alt Hacker NewsCorrect, it's breaks the single prompt, single completion assumption baked into the frameworks. Conceptually it's still prompt/completion but for low latency response you have to do streaming KV cache prefill with a websocket server.
Correct, it's breaks the single prompt, single completion assumption baked into the frameworks. Conceptually it's still prompt/completion but for low latency response you have to do streaming KV cache prefill with a websocket server.
I imagine you have to start decoding many speculative completions in parallel to have true low latency.