#!/bin/sh # Startup script for Tiny Narrator on Hugging Face Spaces free CPU. # Model inference endpoints are configured through environment variables. set -eu echo "[tiny-narrator] Starting Tiny Narrator app on ${GRADIO_SERVER_NAME:-0.0.0.0}:${PORT:-${GRADIO_SERVER_PORT:-7860}}" if [ -n "${LLAMA_CPP_BASE_URL:-}" ]; then echo "[tiny-narrator] Reader brain endpoint configured at ${LLAMA_CPP_BASE_URL}" else echo "[tiny-narrator] LLAMA_CPP_BASE_URL is not configured; reader brain will use deterministic fallback." fi exec python app.py