Cloud Stack Ninja

ONNX Runtime is a high-performance inferencing and training engine for machine learning models. This show focuses on ONNX Runtime for model inference. ONNX Runtime has been widely adopted by a variety of Microsoft products including Bing, Office 365 and Azure Cognitive Services, achieving an average of 2.9x inference speedup. Now we are glad to introduce ONNX Runtime quantization and ONNX Runtime mobile for further accelerating model inference with even smaller model size and runtime size. ONNX Runtime keeps evolving not only for cloud-based inference but also for on-device inference.

Jump To: 

  • [01:02] ONNX and ONNX Runtime overview
  • [02:26] model operationalization with ONNX Runtime
  • [04:04] ONNX Runtime adoption
  • [05:07] ONNX Runtime INT8 quantization for model size reduction and inference speedup
  • [09:46] Demo of ONNX Runtime INT8 quantization
  • [16:00] ONNX Runtime mobile for runtime size reduction

Learn More: 

Don't miss new episodes, subscribe to the AI Show  



Read more here: https://channel9.msdn.com/Shows/AI-Show/Faster-and-Lighter-Model-Inference-with-ONNX-Runtime-from-Cloud-to-Client

Content Attribution

This content was originally published by Seth Juarez at Channel 9, and is syndicated here via their RSS feed. You can read the original post over there.

%d bloggers like this: