CS348Project/llama.cpp
5
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-11-04 09:32:00 +00:00
Code Issues Packages Projects Releases Wiki Activity

kv-cache : fix SWA checks + disable cacheless iSWA (#15811)

Browse Source 
ggml-ci
This commit is contained in:
Georgi Gerganov
2025-09-05 10:39:22 +03:00
committed by GitHub
parent 5d6688de08
commit c610b6c11b
9 changed files with 29 additions and 11 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
src/llama-kv-cache.cpp
View File

@@ -27,10 +27,11 @@ llama_kv_cache::llama_kv_cache(
uint32_t n_seq_max,
uint32_t n_pad,
uint32_t n_swa,
llama_swa_type swa_type,
const layer_filter_cb & filter,
const layer_reuse_cb & reuse) :
model(model), hparams(model.hparams), v_trans(v_trans),
n_seq_max(n_seq_max), n_stream(unified ? 1 : n_seq_max), n_pad(n_pad), n_swa(n_swa) {
n_seq_max(n_seq_max), n_stream(unified ? 1 : n_seq_max), n_pad(n_pad), n_swa(n_swa), swa_type(swa_type) {
GGML_ASSERT(kv_size % n_pad == 0);
@@ -1392,7 +1393,7 @@ ggml_cgraph * llama_kv_cache::build_graph_shift(llm_graph_result * res, llama_co
}
bool llama_kv_cache::is_masked_swa(llama_pos p0, llama_pos p1) const {
return hparams.is_masked_swa(p0, p1);
return llama_hparams::is_masked_swa(n_swa, swa_type, p0, p1);
}
void llama_kv_cache::state_write(llama_io_write_i & io, llama_seq_id seq_id, llama_state_seq_flags flags) const {