root/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggerganov/llama.cpp.git synced 2025-01-03 15:24:35 +00:00
Code Issues Actions 7 Packages Projects Releases Wiki Activity

server : fix non-transformer logic + remove response from /props

Browse Source
This commit is contained in:
Georgi Gerganov 2024-10-12 09:21:41 +03:00
parent 9ec6b49176
commit 0db72b63f5
No known key found for this signature in database
GPG Key ID: 449E073F9DC10735
1 changed files with 3 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
examples/server/server.cpp
View File

@ -2082,10 +2082,13 @@ struct server_context {
// keep only the common part
int p0 = slot.n_past;
if (!llama_kv_cache_seq_rm(ctx, slot.id + 1, p0, -1)) {
// could not partially delete (likely using a non-Transformer model)
llama_kv_cache_seq_rm(ctx, slot.id + 1, -1, -1);
p0 = 0;
// there is no common part left
slot.n_past = 0;
slot.n_past_se = 0;
@ -2773,7 +2776,6 @@ int main(int argc, char ** argv) {
const auto handle_props = [&ctx_server, &res_ok](const httplib::Request &, httplib::Response & res) {
json data = {
{ "system_prompt", "[unavailable]" },
{ "default_generation_settings", ctx_server.default_generation_settings_for_props },
{ "total_slots", ctx_server.params.n_parallel },
{ "chat_template", llama_get_chat_template(ctx_server.model) },