vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

llama : KV cache view API + better KV cache management (#4170)

Browse source 
* llama : keep track of used KV cells + better KV cache management

* llama : zero KV cache used upon clear

ggml-ci

* llama : allow exporting a view of the KV cache (#4180)

* Allow exporting a view of the KV cache

* Allow dumping the sequences per cell in common

* Track max contiguous cells value and position as well

* Fix max contiguous empty cells index calculation

Make dump functions deal with lengths or sequences counts > 10 better

* Fix off by one error in dump_kv_cache_view

* Add doc comments for KV cache view functions

Eliminate cell sequence struct; use llama_seq_id directly

Minor cleanups

* common : add -dkvc arg for enabling kv cache dumps

---------

Co-authored-by: Kerfuffle <44031344+KerfuffleV2@users.noreply.github.com>
This commit is contained in:
Georgi Gerganov 2023-11-23 19:07:56 +02:00 • committed by GitHub
parent d103d935c0
commit 6b0a7420d0
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
5 changed files with 277 additions and 7 deletions
Download patch file Download diff file Expand all files Collapse all files

11
common/common.h
View file

@ -122,6 +122,7 @@ struct gpt_params {
bool numa = false; // attempt optimizations that help on some NUMA systems
bool verbose_prompt = false; // print prompt tokens before generation
bool infill = false; // use infill mode
bool dump_kv_cache = false; // dump the KV cache contents for debugging purposes
// multimodal models (see examples/llava)
std::string mmproj = ""; // path to multimodal projector
@ -218,3 +219,13 @@ std::string get_sortable_timestamp();
void dump_non_result_info_yaml(
FILE * stream, const gpt_params & params, const llama_context * lctx,
const std::string & timestamp, const std::vector<int> & prompt_tokens, const char * model_desc);
//
// KV cache utils
//
// Dump the KV cache view with the number of sequences per cell.
void dump_kv_cache_view(const llama_kv_cache_view & view, int row_size = 80);
// Dump the KV cache view showing individual sequences in each cell (long output).
void dump_kv_cache_view_seqs(const llama_kv_cache_view & view, int row_size = 40);