Doc2Vec -- "tag '23943' not seen in training corpus/invalid" 错误

 

训练Doc2Vec模型之后,想要索取文档向量,却出现意外报错:

    for idx, docvec in enumerate(model.docvecs):
        if idx < 17600:
            x_train.append(docvec)
        else:
            x_test.append(docvec)

  

错误情况:

KeyError: "tag '23943' not seen in training corpus/invalid"

 

解决方法:

① model.docvecs.doctag_syn0 (即将弃用)

Or

② model.docvecs.vectors_docs

  

posted @ 2020-01-04 19:06  Harp_Yestar  阅读(754)  评论(0编辑  收藏  举报