论伯特的句子嵌入

Question

我试图对从预先训练的BERT模型中获得的张量（句子嵌入）进行k-均值聚类。在

from sklearn.cluster import KMeans

embedding = BERTembeddingGenerator.generateSentenceEmbedding(sentence)
embeddingMapping[embedding] = sentence
sentenceEmbeddingsList = list(embeddingMapping.keys())
model = KMeans(n_clusters=10, init='k-means++', max_iter=100, n_init=1)
labels = model.fit_predict(sentenceEmbeddingsList)

这会引发错误

^{pr2}$

句子嵌入的大小为768

我试着把张量转换成numpy数组。有没有任何关于我错在哪里的提示