An Implementation on Building Advanced Multi-Endpoint Machine Learning APIs with LitServe: Batching, Streaming, Caching, and Local Inference