lstm-for-epf TypeError : Tensors in list passed to 'values' of 'Concat' Op have types [float64, float32] that don't all match

[ad34]

Hi,

I m trying to run the script lstm-for-epf.py with tensorflow 0.10

the execution abort due to the following error

File ".../tensorflow/TensorFlow-Tutorials-for-Time-Series-master/lstm_predictor.py", line 117, in lstm_model
output, layers = tf.nn.rnn(stacked_lstm, x, dtype=dtypes.float32)
File "/Library/Python/2.7/site-packages/tensorflow/python/ops/rnn.py", line 219, in rnn
(output, state) = call_cell()
File "/Library/Python/2.7/site-packages/tensorflow/python/ops/rnn.py", line 206, in
call_cell = lambda: cell(input_, state)
File "/Library/Python/2.7/site-packages/tensorflow/python/ops/rnn_cell.py", line 812, in call
cur_inp, new_state = cell(cur_inp, cur_state)
File "/Library/Python/2.7/site-packages/tensorflow/python/ops/rnn_cell.py", line 310, in call
concat = _linear([inputs, h], 4 * self._num_units, True)
File "/Library/Python/2.7/site-packages/tensorflow/python/ops/rnn_cell.py", line 906, in _linear
res = math_ops.matmul(array_ops.concat(1, args), matrix)
File "/Library/Python/2.7/site-packages/tensorflow/python/ops/array_ops.py", line 697, in concat
name=name)
File "/Library/Python/2.7/site-packages/tensorflow/python/ops/gen_array_ops.py", line 414, in _concat
values=values, name=name)
File "/Library/Python/2.7/site-packages/tensorflow/python/framework/op_def_library.py", line 437, in apply_op
raise TypeError("%s that don't all match." % prefix)
TypeError: Tensors in list passed to 'values' of 'Concat' Op have types [float64, float32] that don't all match.

there is a float64/32 conversion issue.

adding this in _lstm_model fix the issue.

X = tf.cast(X,tf.float32)

there is probably another better fix, I just started playing with the code :)

[arnoutaertgeerts]

I think this is a good solution. TensorFlow works best with the float32 type but the np.sin function creates a float64 type array. This causes an error in TensorFlow >0.9.

[ad34]

@arnoutaertgeerts after making the change I still have errors. as the project dont seems to be active anymore I will probably try to port it to tensorflow >0.9 on my side.

[tgjeon]

Thanks for the interesting on this github.
Now, I'm working on changing this repository into Keras.

[ad34]

@tgjeon nice !

Hello - was this issue ever fixed? I have the same problem :/ I would like to use this code.

[fayazahasan]

@ashleymaeconard I am working on time series predictions, facing similar issues as yours. Did you find a solution??

[0b01]

@ashleymaeconard @fayazahasan

The original poster added the fix in his post.

Add this line...

X = tf.cast(X,tf.float32)

to...

TensorFlow-Tutorials-for-Time-Series/lstm_predictor.py

Line 114 in 4672e50

def _lstm_model(X, y):

[0b01]

I encountered this error from using google/seq2seq. This answer also applies.