Windows 11 DeepSpeed若干报错问题解决

Windows 11 DeepSpeed若干报错问题解决

1. 问题：Unable to pre-compile async_io

问题描述

 Preparing metadata (setup.py) ... error
  error: subprocess-exited-with-error

  × python setup.py egg_info did not run successfully.
  │ exit code: 1
  ╰─> [17 lines of output]
      test.c
      LINK : fatal error LNK1181: 无法打开输入文件“aio.lib”
      Traceback (most recent call last):
        File "<string>", line 2, in <module>
        File "<pip-setuptools-caller>", line 34, in <module>
        File "C:\Users\Administrator\AppData\Local\Temp\pip-install-4_iivh0f\deepspeed_07db96ef892c4625952709f60a01fd85\setup.py", line 163, in <module>
          abort(f"Unable to pre-compile {op_name}")
        File "C:\Users\Administrator\AppData\Local\Temp\pip-install-4_iivh0f\deepspeed_07db96ef892c4625952709f60a01fd85\setup.py", line 51, in abort
          assert False, msg
      AssertionError: Unable to pre-compile async_io
      Setting ds_accelerator to cuda (auto detect)
      [WARNING] Torch did not find cuda available, if cross-compiling or running with cpu only you can ignore this message. Adding compute capability for Pascal, Volta, and Turing (compute capabilities 6.0, 6.1, 6.2)
      DS_BUILD_OPS=1
       [WARNING]  async_io requires the dev libaio .so object and headers but these were not found.
       [WARNING]  If libaio is already installed (perhaps from source), try setting the CFLAGS and LDFLAGS environment variables to where it can be found.
       [WARNING]  One can disable async_io with DS_BUILD_AIO=0
       [ERROR]  Unable to pre-compile async_io
      [end of output]

解决方法

1. 首先通过git下载deepspeed：

git clone https://github.com/microsoft/DeepSpeed.git

2. 运行如下命令：

Set-Item Env:\DS_BUILD_OPS 0

3. 运行cmd或PowerShell，进入deepspeed文件夹下，运行build_win.bat编译DeepSpeed。
4. 进入dist文件夹，运行pip install命令进行安装，即可安装成功

 pip install .\deepspeed-0.14.1+unknown-py3-none-any.whl

2. 问题：ModuleNotFoundError: No module named ‘deepspeed.accelerator’

问题描述

    ...
    raise RuntimeError(
RuntimeError: Failed to import transformers.models.llama.modeling_llama because of the following error (look up to see its traceback):
No module named 'deepspeed.accelerator'

解决方法

用github上下载的deepspeed包进行安装的，是由于离线安装和在线的pip install安装方法在加载该包时方法不一样，因此会报如上错误。到离线包的位置找到accelerator，将该包拷贝到相应位置即可。

3.问题：ModuleNotFoundError: No module named ‘deepspeed.ops.op_builder’

问题描述

 raise RuntimeError(
RuntimeError: Failed to import transformers.models.llama.modeling_llama because of the following error (look up to see its traceback):
No module named 'deepspeed.ops.op_builder'

解决方法

出现该问题是因为

• 从pypi安装的torch、torchvivsion、torchaudio，没有直接从pytorch官网安装。
• 安装的CPU版本的torch，没有安装GPU版本的torch

可以运行如下命令进行安装：

pip install torch==1.9.1+cu111 torchvision==0.10.1+cu111 torchaudio==0.9.1 -f https://download.pytorch.org/whl/torch_stable.html

安装完成以后没有该报错了。