Inference API with batched GPU serving, the answer

Check out shared "Inference API with batched GPU serving, the answer" scene on Excalidraw+