Accelerate TensorFlow* Model Inference on CPUs with Intel® AI Technology

This training session focuses on:

  • Intel® Optimization of TensorFlow* on an Intel® Xeon® platform
  • AI model optimization quantification tool: Intel® Neural Compressor

A demo shows the following process:

  1. Train and get an FP32 TensorFlow model.
  2. Use the Intel Neural Compressor to quantize and optimize the FP32 model to get an INT8 model.
  3. Test and compare the performance improvement and accuracy loss of FP32 and INT8 models on an Intel Xeon platform with Intel® Deep Learning Boost technology in the Intel® Developer Cloud.


Zhang (Neo) Jianyu is a senior software engineer of Intel® AI software solutions. He focuses on AI solutions and performance optimization on Intel® platforms (CPUs and GPUs). He has a master's degree in pattern recognition and AI from Northwestern Polytechnical University. Zhang has experience in AI, virtualization, communication, and embedded software development.