search

thunlp / TensorFlow-Summarization

MIT License
389 stars 110 forks source link

Cannot work with higher TF version #20

Open bangnguyen89 opened 6 years ago

bangnguyen89 commented 6 years ago

I tried to run your code with higher TF version (1.5) but there are something wrong with these codes: decoder_cell = tf.contrib.seq2seq.DynamicAttentionWrapper( decoder_cell, attention, state_size * 2) wrapper_state = tf.contrib.seq2seq.DynamicAttentionWrapperState( self.init_state, self.prev_att)

DynamicAttentionWrapper and DynamicAttentionWrapperState already deprecated in this TF version 1.5. They replaced by AttentionWrapper and AttentionWrapperState but the input parameters are quiet complexity Could you please help us update your code to compatible with new TF version Thank you so much

Sparsh-Bansal commented 5 years ago

This may be because the pretrained model is trained on version 1.0 and 1.1