ValueError: Layer question_attn_gru_1 expects 2 inputs, but it received 7 input tensors

[anand-ar]

Hi,

I used your trained model and ran predict script - I get the the above errror

ValueError: Layer question_attn_gru_1 expects 2 inputs, but it received 7 input tensors. Input received: [<tf.Tensor 'uP/cond/Merge:0' shape=(?, ?, 90) dtype=float32>, <tf.Tensor 'uQ/cond/Merge:0' shape=(?, ?, 90) dtype=float32>, <tf.Tensor 'mul:0' shape=(90, 45) dtype=float32>, <tf.Tensor 'mul_1:0' shape=(45, 45) dtype=float32>, <tf.Tensor 'mul_2:0' shape=(90, 45) dtype=float32>, <tf.Tensor 'mul_3:0' shape=(45, 1) dtype=float32>, <tf.Tensor 'mul_4:0' shape=(180, 180) dtype=float32>]

what needs to be done

Thanks

[mahnerak]

We have a problem with Keras > 2.0.8.
Consider using Keras 2.0.6 untill we fix the bug.

[anand-ar]

switching to 2.06 avoided that issue - now this error has come up
Using TensorFlow backend.
Preparing model...Traceback (most recent call last):
File "predict.py", line 33, in
model = load_model(args.model, custom_objects())
File "/Users/arunachalam/anaconda/lib/python2.7/site-packages/Keras-2.0.6-py2.7.egg/keras/models.py", line 233, in load_model
File "/Users/arunachalam/anaconda/lib/python2.7/site-packages/Keras-2.0.6-py2.7.egg/keras/models.py", line 307, in model_from_config
File "/Users/arunachalam/anaconda/lib/python2.7/site-packages/Keras-2.0.6-py2.7.egg/keras/layers/init.py", line 54, in deserialize
File "/Users/arunachalam/anaconda/lib/python2.7/site-packages/Keras-2.0.6-py2.7.egg/keras/utils/generic_utils.py", line 139, in deserialize_keras_object
File "/Users/arunachalam/anaconda/lib/python2.7/site-packages/Keras-2.0.6-py2.7.egg/keras/engine/topology.py", line 2450, in from_config
File "/Users/arunachalam/anaconda/lib/python2.7/site-packages/Keras-2.0.6-py2.7.egg/keras/engine/topology.py", line 2447, in process_layer
File "/Users/arunachalam/anaconda/lib/python2.7/site-packages/Keras-2.0.6-py2.7.egg/keras/engine/topology.py", line 571, in call
File "/Users/arunachalam/rnet/R-NET-in-Keras/layers/QuestionPooling.py", line 26, in build
assert(isinstance(input_shape, list) and len(input_shape) == 5)

still using your model

Thanks

[MartinXPN]

That error occurred because QuestionPooling layer was updated in later commits to be consistent with R-Net paper. However the provided model was built with previous implementation. To avoid this error and use the model without errors run the following command
git checkout 715221945416f312c253d88bf9ca2c1debef61d3
and then run the prediction script

[anand-ar]

Ok, prediction script worked with patch on the complete dataset

Then I wanted to just check one topic - super_bowl_50 - cut the input to just have super_bowl_50 passage and questions - parsed and preprocessed - looks like some keras compatibility issue or your implementation requires complete data set

prediction script had this error

Preparing model...Done!
Loading data...Done!
Running predicting model...Traceback (most recent call last):
File "predict.py", line 54, in
verbose=1)
File "/Users/arunachalam/anaconda/lib/python2.7/site-packages/Keras-2.0.6-py2.7.egg/keras/legacy/interfaces.py", line 87, in wrapper
File "/Users/arunachalam/anaconda/lib/python2.7/site-packages/Keras-2.0.6-py2.7.egg/keras/engine/training.py", line 2085, in predict_generator
File "/Users/arunachalam/anaconda/lib/python2.7/site-packages/Keras-2.0.6-py2.7.egg/keras/engine/training.py", line 1622, in predict_on_batch
File "/Users/arunachalam/anaconda/lib/python2.7/site-packages/Keras-2.0.6-py2.7.egg/keras/engine/training.py", line 78, in _standardize_input_data
ValueError: Error when checking model : the list of Numpy arrays that you are passing to your model is not the size the model expected. Expected to see 4 arrays but instead got the following list of 2 arrays: [array([[[-0.41758001, -0.16617 , -0.19927999, ..., -0.155 ,
0.50695002, 0.068692 ],
[-0.26901001, 0.63352001, 0.36471999, ..., 0.43628001,
0.53119999, -0.303649...

[MartinXPN]

Have you included char-level-embeddings while parsing? As far as I understand the model that you are trying to run requires them (vectors + embeddings => 4 arrays)

[shwetgarg]

I used your trained model and ran predict script after following 2 steps:

1. installed Keras 2.0.6
2. git checkout 7152219

But I am getting following error:

File "predict.py", line 54, in
verbose=1)
File "/dev/shm/shweta/software/anaconda2/lib/python2.7/site-packages/keras/legacy/interfaces.py", line 87, in wrapper
return func(*args, **kwargs)
File "/dev/shm/shweta/software/anaconda2/lib/python2.7/site-packages/keras/engine/training.py", line 2085, in predict_generator
outs = self.predict_on_batch(x)
File "/dev/shm/shweta/software/anaconda2/lib/python2.7/site-packages/keras/engine/training.py", line 1622, in predict_on_batch
self._feed_input_shapes)
File "/dev/shm/shweta/software/anaconda2/lib/python2.7/site-packages/keras/engine/training.py", line 140, in _standardize_input_data
str(array.shape))
ValueError: Error when checking : expected Q_vecs to have shape (None, None, 300) but got array with shape (100, 30, 25)

What needs to be done.

Thanks

[TarunTater]

@shwetgarg - had the same set of errors. Were you able to solve the last error? Thanks

[soni-sarvesh]

@shwetgarg I am also getting the same error. Could you resolve the issue?

[blurLake]

I have similar error with Keras 2.3.1.

[edmondja]

I have similar error with Keras 2.3.1.

+1

[sarahosini]

Hi
I ran a code for my dataset that I was able to learn several times. Now I want to learn it again. It gives the following error if I have not changed the code.

/usr/local/lib/python3.6/dist-packages/tensorflow/python/keras/engine/training.py:805 train_function *
return step_function(self, iterator)
/usr/local/lib/python3.6/dist-packages/tensorflow/python/keras/engine/training.py:795 step_function **
outputs = model.distribute_strategy.run(run_step, args=(data,))
/usr/local/lib/python3.6/dist-packages/tensorflow/python/distribute/distribute_lib.py:1259 run
return self._extended.call_for_each_replica(fn, args=args, kwargs=kwargs)
/usr/local/lib/python3.6/dist-packages/tensorflow/python/distribute/distribute_lib.py:2730 call_for_each_replica
return self._call_for_each_replica(fn, args, kwargs)
/usr/local/lib/python3.6/dist-packages/tensorflow/python/distribute/distribute_lib.py:3417 _call_for_each_replica
return fn(*args, **kwargs)
/usr/local/lib/python3.6/dist-packages/tensorflow/python/keras/engine/training.py:788 run_step **
outputs = model.train_step(data)
/usr/local/lib/python3.6/dist-packages/tensorflow/python/keras/engine/training.py:754 train_step
y_pred = self(x, training=True)
/usr/local/lib/python3.6/dist-packages/tensorflow/python/keras/engine/base_layer.py:998 call
input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)
/usr/local/lib/python3.6/dist-packages/tensorflow/python/keras/engine/input_spec.py:207 assert_input_compatibility
' input tensors. Inputs received: ' + str(inputs))

ValueError: Layer model expects 1 input(s), but it received 2 input tensors. Inputs received: [<tf.Tensor 'IteratorGetNext:0' shape=(None, None, None, None) dtype=uint8>, <tf.Tensor 'IteratorGetNext:1' shape=(None, None, None) dtype=float32>]

Can you guide me?

[JosephTeja]

Hi
I ran a code for my dataset that I was able to learn several times. Now I want to learn it again. It gives the following error if I have not changed the code.

/usr/local/lib/python3.6/dist-packages/tensorflow/python/keras/engine/training.py:805 train_function *
return step_function(self, iterator)
/usr/local/lib/python3.6/dist-packages/tensorflow/python/keras/engine/training.py:795 step_function **
outputs = model.distribute_strategy.run(run_step, args=(data,))
/usr/local/lib/python3.6/dist-packages/tensorflow/python/distribute/distribute_lib.py:1259 run
return self._extended.call_for_each_replica(fn, args=args, kwargs=kwargs)
/usr/local/lib/python3.6/dist-packages/tensorflow/python/distribute/distribute_lib.py:2730 call_for_each_replica
return self._call_for_each_replica(fn, args, kwargs)
/usr/local/lib/python3.6/dist-packages/tensorflow/python/distribute/distribute_lib.py:3417 _call_for_each_replica
return fn(*args, **kwargs)
/usr/local/lib/python3.6/dist-packages/tensorflow/python/keras/engine/training.py:788 run_step **
outputs = model.train_step(data)
/usr/local/lib/python3.6/dist-packages/tensorflow/python/keras/engine/training.py:754 train_step
y_pred = self(x, training=True)
/usr/local/lib/python3.6/dist-packages/tensorflow/python/keras/engine/base_layer.py:998 call
input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)
/usr/local/lib/python3.6/dist-packages/tensorflow/python/keras/engine/input_spec.py:207 assert_input_compatibility
' input tensors. Inputs received: ' + str(inputs))

ValueError: Layer model expects 1 input(s), but it received 2 input tensors. Inputs received: [<tf.Tensor 'IteratorGetNext:0' shape=(None, None, None, None) dtype=uint8>, <tf.Tensor 'IteratorGetNext:1' shape=(None, None, None) dtype=float32>]

Can you guide me?

@sarahosini i'm havinf the same problem. did you get it resolved?

[itskarthicklakshmanan]

Hi
I ran a code for my dataset that I was able to learn several times. Now I want to learn it again. It gives the following error if I have not changed the code.

/usr/local/lib/python3.6/dist-packages/tensorflow/python/keras/engine/training.py:805 train_function *
return step_function(self, iterator)
/usr/local/lib/python3.6/dist-packages/tensorflow/python/keras/engine/training.py:795 step_function **
outputs = model.distribute_strategy.run(run_step, args=(data,))
/usr/local/lib/python3.6/dist-packages/tensorflow/python/distribute/distribute_lib.py:1259 run
return self._extended.call_for_each_replica(fn, args=args, kwargs=kwargs)
/usr/local/lib/python3.6/dist-packages/tensorflow/python/distribute/distribute_lib.py:2730 call_for_each_replica
return self._call_for_each_replica(fn, args, kwargs)
/usr/local/lib/python3.6/dist-packages/tensorflow/python/distribute/distribute_lib.py:3417 _call_for_each_replica
return fn(*args, **kwargs)
/usr/local/lib/python3.6/dist-packages/tensorflow/python/keras/engine/training.py:788 run_step **
outputs = model.train_step(data)
/usr/local/lib/python3.6/dist-packages/tensorflow/python/keras/engine/training.py:754 train_step
y_pred = self(x, training=True)
/usr/local/lib/python3.6/dist-packages/tensorflow/python/keras/engine/base_layer.py:998 call
input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)
/usr/local/lib/python3.6/dist-packages/tensorflow/python/keras/engine/input_spec.py:207 assert_input_compatibility
' input tensors. Inputs received: ' + str(inputs))

ValueError: Layer model expects 1 input(s), but it received 2 input tensors. Inputs received: [<tf.Tensor 'IteratorGetNext:0' shape=(None, None, None, None) dtype=uint8>, <tf.Tensor 'IteratorGetNext:1' shape=(None, None, None) dtype=float32>]

Can you guide me?

@sarahosini, I am facing the same issue! any solution found for this error!

[smitshah]

Hi
I ran a code for my dataset that I was able to learn several times. Now I want to learn it again. It gives the following error if I have not changed the code.
/usr/local/lib/python3.6/dist-packages/tensorflow/python/keras/engine/training.py:805 train_function *
return step_function(self, iterator)
/usr/local/lib/python3.6/dist-packages/tensorflow/python/keras/engine/training.py:795 step_function **
outputs = model.distribute_strategy.run(run_step, args=(data,))
/usr/local/lib/python3.6/dist-packages/tensorflow/python/distribute/distribute_lib.py:1259 run
return self._extended.call_for_each_replica(fn, args=args, kwargs=kwargs)
/usr/local/lib/python3.6/dist-packages/tensorflow/python/distribute/distribute_lib.py:2730 call_for_each_replica
return self._call_for_each_replica(fn, args, kwargs)
/usr/local/lib/python3.6/dist-packages/tensorflow/python/distribute/distribute_lib.py:3417 _call_for_each_replica
return fn(*args, **kwargs)
/usr/local/lib/python3.6/dist-packages/tensorflow/python/keras/engine/training.py:788 run_step **
outputs = model.train_step(data)
/usr/local/lib/python3.6/dist-packages/tensorflow/python/keras/engine/training.py:754 train_step
y_pred = self(x, training=True)
/usr/local/lib/python3.6/dist-packages/tensorflow/python/keras/engine/base_layer.py:998 call
input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)
/usr/local/lib/python3.6/dist-packages/tensorflow/python/keras/engine/input_spec.py:207 assert_input_compatibility
' input tensors. Inputs received: ' + str(inputs))

ValueError: Layer model expects 1 input(s), but it received 2 input tensors. Inputs received: [<tf.Tensor 'IteratorGetNext:0' shape=(None, None, None, None) dtype=uint8>, <tf.Tensor 'IteratorGetNext:1' shape=(None, None, None) dtype=float32>]

Can you guide me?

@sarahosini, I am facing the same issue! any solution found for this error!

@sarahosini, I am also getting the same error. Has any solution been found for this!