llama.cpp
llama.cpp copied to clipboard
Does the app installed via examples/llama.android have the large model only in completion mode, without chat mode?
As shown in the image, I selected the qwen3 1.7b_q8 model. When I ask "what is apple," it operates in completion mode; when I ask "apple is," it still remains in completion mode. Is the app installed via examples/llama.android only capable of text completion, not chat? Chat mode seems more important.