EngineX-Ascend/enginex-ascend-910-llama.cpp
10
0
Fork 0
Code Issues Pull Requests Actions 4 Projects Releases Wiki Activity

kv-cache : remove LLAMA_SET_ROWS checks (#15505)

Browse Source 
ggml-ci
This commit is contained in:
Georgi Gerganov
2025-08-28 12:27:02 +03:00
committed by GitHub
parent 64387f6e95
commit 8a4280ce43
7 changed files with 23 additions and 141 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
src/llama-graph.cpp
View File

@@ -314,8 +314,6 @@ bool llm_graph_input_attn_kv::can_reuse(const llm_graph_params & params) {
res &= self_kq_mask->ne[0] == mctx->get_n_kv();
res &= self_kq_mask->ne[1] == GGML_PAD(params.ubatch.n_tokens, GGML_KQ_MASK_PAD);
res &= mctx->get_supports_set_rows(); // TODO: tmp
return res;
}
@@ -350,8 +348,6 @@ bool llm_graph_input_attn_kv_iswa::can_reuse(const llm_graph_params & params) {
res &= self_kq_mask_swa->ne[0] == mctx->get_swa()->get_n_kv();
res &= self_kq_mask_swa->ne[1] == GGML_PAD(params.ubatch.n_tokens, GGML_KQ_MASK_PAD);
res &= mctx->get_base()->get_supports_set_rows(); // TODO: tmp
return res;
}