search

dongjun-Lee / text-classification-models-tf

Tensorflow implementations of Text Classification Models.
505 stars 164 forks source link

Word-level CNN model's result is poor in SST-2 dataset. #9

Closed mJackie closed 5 years ago

mJackie commented 5 years ago

Have you test the SST-2 dataset ? I have run word_cnn model on SST-2 dataset . Only get 0.38 accuracy. But the accuray is 0.45 in paper. How to change the model to improve the accuracy? Is something wrong? The word_rnn model can get 0.45 accuracy, why diff is so large?

mJackie commented 5 years ago

change the word embedding size smaller can get higher score i think it is because the sst-2 dataset is so small , when use big embedding size, the representation is poor