CS348Project/llama.cpp
5
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-10-28 08:31:25 +00:00
Code Issues Packages Projects Releases Wiki Activity

server : send partial stop string when <EOG> is reached (#15007)

Browse Source
This commit is contained in:
matteo
2025-10-23 11:32:24 +02:00
committed by GitHub
parent 9de9672adb
commit 8cf6b42d46
1 changed files with 1 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
tools/server/server.cpp
View File

@@ -2839,7 +2839,7 @@ struct server_context {
slot.generated_text.begin() + pos + stop_pos, slot.generated_text.begin() + pos + stop_pos,
slot.generated_text.end()); slot.generated_text.end());
pos = std::min(slot.n_sent_text, slot.generated_text.size()); pos = std::min(slot.n_sent_text, slot.generated_text.size());
} else if (slot.has_next_token) { } else if (slot.has_next_token && !llama_vocab_is_eog(vocab, result.tok) ) {
stop_pos = slot.find_stopping_strings(str_test, token_str.size(), false); stop_pos = slot.find_stopping_strings(str_test, token_str.size(), false);
send_text = stop_pos == std::string::npos; send_text = stop_pos == std::string::npos;
} }