Kaggle-ValueError: Input contains NaN, infinity or a value too large for dtype('float32').

在做kaggle的这个word2vec的教程时候遇到这个问题，解决办法是 from sklearn.preprocessing import Imputer trainDataVecs = Imputer().fit_transform(trainDataVecs) 或者给nwords+1以避免出现无限大的情况featureVec = np.divide(featureVec, np.float32(nwords)+1.0)