InternLM / InternEvo

InternEvo is an open-sourced lightweight training framework aims to support model pre-training without the need for extensive dependencies.
https://internevo.readthedocs.io/zh-cn/latest/?badge=latest
Apache License 2.0
310 stars 52 forks source link

[Bug] 昇腾910安装internLM环境时报错需要nvcc #298

Closed tungsten106 closed 2 months ago

tungsten106 commented 3 months ago

描述该错误

$ pip install InternEvo Collecting InternEvo Downloading InternEvo-0.5.2-py3-none-any.whl.metadata (5.6 kB) Collecting transformers<4.30.0 (from InternEvo) Downloading transformers-4.29.2-py3-none-any.whl.metadata (112 kB) Collecting sentencepiece (from InternEvo) Downloading sentencepiece-0.2.0-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl.metadata (7.7 kB) Requirement already satisfied: numpy in /root/miniconda3/envs/torch_npu/lib/python3.10/site-packages (from InternEvo) (1.26.4) Collecting tqdm (from InternEvo) Downloading tqdm-4.66.5-py3-none-any.whl.metadata (57 kB) Requirement already satisfied: psutil in /root/miniconda3/envs/torch_npu/lib/python3.10/site-packages (from InternEvo) (6.0.0) Collecting packaging (from InternEvo) Downloading packaging-24.1-py3-none-any.whl.metadata (3.2 kB) Collecting pre-commit (from InternEvo) Downloading pre_commit-3.8.0-py2.py3-none-any.whl.metadata (1.3 kB) Collecting ninja (from InternEvo) Downloading ninja-1.11.1.1-py2.py3-none-manylinux2014_aarch64.manylinux_2_17_aarch64.whl.metadata (5.3 kB) Collecting gputil (from InternEvo) Downloading GPUtil-1.4.0.tar.gz (5.5 kB) Preparing metadata (setup.py) ... done Collecting pytest (from InternEvo) Downloading pytest-8.3.2-py3-none-any.whl.metadata (7.5 kB) Collecting boto3 (from InternEvo) Downloading boto3-1.34.159-py3-none-any.whl.metadata (6.6 kB) Collecting botocore (from InternEvo) Downloading botocore-1.34.159-py3-none-any.whl.metadata (5.7 kB) Collecting pyecharts (from InternEvo) Downloading pyecharts-2.0.6-py3-none-any.whl.metadata (1.3 kB) Collecting py-libnuma (from InternEvo) Downloading py_libnuma-1.2-py3-none-any.whl.metadata (3.0 kB) Collecting pynvml (from InternEvo) Downloading pynvml-11.5.3-py3-none-any.whl.metadata (8.8 kB) Collecting tensorboard (from InternEvo) Downloading tensorboard-2.17.0-py3-none-any.whl.metadata (1.6 kB) Collecting rotary-emb (from InternEvo) Downloading rotary_emb-0.5.2.tar.gz (3.6 kB) Preparing metadata (setup.py) ... error error: subprocess-exited-with-error

× python setup.py egg_info did not run successfully. │ exit code: 1 ╰─> [22 lines of output] Traceback (most recent call last): File "", line 2, in File "", line 34, in File "/tmp/pip-install-5lvk869c/rotary-emb_d0707a936625419183e4f10e232debec/setup.py", line 90, in raise_if_cuda_home_none("rotary_emb") File "/tmp/pip-install-5lvk869c/rotary-emb_d0707a936625419183e4f10e232debec/setup.py", line 44, in raise_if_cuda_home_none raise RuntimeError( RuntimeError: rotary_emb was requested, but nvcc was not found. Are you sure your environment has nvcc available? If you're installing within a container from https://hub.docker.com/r/pytorch/pytorch, only images whose names contain 'devel' will provide nvcc.

  Warning: Torch did not find available GPUs on this system.
   If your intention is to cross-compile, this is not an error.
  By default, Apex will cross-compile for Pascal (compute capabilities 6.0, 6.1, 6.2),
  Volta (compute capability 7.0), Turing (compute capability 7.5),
  and, if the CUDA version is >= 11.0, Ampere (compute capability 8.0).
  If you wish to cross-compile for a single specific architecture,
  export TORCH_CUDA_ARCH_LIST="compute capability" before running setup.py.

环境信息

torch==2.1.0  
torch-npu==2.1.0.post3
python==3.10
CANN 8.0.RC1

安装方法如下:

pip install torch==2.1.0 torchvision torchaudio --index-url https://download.pytorch.org/whl/cpu
pip3 install torch-npu==2.1.0.post3
Looking in indexes: https://download.pytorch.org/whl/cpu

其他信息

环境安装好后执行到 pip install InternEvo 时报错。已配置好torch和torch-npu环境。

sunpengsdu commented 3 months ago

@li126com

li126com commented 3 months ago

pip install InternEvo目前仅支持GPU,NPU可以源码安装InternEvo。后续会加一个可选项对rotary_emb 的安装