llama / Dockerfile
Leon4gr45's picture
Create Dockerfile
1fe2763 verified
Raw History Blame
502 Bytes
FROM ghcr.io/ggml-org/llama.cpp:server
EXPOSE 7860
ENTRYPOINT ["/app/llama-server"]
CMD [
"--host", "0.0.0.0",
"--port", "7860",
"--hf-repo", "XHToken/Spark-X2.5-1.7B-GGUF:Q8_0",
"--alias", "spark-x2.5-1.7b",
"--ctx-size", "32768",
"--parallel", "1",
"--threads", "2",
"--threads-batch", "2",
"--batch-size", "512",
"--ubatch-size", "256",
"--cache-type-k", "q8_0",
"--cache-type-v", "q8_0",
"--flash-attn", "auto",
"--cont-batching",
"--metrics",
"--no-webui"
]