ikawrakow/ik_llama.cpp
1
0
Fork 0
mirror of https://github.com/ikawrakow/ik_llama.cpp.git synced 2026-02-27 08:34:09 +00:00
Code Issues Packages Projects Releases Wiki Activity

server: enable checkpoint for recurrent models (#1310)

Browse Source 
* server: enable checkpoint for recurrent models

create checkpoint after cancel

fix ban string and rm context during rewind

add checkpoint interval

only save recurrent cache

* save checkpoint during pp

---------

Co-authored-by: firecoperana <firecoperana>
This commit is contained in:
firecoperana
2026-02-25 23:51:18 -06:00
committed by GitHub
parent 216f44363f
commit 3fac78c48b
11 changed files with 204 additions and 111 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
examples/server/server-task.cpp
View File

@@ -1117,7 +1117,7 @@ bool server_prompt_cache::load(server_prompt& prompt, const server_tokens& token
if (it_best != states.end()) {
LLAMA_LOG_INFO(" - found better prompt with f_keep = %.3f, sim = %.3f, n_keep = %d, n_discarded_prompt = %d\n", f_keep_best, sim_best, it_best->n_kept_prompt, it_best->n_discarded_prompt);
const size_t size = it_best->data.size();
const size_t n = llama_state_seq_set_data(ctx, it_best->data.data(), size, id_slot);
const size_t n = llama_state_seq_set_data(ctx, it_best->data.data(), size, id_slot, 0);
if (n != size) {
LLAMA_LOG_INFO("failed to restore state with size %zu\n", size);
return false;