New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Seq2Seq(Attention)Input Shape Question #31
Comments
I would also like to ask if this is attention can be used for LSTM timing prediction. The main problem that bothers me is this n_class = len (word_dict). Can this be considered as a different feature of input? In the end, your project is very good. It has benefited a lot. It's very good. Thank you very much. |
'S i want a beer', 'i want a beer E',Is this the last attention mechanism not aligned? |
I have the same question. Do you solve it ? Oh, I have solve it.You are right.However, batch_size=1,so it has no effect。 |
i think the code is only for batch)size=1,isn't is ? |
Seq2Seq(Attention)\Seq2Seq(Attention)-Tensor.py
The shape of the input should be [max_time, batch_size,...]. The input = tf. transpose (dec_inputs, [1, 0, 2]) has already been transformed. In tf. expand_dims (inputs [i], 1), the expansion is indeed one dimension. It seems that there should be zero dimension expansion here. Although the final shape is correct, whether it is intentional or not is here. What about a little trick?
The text was updated successfully, but these errors were encountered: