136 140 days ago
slaren
slaren approved these changes on 2025-05-23
ggerganov kv-cache : rework kv_cell
be9558e3
ggerganov kv-cells : use "shift" instead of "delta" consistently
71be7e50
ggerganov llama : add llama_max_parallel_sequences()
7b3f12a8
ggerganov kv-cells : update comments [no ci]
6221dd29
ggerganov context : fail upon construction if sequences exceed max value
43b40d3f
ggerganov kv-cells : get_pos() -> pos_get() + comments
f71e737a
ggerganov ggerganov force pushed from 1ec785c7 to 0dc48042 139 days ago
ggerganov kv-cells : fix tracking of "used" cells
dd394a69
ggerganov ggerganov force pushed from 0dc48042 to dd394a69 139 days ago
ggerganov
ggerganov ggerganov merged de2ef53a into master 138 days ago
ggerganov ggerganov deleted the gg/kv-cache-simplify-part2 branch 138 days ago

Login to write a write a comment.

Login via GitHub

Reviewers
Assignees
No one assigned
Labels
Milestone