ONNX (Open Neural Network Exchange) was established in December 2017 as an open source format for machine learning models (Deep Learning and traditional ML). Backed by support from over 20 industry leading companies including Microsoft, Facebook, Amazon, Intel, NVIDIA, and more, ONNX provides data scientists with the choice to select the right tools for their task, and offers software and hardware developers a common standard to build optimizations on. We will discuss the scenarios that ONNX enables with a technical overview of the format itself.
There are several ways to obtain an ONNX model, including selecting popular pre-trained models from the ONNX Model Zoo, exporting/converting an existing model trained on another framework (including PyTorch/Caffe2, CNTK, Keras, Scikit-Learn, Tensorflow, Chainer, and more), or training a new model using services such as Azure Machine Learning or Azure Custom Vision Service. We will demystify the process and show several examples of how this can be done easily.
The ONNX model can then be operationalized using an inference runtime such as ONNX Runtime on a variety of hardware endpoints. Hardware companies are plugging in their accelerators to provide maximum efficiency in latency and resource utilization on cloud and edge. We will discuss how Intel, NVidia, and others are participating and the performance gains we are seeing on our own models at Microsoft.
Henry Zeng is a principal program manager in the Cloud AI Group at Microsoft, where he works with engineering team, partners and customers to ensure the success of ML platform. He has been in AI and data area for more than 10 years from database, NoSQL, Hadoop ecosystem, machine learning to deep learning. Prior to this role, he was the lead AI solution architect in Microsoft China working with partners and customer to land AI solutions in manufactory, retail, education and public service etc with Microsoft AI offerings. Henry holds a MS in computer science from Wuhan University.
Emma Ning is senior Program Manager in Microsoft Cloud&AI ML Platform team, focusing on AI model operationalization and acceleration with ONNX/ONNXRuntime in support of Microsoft’s strategic investment for open and interoperable AI. She had been driving search engine experience for more than 5 years and later on 2 years on exploring adoption of AI among various businesses. Emma holds a MS in computer science from Institute of Computing Technology, Chinese Academy of Sciences.
Help us make this conference the best it can be for you. Have questions you'd like this speaker to address? Suggestions for issues that deserve extra attention? Feedback that you'd like to share with the speaker and other attendees?
Join the conversation here (requires login)
©2019, O'Reilly Media, Inc. • (800) 889-8969 or (707) 827-7019 • Monday-Friday 7:30am-5pm PT • All trademarks and registered trademarks appearing on oreilly.com are the property of their respective owners. • email@example.com