Modrepo

# Inference server code @server.on_startup async def load_production_model(): # Modrepo handles the download and caching model = await modrepo.get_model("sentiment_net", stage="production") app.state.model = model