llama-cli
Command-line interface for LLM inference. tldr-pages
On Linux
Linux
Start up LLM model and interface
llama-cli --model path/to/modelIdentifyStart with setup prompt context size
llama-cli --model path/to/model --ctx-size numberIdentifyStart with set offload number of layer gpu layers
llama-cli --model path/to/model --n-gpu-layers autoIdentifyStart with sending content as system prompt from a file
llama-cli --model path/to/model --system-prompt-file path/to/promptIdentifyStart with sending content in file as first prompt
llama-cli --model path/to/model --file path/to/promptIdentify