tqcq/llama.cpp
0
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-06-27 12:05:03 +00:00
Code Issues Packages Projects Releases Wiki Activity

kv-cache : simplify the interface (#13660)

Browse Source 
* kv-cache : simplify the interface

ggml-ci

* context : revert llama_batch_allocr position change

ggml-ci
This commit is contained in:
Georgi Gerganov
2025-05-21 15:11:13 +03:00
committed by GitHub
parent b44890df2e
commit 797f2ac062
9 changed files with 89 additions and 153 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
src/llama-batch.cpp
View File

@ -1,5 +1,6 @@
#include "llama-batch.h"
#include <cassert>
#include <cstring>
#include <algorithm>
@ -281,9 +282,10 @@ llama_batch_allocr::llama_batch_allocr(struct llama_batch in_batch, llama_pos p0
batch = in_batch;
GGML_ASSERT(batch.n_tokens > 0);
if (!batch.pos) {
assert(p0 >= 0);
pos.resize(batch.n_tokens);
for (int32_t i = 0; i < batch.n_tokens; i++) {
pos[i] = i + p0;
pos[i] = p0 + i;
}
batch.pos = pos.data();
}