hui qin

减小学习率看看。

> > > @rogerwwww I have already used conda to set enviroment which is include opencv-nopython, cython numpy, opencv-mat and instaboost.But when i try to import instaboost, i met this...

> > > > > @rogerwwww I have already used conda to set enviroment which is include opencv-nopython, cython numpy, opencv-mat and instaboost.But when i try to import instaboost, i...

In normal condition, we really can not change the recurrent_activation with 'elu'. But I tested your SRU, it worked in 'elu'. And I got a better performance than 'sigmoid'. Besides,...

I tested MinimalRNNCell in imdb dataset. It was faster than SRU and got a better accuracy score. : ( ---------------------------------------------------------------------------------------------------------- MinimalRNNCell 5s - loss: 0.6160 - acc: 0.7472 - val_loss:...

That is ok. I will wait for you. Besides, I found a funny things. I used two selu activations in SRU. My codes likes these: rnn_layer1 = SRU(16,dropout=0.,recurrent_dropout=0.,activation='selu',implementation=0,\ unit_forget_bias=True, unroll=True,recurrent_activation='selu')...

When I changed the batch_size into 16, then SRU got 0.80 score. But the SRU program became very slow. My log is here: Epoch 1/5 64s - loss: 0.5681 -...

抱歉回答晚了。sru和gru用法基本上相似。sru确实快，但是评价指标比gru差多了，分数掉了一截。如果你能忍受掉分，还是可以用用的。

Thanks for your reply. So we should prepare the premise_mask and hyp_mask datas in tokenizer ,then put them into dataset. Is that right?

Thanks. Supposed I have the following datas: passage_text='Woman: Has Tom moved to the downtown? Man: No. He is still living in the country.' question_text='Where does Tom live? ' answer_text='In the...