启用IPEX所需的主要变更包括：

# Import the library:

    import intel_extension_for_pytorch as ipex
    #Apply the optimizations to the model for its datatype:
    model = ipex.optimize(model)
    #torch.channels_last should be applied to both of the model object and data to raise CPU resource usage efficiency.
    model = model.to(memory_format=torch.channels_last)
    data = data.to(memory_format=torch.channels_last)

另外，请查看https://intel.github.io/intel-extension-for-pytorch/latest/tutorials/examples.html了解IPEX示例。请查看IPEX官方页面https://www.intel.com/content/www/us/en/developer/tools/oneapi/extension-for-pytorch.html。

要启用Intel Extension for Pytorch，您只需在代码中给予以下内容：

import intel_extension_for_pytorch as ipex

导入上述内容扩展了PyTorch，优化了英特尔硬件上的性能提升
之后，您必须在代码中添加此代码

model = model.to(ipex.DEVICE)

首先，您需要将Trainer对象子类化，并创建一个自定义优化器，如Hugging Face docs
用于使用intel_extension_for_pytorch的API发生了一些变化，要使用该库，您只需执行以下操作：

import intel_extension_for_prytorch as ipex

model, optimizer = ipex.optimize(model, optimizer=optimizer)

目前，Transformers 4.21已支持IPEX。IPEX图形优化采用JIT模式

python run_qa.py
    --model_name_or_path csarron/bert-base-uncased-squad-v1 \
    --dataset_name squad \
    --do_eval \
    --max_seq_length 384 \
    --doc_stride 128 \
    --output_dir /tmp/ \
    --no_cuda \
    --use_ipex \
    --jit_mode_eval