happyz
/
llama.cpp
mirror of https://github.com/ggerganov/llama.cpp.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Allow partial success of seq_rm for hybrid memory

This commit is contained in:
Ivan Chikish 2026-02-16 20:18:02 +03:00
parent acb7c79069
commit 2124862e9b
1 changed files with 8 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

14
src/llama-memory-hybrid.cpp
View File

@ -130,12 +130,14 @@ void llama_memory_hybrid::clear(bool data) {
}
bool llama_memory_hybrid::seq_rm(llama_seq_id seq_id, llama_pos p0, llama_pos p1) {
// Try removing from the recurrent cache first since it may fail. If it does
// fail, the cache will not have been mutated.
if (!mem_recr->seq_rm(seq_id, p0, p1)) {
return false;
}
return mem_attn->seq_rm(seq_id, p0, p1);
bool res = true;
// Allow partial success when only mem_attn is modified
// This permits making cheap checkpoints if only mem_recr is saved and restored
// In that case it would require mem_attn to be manually trimmed
res &= mem_recr->seq_rm(seq_id, p0, p1);
res &= mem_attn->seq_rm(seq_id, p0, p1);
return res;
}
void llama_memory_hybrid::seq_cp(llama_seq_id seq_id_src, llama_seq_id seq_id_dst, llama_pos p0, llama_pos p1) {