mirror of
https://github.com/abetlen/llama-cpp-python.git
synced 2023-09-07 17:34:22 +03:00
I think this is actually supported (its in the arguments of `LLama.__call__`, which is how the completion is invoked). decision: mark as supported