Saw this post trending on twitter and thought it might be a good resource for beginners / intermediates.

tweet: https://twitter.com/GokuMohandas/status/1202411040295645184

course: https://practicalai.me

The tweet has details on why the author changed from PyTorch to TensorFlow and it’s one of the few times I see concrete reasons for this specific use case. I’m an avid PyTorch dev (for research and production) but I’m willing to give TF a shot…but GOD I remember hating gradient tape.

Has anyone tried TF2.0 yet and liked it?