gaotongxiao/latentvariabledialoguemodel

latentvariabledialoguemodel's Issues

Word dropout is not implemented

The paper uses word dropout in the decoder. However, in the implementation, you directly drop the z value. Please follow the paper and implement the word dropout.
Here is an nmt project with word dropout : https://github.com/EdinburghNLP/nematus/blob/master/nematus/nmt.py
Introduction of word-dropout:

1. Sec 3.1 in https://cs.umd.edu/~miyyer/pubs/2015_acl_dan.pdf ( corresponding code )
2. Sec 5.3 in https://arxiv.org/pdf/1603.00375.pdf

A related paper for your reference

Here is a very similar paper that generates sentences with latent variable.
https://arxiv.org/pdf/1511.06349.pdf
Code[tf]: https://github.com/Chung-I/Variational-Recurrent-Autoencoder-Tensorflow
Code[pytorch]: https://github.com/kefirski/pytorch_RVAE

Could not open model\model: Unknown: NewRandomAccessFile failed to Create/Open: model\model : �ܾ����ʡ� ; Input/output error

when I start to run main.py to train, I got this problem:
2019-04-10 15:18:43.911854: W T:\src\github\tensorflow\tensorflow\core\util\tensor_slice_reader.cc:95] Could not open model\model: Unknown: NewRandomAccessFile failed to Create/Open: model\model : �ܾ����ʡ� ; Input/output error 2019-04-10 15:18:43.922905: W T:\src\github\tensorflow\tensorflow\core\util\tensor_slice_reader.cc:95] Could not open model\model: Unknown: NewRandomAccessFile failed to Create/Open: model\model : �ܾ����ʡ� ; Input/output error 2019-04-10 15:18:43.924309: W T:\src\github\tensorflow\tensorflow\core\framework\op_kernel.cc:1318] OP_REQUIRES failed at save_restore_tensor.cc:170 : Data loss: Unable to open table file model\model: Unknown: NewRandomAccessFile failed to Create/Open: model\model : �ܾ����ʡ� ; Input/output error