💡 [REQUEST] - Write a tutorial about how to leverage AMX with PyTorch on the 4th Gen of Xeon

[mingfeima]

🚀 Descirbe the improvement or the new tutorial

The 4th Generation Intel® Xeon® Scalable Processor platform is an unique, scalable platform optimized for different workloads acceleration on AI. The new built-in AI acceleration engine, Intel® Advanced Matrix Extensions (AMX) is able to accelerate a variety of AI Inference and Training workloads (NLP, recommendation systems, image recognition…) with BF16 and INT8 datatype.

PyTorch has enabled AMX support for computation intensive operators, e.g. Conv2d, ConvTranspose2d, Linear, MatMul, bmm with torch.bfloat16 datatype and int8 on the quantization backend. It is better to write a tutorial to tell users how to leverage AMX on PyTorch.

Existing tutorials on this topic

No response

Additional context

We aim to complete the document as part of PyTorch Docathon 2023. cc @jgong5 @XiaobingSuper @sanchitintel @ashokei @jingxu10 @ZailiWang @ZhaoqiongZ @leslie-fang-intel @Xia-Weiwen @sekahler2 @CaoE @zhuhaozhe @Valentine233 @sekyondaMeta @svekars @carljparker @NicolasHug @kit1980 @subramen @CaoE

[CaoE]

/assigntome