I am on lesson 1. I executed the notebook cell that generated 10 tokens and timed the process, which took 90+ sec on the Deeplearning hardware and 1.x seconds for the instructor. Is that expected? Is the instructor running on hardware that is much more performant than the students? That is fine, of course, but I want to adjust my expectations accordingly. Thanks for what looks like a great course!
Gordon
Hey @Erlebach, when we recorded the lessons, we were running on the same hardware and environment that the learners would be using. It might be the case that these instances are shared by multiple learners or something else that’s eating up resources in the background (I am not familiar with their process for managing the hardware), so it may be worth restarting the session and seeing if you get better results.