ONNX and ONNX Runtime 

OMG, please show the slides more than the presenter. I can never finish reading most of the slides because they are shown for only a few seconds. This is very, very, very annoying and disappointing.

Really awesome presentation skills, simply precise and clear.

very informative talk!

Good presentation badly edited, the ratio of presenter footage to slides is awful. Some slides are barely on the screen for a second.

Thanks for the video! I am interested in adding a new custom operator to onnxruntime in C++, but there isn't any example for it. Does anyone know of one?

39:03 references

Too short a time showing slides. Constantly need to backup & pause to digest slides.

Nice work.

Can I train it on Pytorch in Python and run the model in in Java, Javascript, C++, and more? I am talking about RNN, Transformer, fastRCNN, and more advance model.

I cant get mlmodel to onnx converter to work on mac

Pranav Sharma bhai, please bring toilets to us in India