version: '3.8'

# 1-Click OpenAI-Compatible API Server using llama.cpp
# Endpoint: http://localhost:8080/v1/chat/completions
services:
  slm-api:
    image: ghcr.io/ggerganov/llama.cpp:server
    container_name: sortmark-dataaudit-0.5b-api
    restart: unless-stopped
    ports:
      - "8080:8080"
    volumes:
      - ./:/models
    command:
      - "-m"
      - "/models/SORTMARK-DataAudit-0.5B-Q4_K_M.gguf"
      - "-c"
      - "2048"
      - "--host"
      - "0.0.0.0"
      - "--port"
      - "8080"
      - "--temp"
      - "0.2"
      - "-np"
      - "2"