thek0tyara/llama-cpp-turboquant
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions 4

server : fix off-by-1 in server_tokens::size_up_to_pos() (#20279)

Browse source 
* server : fix off-by-1 in server_tokens::size_up_to_pos()

* cont : fix typo [no ci]
This commit is contained in:
Georgi Gerganov 2026-03-09 16:43:38 +02:00 • committed by GitHub
parent f76565db92
commit d6e1556499
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
3 changed files with 7 additions and 7 deletions
Download patch file Download diff file Expand all files Collapse all files

4
tools/server/server-common.cpp
View file

@ -276,7 +276,7 @@ llama_pos server_tokens::pos_next(int64_t n_tokens) const {
size_t server_tokens::size_up_to_pos(llama_pos max_pos) const {
if (!has_mtmd) {
return std::min((size_t)(max_pos + 1), tokens.size());
return std::min((size_t)max_pos, tokens.size());
}
size_t idx = 0;
@ -296,7 +296,7 @@ size_t server_tokens::size_up_to_pos(llama_pos max_pos) const {
idx++;
}
if (pos > max_pos) {
if (pos >= max_pos) {
break;
}
}