提示pytorch不是cuda版本

Question

运行embedding_cls = RagEmbedding()的时候提示如下问题，检查了下torch不是cuda版本

Answer 1

阿基米口 2025-03-25 11:30:31

 RagEmbedding 应该是 device 选的cpu， 请提供下完整错误截图，和  RagEmbedding的调用

• 提问者 慕仙3296778 #1

  RagEmbedding代码：
  from langchain.embeddings.huggingface import HuggingFaceEmbeddings
  class RagEmbedding(object):
      def __init__(self, model_path="../RAG MODEL/BGE-M3/", 
                   device="cuda"):
          self.embedding = HuggingFaceEmbeddings(model_name=model_path,
                                                 model_kwargs={"device": "cuda"})
      def get_embedding_fun(self):
          return self.embedding

• 阿基米口 回复 提问者 慕仙3296778 #2

  HuggingFaceEmbeddings(model_name=model_path,
                                                 model_kwargs={"device": "cuda"}) If you don't have a GPU, you can change it to CPU HuggingFaceEmbeddings(model_name=model_path,
                                                 model_kwargs={"device": "cuda"})  如果是没有gpu的环境，这里可以改成cpu    {"device": "cpu"}