Everything You Want to Know About ONNX

Open Neural Network Exchange (ONNX) is an open format built to represent machine learning models. It defines the building blocks of machine learning and deep learning models along with a common file format to enable AI developers to use models with a variety of frameworks, tools, runtimes, and compilers. Watch this webinar recording to understand what is ONNX and how to get started with it.

[rahulkalidindi]

Always loved how you explain each topic with such detail. Appreciate if you post an article or video on how you keep up with all the technologies and best sources to read about them.

[qiulinzhang]

Excellent works for beginners like me, the way from the big macro architecture and their inner relations to the small code details makes me understand onnx more clearly. I can say this is one of the best material for me to take one step into model deloyment. Thank you.

[jameslu]

Explained a quite involved subject in a very clear manner, job well done!

[ravitan]

Fantastic explanation. When I started I had no clue what ONNX is, but after watching this one hour video I am pretty confident before start to working with ONNX. I, however, request you to please upload video on how to use ONNX model for production purposes like running it on mobile phones.

[asadraza]

Thanks for the video. it is very informative and helpful in understanding ONNX

[harshalnayanrathi]

A great tutorial on ONXX.
Could you please make the code available🙏

[prudvi]

Thanks for this amazing video! I was able to convert the .onnx model to .pb but it doesn't show much on netron and the inference is incorrect. Any suggestions where I might be messing up? (was able to get proper inference on the .onnx model converted from .pth)

[tmbrye]

I've been tasked with writing a scoring program for any ONNX model that I'm giving. My question is like in your example you used the Iris dataset. I did not see anywhere in the model where it defines that your float array must be in the exact order of say sepal length, sepal width, petal length, and petal width. Is this information not available in the model and if sharing a model with others must you give them this information by hand or is it contained in the model somewhere? Thanks.

[anuradhadas]

"The name 'add_4:0' refers to a Tensor which does not exist. The operation, 'add_4', does not exist in the graph."

This is the error am getting whenever am trying to run the final infer through tensorflow. (1.02.46)
Any idea how to solve this? Or any material I can refer to understand how to define these names?

[akash_a_desai]

can you please explain how to do inference of yolo v3-v4 models?

[kashitaniranjanudayangag.w]

when i call to export onnx to tf model, it gives following error. can you please help? x = input_dict[node.inputs[0]]

KeyError: 'input.1'

[Sydney]

How do you remove ONNX from a Windows 10 PC ?

[QuotableWebisodes]

Awesome webinar. I wanted some help regarding ONNX runtime and I’m unable to get it anywhere on the internet. I request you to reply to this comment if you will be able to help sir.

[bgovi.]

What language is the main source code written in?