microsoft / DeepSpeed

DeepSpeed is a deep learning optimization library that makes distributed training and inference easy, efficient, and effective.
https://www.deepspeed.ai/
Apache License 2.0
33.6k stars 3.94k forks source link

[REQUEST]win10 install fail build_win.bat #5698

Open shark-xiake opened 3 days ago

shark-xiake commented 3 days ago

when I run build_win.bat, I get a deepspeed whl file finnally. The matter seems to have been resolved. However, when I ran the program, the following issue occurred

File "D:\anaconda3\envs\llm\lib\site-packages\sat__init__.py", line 6, in from .training.deepspeed_training import training_main File "D:\anaconda3\envs\llm\lib\site-packages\sat\training\deepspeed_training.py", line 27, in import deepspeed File "D:\anaconda3\envs\llm\lib\site-packages\deepspeed__init.py", line 25, in from . import ops File "D:\anaconda3\envs\llm\lib\site-packages\deepspeed\ops\init.py", line 11, in from . import transformer File "D:\anaconda3\envs\llm\lib\site-packages\deepspeed\ops\transformer\init.py", line 8, in from ...model_implementations.transformers.ds_transformer import DeepSpeedTransformerInference File "D:\anaconda3\envs\llm\lib\site-packages\deepspeed\model_implementations\transformers\ds_transformer.py", line 17, in from deepspeed.ops.transformer.inference.triton.mlp import TritonMLP File "D:\anaconda3\envs\llm\lib\site-packages\deepspeed\ops\transformer\inference\init__.py", line 7, in from ....model_implementations.transformers.ds_transformer import DeepSpeedTransformerInference ImportError: cannot import name 'DeepSpeedTransformerInference' from partially initialized module 'deepspeed.model_implementations.transformers.ds_transformer' (most likely due to a circular import) (D:\anaconda3\envs\llm\lib\site-packages\deepspeed\model_implementations\transformers\ds_transformer.py)

loadams commented 2 days ago

@shark-xiake can you please share the program you are running as a repro case?

shark-xiake commented 2 days ago

I tested this project (https://github.com/lzw-lzw/RemoteGLM). Everything is normal on the Linux system