Inference planner

Work out how many GPUs a language model needs, and how fast it can answer, as you change its size, precision and users. This tool needs JavaScript, so turn it on to use it. Everything you need to answer the exam question is explained in the text above.