In Tensorflow, all the functions of the decoder-encoder ( https://github.com/tensorflow/tensorflow/blob/master/tensorflow/python/ops/seq2seq.py ) use a unidirectional implementation of the RNN block.
How can we implement a bi-directional encoder ( http://arxiv.org/abs/1409.0473 or similar systems) in Tensorflow so that it can be both immediately and backward in the setup of the encoder-decoder?
No one has answered this question yet.
See related questions:
Source: https://habr.com/ru/post/1649578/More articles:Search for bash file inside PyQt4 GUI - bashCan I run a downloadable book on ec2 only with ssh keyfile no aws key / pair? - amazon-ec2Can I set the default value when unpacking? - pythonEmbedding JavaScript in Math.pow - javascriptCalculating the time between text field interaction - pythonDoes Kafka support ELB in front of a cluster of brokers? - apache-kafkaUnused variables and '_' - swiftnode.js noobie is trying to follow the tutorial - need to change jade link to pug - jsonformat link value in attribute android - androidфункция для возврата копии np.array с замененными элементами - pythonAll Articles