I’d love to see a write-up or explanation on how its was built. I work with some High School students that are trying to do something like this for their AP Comp Sci class.
I saw the episode 4, season 4 last night and the whole sequence is hilarious, and enjoyed the accurate tech talk about “it’s a good classifier at core, just need more training” that follows.
Which make sense if the HBO guy behind it was a member of your Part#1 cohort, impressive
For those without HBO access, here’s a small 1min extract.
And here’s an extended 3 min version, with a special wink wink message for you AWS users not building your own DL box ^!^
Can’t wait for the episodes talking about “basics” like over-fitting, data augmentation, BatchNorm and pseudo-labels.
With @jeremy and @rachel in cameo appearances, muahahaha !
EPB
PS: click on the YouTube icon to get the original video with max resolution.