llama.cpp
kv cache slot search improvements
#3493
Merged

kv cache slot search improvements #3493

KerfuffleV2
KerfuffleV2 kv cache slot search improvements
abafd01e
KerfuffleV2
KerfuffleV2 commented on 2023-10-05
KerfuffleV2
KerfuffleV2 commented on 2023-10-05
KerfuffleV2
KerfuffleV2 KerfuffleV2 requested a review from ggerganov ggerganov 1 year ago
KerfuffleV2 Use n_ctx in kv find slot for consistency
3144563d
ggerganov
KerfuffleV2
ggerganov
KerfuffleV2 Ensure kv cache head points to a valid slot in llama_decode internal
465b8f4f
ggerganov
ggerganov approved these changes on 2023-10-06
KerfuffleV2
KerfuffleV2 KerfuffleV2 merged 9ca79d5c into master 1 year ago
KerfuffleV2 KerfuffleV2 deleted the feat-kv_cache_improvements branch 1 year ago

Login to write a write a comment.

Login via GitHub

Reviewers
Assignees
No one assigned
Labels
Milestone