A very simple interactive demo to understand the common LLM samplers. Released under the Apache License, version 2.0.
git clone https://github.com/ggerganov/llama.cpp
make -C llama.cpp server
# json array of strings (prompts)
$EDITOR in.json
systemd-run --user ./llama.cpp/server --host 127.0.0.1 --port 24498 -c 8192 -np 16 -cb -m foo.gguf
make out.json