mirror of
https://github.com/cloudstack-llc/ollama.git
synced 2026-07-16 10:34:26 -04:00
llm: add back check for empty token cache
This commit is contained in:
Vendored
+1
-1
@@ -1032,7 +1032,7 @@ struct llama_server_context
|
||||
slot.has_next_token = false;
|
||||
}
|
||||
|
||||
if (llama_token_is_eog(model, result.tok))
|
||||
if (!slot.cache_tokens.empty() && llama_token_is_eog(model, result.tok))
|
||||
{
|
||||
slot.stopped_eos = true;
|
||||
slot.has_next_token = false;
|
||||
|
||||
Reference in New Issue
Block a user