3 ms·
It's literally the model's output, in real-time. The GPT-3 API lets you turn streaming on or off, but if you don't enable it you have to wait much longer for th
by jw1224 4y ago
It's literally the model's output, in real-time. The GPT-3 API lets you turn streaming on or off, but if you don't enable it you have to wait much longer for the completion to finish entirely, before getting anything back.