I just did a ten minute sampling of the material: really good stuff. Gluon seems like a good API that supports a lot of languages. A little off topic, but I s

I just did a ten minute sampling of the material: really good stuff. Gluon seems like a good API that supports a lot of languages. A little off topic, but I started using TensorFlow (and later Keras) when it was first released, and it ended up being my only tool for DL work (including managing a DL team) for four+ years. In hind site I really wish I had also learned PyTorch, Gluon, etc. It doesn’t matter now, I am retired and have gone back to my roots: studying knowledge representation (including linked data/semantic web) and more ‘classic’ AI. Given my experience, I would encourage people to cast a wider net and experiment with more DL libraries.