Skip to content

Update LSTMModel.py #4

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.

Sign up for GitHub

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

Open
stevenhurwitt wants to merge 1 commit into
base: master
Choose a base branch
from
Open

Update LSTMModel.py #4

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
12 changes: 6 additions & 6 deletions code/LSTMModel.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,7 +1,7 @@
# inspired by https://github.com/hunkim/word-rnn-tensorflow

import tensorflow as tf
from tensorflow.python.ops import rnn_cell, seq2seq
#from tensorflow.python.ops import rnn_cell, seq2seq
import numpy as np

import constants as c
Expand Down Expand Up @@ -45,8 +45,8 @@ def build_graph(self, test):
# LSTM Cells
##

lstm_cell = rnn_cell.BasicLSTMCell(self.cell_size)
self.cell = rnn_cell.MultiRNNCell([lstm_cell] * self.num_layers)
lstm_cell = tf.contrib.rnn.BasicLSTMCell(self.cell_size)
self.cell = tf.contrib.rnn.MultiRNNCell([lstm_cell] * self.num_layers)

##
# Data
Expand All @@ -71,7 +71,7 @@ def build_graph(self, test):
# The split splits this tensor into a seq_len long list of 3D tensors of shape
# [batch_size, 1, rnn_size]. The squeeze removes the 1 dimension from the 1st axis
# of each tensor
inputs_split = tf.split(1, self.seq_len, input_embeddings)
inputs_split = tf.split(input_embeddings, self.seq_len, 1)
inputs_split = [tf.squeeze(input_, [1]) for input_ in inputs_split]


Expand Down Expand Up @@ -100,7 +100,7 @@ def loop(prev, _):
self.cell,
loop_function=loop if test else None,
scope='lstm_vars')
lstm_outputs = tf.reshape(tf.concat(1, lstm_outputs_split), [-1, self.cell_size])
lstm_outputs = tf.reshape(tf.concat(lstm_outputs_split, 1), [-1, self.cell_size])

# outputs looks like this:
# [
Expand Down Expand Up @@ -138,7 +138,7 @@ def loop(prev, _):
# Train
##

total_loss = seq2seq.sequence_loss_by_example([logits],
total_loss = tf.contrib.legacy_seq2seq.sequence_loss_by_example([logits],
[tf.reshape(self.targets, [-1])],
[tf.ones([self.batch_size * self.seq_len])],
self.vocab_size)
Expand Down