ikawrakow/ik_llama.cpp
1
0
Fork 0
mirror of https://github.com/ikawrakow/ik_llama.cpp.git synced 2026-04-21 06:59:21 +00:00
Code Issues Packages Projects Releases Wiki Activity

Revert kv cache name changes

Browse Source
This commit is contained in:
firecoperana
2026-01-17 08:42:43 -06:00
parent 79185862e5
commit 93abfb66ab
15 changed files with 62 additions and 62 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
common/speculative.cpp
View File

@@ -258,14 +258,14 @@ std::vector<llama_token> llama_speculative_gen_draft(
}
if (reuse_i > 0) {
llama_memory_seq_rm (ctx_dft, 0, 0, reuse_i);
llama_memory_seq_add(ctx_dft, 0, reuse_i, -1, -reuse_i);
llama_kv_cache_seq_rm (ctx_dft, 0, 0, reuse_i);
llama_kv_cache_seq_add(ctx_dft, 0, reuse_i, -1, -reuse_i);
prompt_dft.erase(prompt_dft.begin(), prompt_dft.begin() + reuse_i);
}
if (reuse_n < (int) prompt_dft.size()) {
llama_memory_seq_rm (ctx_dft, 0, reuse_n, -1);
llama_kv_cache_seq_rm (ctx_dft, 0, reuse_n, -1);
prompt_dft.erase(prompt_dft.begin() + reuse_n, prompt_dft.end());
}