GPU llama3.2运行环境问题

问题 1

from transformers import pipeline
  File "<frozen importlib._bootstrap>", line 1075, in _handle_fromlist
  File "/torch/venv3/pytorch/lib/python3.10/site-packages/transformers/utils/import_utils.py", line 1754, in __getattr__
    module = self._get_module(self._class_to_module[name])
  File "/torch/venv3/pytorch/lib/python3.10/site-packages/transformers/utils/import_utils.py", line 1766, in _get_module
    raise RuntimeError(

RuntimeError: Failed to import transformers.pipelines because of the following error (look up to see its traceback):
Failed to import transformers.generation.utils because of the following error (look up to see its traceback):

该问题主要是accelerate和transformer版本不匹配导致

推荐使用一下版本运行

ubuntu22.04的系统环境accelerate 1.2.1

软件名	版本
accelerate	1.2.1
torch	2.4.0+cu124
torchaudio	2.4.0+cu124
torchvision	0.19.0+cu124
transformer_engine          transformer_engine_cu12      transformers            transformers-stream-generator	1.13.0 1.13.0 4.46.3 0.0.5
torch	2.2.2
torchvision	0.15.0a0

ubuntu20.04系统环境 accelerate 1.0.1 transformer-engine 0.6.0 transformers 4.46.3

软件名	版本
accelerate	1.0.1
transformer-engine	0.6.0
transformers	4.46.3
transformer-engine	0.6.0

问题2

ImportError: cannot import name 'MllamaForConditionalGeneration' from 'transformers' (/torch/venv3/pytorch/lib/python3.10/site-packages/transformers/__init__.py)

该问题主要是transformers版本太低导致（没有包含llama3.2的多模态代码），升级transformer4.43.0以上版本