For AI training in NanKai University, PRC on Dec 4th, 2022

[zhuo-yoyowz]

Hi all, please leave your inference results and questions below. Thank you~

[NkuPhyHenry]

[intelyoungway]

Fantastic!

[raymondlo84]

WOW!

[MITNKU]

[zhuo-yoyowz]

Cool~

[intelyoungway]

Good work!

[MITNKU]

老师可以咨询一下，就是如果支持Interl的AVX-512 指令集的至强处理器，能否运用intel的某个类似于cuda这类的驱动，实现在CPU上可以高校训练模型，比较好奇目前Intel支持的高效模型训练的方法。

[intelyoungway]

OpenVINO最重要的作用是模型部署，以期达到最优的推理性能；
模型训练相关优化可以借助Intel oneAPI等相关工具进行尝试：比如这个 https://github.com/IntelAI/models/tree/master/quickstart/image_recognition/pytorch/resnet50/training/cpu

[MITNKU]

好的，谢谢老师讲解~

[qianqiancome]

[qianqiancome]

Run on my cpu~

[intelyoungway]

Nice work! You mastered the major point of AUTO-Device part:
1 throughput hint mode always focus on maximizing throughput;
2 while latency hint mode always focus on real-time!