CS348Project/llama.cpp
5
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-10-27 08:21:30 +00:00
Code Issues Packages Projects Releases Wiki Activity

kv-cache : fix k-shift for multiple streams

Browse Source 
ggml-ci
This commit is contained in:
Georgi Gerganov
2025-07-17 20:18:36 +03:00
parent 01612b7409
commit 05baa62a73
1 changed files with 1 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
src/llama-kv-cache-unified.cpp
View File

@@ -1261,7 +1261,7 @@ void llama_kv_cache_unified::set_input_k_shift(ggml_tensor * dst) const {
const auto & cells = v_cells[s];
for (uint32_t i = 0; i < cells.size(); ++i) {
data[i] = cells.is_empty(i) ? 0 : cells.get_shift(i);
data[s*cells.size() + i] = cells.is_empty(i) ? 0 : cells.get_shift(i);
}
}
}