Hub: tf2-preview: ValueError: Arguments and signature arguments do not match: 4 3

[HanGuo97]

Code to reproduce the error (copied from https://tfhub.dev/google/tf2-preview/nnlm-en-dim128/1):

import tensorflow_hub as hub

embed = hub.load("https://tfhub.dev/google/tf2-preview/nnlm-en-dim128/1")
embeddings = embed(["cat is on the mat", "dog is in the fog"])

Re-run the last line again, or wrap the list input with tf.constant, seems to solve the issue.

I'm using Alpha preview of TF 2.0.

[arnoegw]

Problem confirmed.

[andresusanopinto]

This has been fixed in March 5th in https://github.com/tensorflow/tensorflow/commit/46ebf0267ad8af92bc7a7331b96ec742e00d296f#diff-fd05cc53efd6ecee89775b1fc1f0a4bb

The fix will be in the next 2.0 release, but for now you have to install from nightly builds.

[andresusanopinto]

E.g.

!pip install tf-nightly-2.0-preview

Instead of the "pip install tensorflow==2.0.0a0"

[HanGuo97]

Thanks for the response. So for the current version, are the results obtained from wrapping the input with tf.constant workaround correct? I'm totally fine with an extra line of code if the results are correct.

[arnoegw]

Yes, I think passing a tf.constant will work just fine

[cottrell]

I this actually fixed? I am hitting same errors but with 100 and 102.

    396       raise ValueError(
    397           "Arguments and signature arguments do not match: %s %s " %
--> 398           (len(args), len(list(self.signature.input_arg))))
    399
    400     function_call_options = ctx.function_call_options

ValueError: Arguments and signature arguments do not match: 100 102

Am on latest nightly 2.0.

[davidemariani]

Hello, I am having the same issue using tensorflow 2.0.0-alpha0 , getting the following error from a jupyter notebook:

---

ValueError Traceback (most recent call last)
in
2 tr = tf.constant(examples)
3 BATCH_SIZE = 32
----> 4 model.fit(tr, epochs=5, steps_per_epoch=math.ceil(len(train)/BATCH_SIZE))

~\Anaconda3\envs\thesis_tensorflow\lib\site-packages\tensorflow\python\keras\engine\training.py in fit(self, x, y, batch_size, epochs, verbose, callbacks, validation_split, validation_data, shuffle, class_weight, sample_weight, initial_epoch, steps_per_epoch, validation_steps, validation_freq, max_queue_size, workers, use_multiprocessing, **kwargs)
871 validation_steps=validation_steps,
872 validation_freq=validation_freq,
--> 873 steps_name='steps_per_epoch')
874
875 def evaluate(self,

~\Anaconda3\envs\thesis_tensorflow\lib\site-packages\tensorflow\python\keras\engine\training_arrays.py in model_iteration(model, inputs, targets, sample_weights, batch_size, epochs, verbose, callbacks, val_inputs, val_targets, val_sample_weights, shuffle, initial_epoch, steps_per_epoch, validation_steps, validation_freq, mode, validation_in_fit, prepared_feed_values_from_dataset, steps_name, **kwargs)
261 # ins can be callable in DistributionStrategy + eager case.
262 actual_inputs = ins() if callable(ins) else ins
--> 263 batch_outs = f(actual_inputs)
264 except errors.OutOfRangeError:
265 if is_dataset:

~\Anaconda3\envs\thesis_tensorflow\lib\site-packages\tensorflow\python\keras\backend.py in __call__(self, inputs)
3215 value = math_ops.cast(value, tensor.dtype)
3216 converted_inputs.append(value)
-> 3217 outputs = self._graph_fn(*converted_inputs)
3218 return nest.pack_sequence_as(self._outputs_structure,
3219 [x.numpy() for x in outputs])

~\Anaconda3\envs\thesis_tensorflow\lib\site-packages\tensorflow\python\eager\function.py in __call__(self, args, *kwargs)
556 raise TypeError("Keyword arguments {} unknown. Expected {}.".format(
557 list(kwargs.keys()), list(self._arg_keywords)))
--> 558 return self._call_flat(args)
559
560 def _filtered_call(self, args, kwargs):

~\Anaconda3\envs\thesis_tensorflow\lib\site-packages\tensorflow\python\eager\function.py in _call_flat(self, args)
625 # Only need to override the gradient in graph mode and when we have outputs.
626 if context.executing_eagerly() or not self.outputs:
--> 627 outputs = self._inference_function.call(ctx, args)
628 else:
629 self._register_gradient()

~\Anaconda3\envs\thesis_tensorflow\lib\site-packages\tensorflow\python\eager\function.py in call(self, ctx, args)
395 raise ValueError(
396 "Arguments and signature arguments do not match: %s %s " %
--> 397 (len(args), len(list(self.signature.input_arg))))
398
399 function_call_options = ctx.get_function_call_options()

ValueError: Arguments and signature arguments do not match: 27 29

[gijswobben]

I'm still seeing this issue with the latest version.. Any updates?

[vbardiovskyg]

This is no longer an issue with the nightly build:
pip install tf-nightly-2.0-preview

It won't start working in alpha until we get a new release.

[ahmedbr]

@vbardiovskyg I'm having same issue with tensorflow-gpu: '2.1.0'. Any solutions ?
ValueError: Arguments and signature arguments do not match. got: 37, expected: 39

[vbardiovskyg]

Hi Ahmed, sorry but I could not reproduce this. I tried tensorflow-gpu==2.1.0 and tensorflow-gpu==2.2.0 and tensorflow==2.2.0.

Are you sure you are importing the correct tensorflow? A typical issue is that there are both tensorflow and tensorflow-gpu installed, and when doing import tensorflow, it is not clear which got imported.

[ahmedbr]

@vbardiovskyg thanks for replying. I only use: tensorflow-gpu: 2.1.0 and keras 2.3.1 for building MLP model. (using import keras instead of import tensorflow.keras )

This is what I'm trying to do:

model.fit_generator(generator=batch_generator(X_train, y_train, divide_on, True), epochs=1,
                                steps_per_epoch=np_steps,
                                verbose=1)

def batch_generator(X, y, batch_size, shuffle):
    number_of_batches = round(X.shape[0] / batch_size)
    print(f'y.shape: {X.shape}')
    print(f'X.shape[0]: {X.shape[0]}')
    print(f'batch_size: {batch_size}')
    print(f'number_of_batches: {number_of_batches}')
    counter = 0
    sample_index = np.arange(X.shape[0])
    if shuffle:
        np.random.shuffle(sample_index)
    while True:
        batch_index = sample_index[batch_size * counter:batch_size * (counter + 1)]
        print(f'counter: {counter}')
        print(f'batch_index: {batch_index}')
        X_batch = X[batch_index, :].toarray()
        y_batch = y[batch_index]
        counter += 1
        print(f'yielded X_batch shape: {X_batch.shape} where counter: {counter}')
        print(f'yielded y_batch shape: {y_batch.shape} where counter: {counter}')
        yield X_batch, y_batch
        if counter == number_of_batches:
            if shuffle:
                np.random.shuffle(sample_index)
            counter = 0

While the result is:

shape of x_train: (1407577, 207)
shape of x_valid: (74084, 207)
Epoch 1/1
y.shape: (1407577, 207)
X.shape[0]: 1407577
batch_size: 64
number_of_batches: 21993
counter: 0
batch_index: [  55578  885043  826530 1085983  596906  990879  150577  916018 1283648
  657662   22295  978107 1330422 1094582 1323311 1386819 1349709  361209
 1314015  112607 1363984  179845  464715  474078  337839 1097602 1173681
  910575 1113526   27510  647144 1081263  678677  567365  149537  936813
  202732  727397   21943  563727  923775  661254  465769  383644   99391
  918760 1152295   23548  511531  661704  590537  855781  654484 1008951
  690456  414501  998788   97076  334435   57625  822775  125039  857771
 1103476]
yielded X_batch shape: (64, 207) where counter: 1
yielded y_batch shape: (64, 1) where counter: 1
counter: 1
batch_index: [ 784817  248893  472646 1010253 1400872  522094  217448 1120737  844047
  837030 1298761  768817  728090  799502  951408 1356634  365863  997856
  752368  558098  732175  571739 1048726  536754 1315847 1032336  413975
  366115  550960  976257 1116407  237934  386406 1160910  113013  163694
  437774  372660 1164205  335589  223032 1364559  626345 1135905  396354
 1120329  567954 1252556  763024  351094  445536  734038 1344405  239830
  869035  728995  288278  238975  699285 1267507  131834   24448  233283
  623641]
Traceback (most recent call last):
  File "mlp.py", line 443, in <module>
yielded X_batch shape: (64, 207) where counter: 2
yielded y_batch shape: (64, 1) where counter: 2
    main()
  File "mlp.py", line 371, in main
    model1 = run_model1(X_train, y_train, X_valid, y_valid)
  File "mlp.py", line 255, in run_model1
    verbose=1)
  File "/path_to_project/venv/lib/python3.6/site-packages/keras/legacy/interfaces.py", line 91, in wrapper
    return func(*args, **kwargs)
  File "/path_to_project/venv/lib/python3.6/site-packages/keras/engine/training.py", line 1732, in fit_generator
    initial_epoch=initial_epoch)
  File "/path_to_project/venv/lib/python3.6/site-packages/keras/engine/training_generator.py", line 220, in fit_generator
    reset_metrics=False)
  File "/path_to_project/venv/lib/python3.6/site-packages/keras/engine/training.py", line 1514, in train_on_batch
    outputs = self.train_function(ins)
  File "/path_to_project/venv/lib/python3.6/site-packages/tensorflow_core/python/keras/backend.py", line 3727, in __call__
    outputs = self._graph_fn(*converted_inputs)
  File "/path_to_project/venv/lib/python3.6/site-packages/tensorflow_core/python/eager/function.py", line 1551, in __call__
    return self._call_impl(args, kwargs)
  File "/path_to_project/venv/lib/python3.6/site-packages/tensorflow_core/python/eager/function.py", line 1591, in _call_impl
    return self._call_flat(args, self.captured_inputs, cancellation_manager)
  File "/path_to_project/venv/lib/python3.6/site-packages/tensorflow_core/python/eager/function.py", line 1692, in _call_flat
    ctx, args, cancellation_manager=cancellation_manager))
  File "/path_to_project/venv/lib/python3.6/site-packages/tensorflow_core/python/eager/function.py", line 527, in call
    (len(args), len(list(self.signature.input_arg))))
ValueError: Arguments and signature arguments do not match. got: 37, expected: 39 

It fails in the second iteration of the loop.

[ahmedbr]

@vbardiovskyg thanks for replying. I only use: tensorflow-gpu: 2.1.0 and keras 2.3.1 for building MLP model. (using import keras instead of import tensorflow.keras )

This is what I'm trying to do:

model.fit_generator(generator=batch_generator(X_train, y_train, divide_on, True), epochs=1,
                                steps_per_epoch=np_steps,
                                verbose=1)

def batch_generator(X, y, batch_size, shuffle):
    number_of_batches = round(X.shape[0] / batch_size)
    print(f'y.shape: {X.shape}')
    print(f'X.shape[0]: {X.shape[0]}')
    print(f'batch_size: {batch_size}')
    print(f'number_of_batches: {number_of_batches}')
    counter = 0
    sample_index = np.arange(X.shape[0])
    if shuffle:
        np.random.shuffle(sample_index)
    while True:
        batch_index = sample_index[batch_size * counter:batch_size * (counter + 1)]
        print(f'counter: {counter}')
        print(f'batch_index: {batch_index}')
        X_batch = X[batch_index, :].toarray()
        y_batch = y[batch_index]
        counter += 1
        print(f'yielded X_batch shape: {X_batch.shape} where counter: {counter}')
        print(f'yielded y_batch shape: {y_batch.shape} where counter: {counter}')
        yield X_batch, y_batch
        if counter == number_of_batches:
            if shuffle:
                np.random.shuffle(sample_index)
            counter = 0

While the result is:

shape of x_train: (1407577, 207)
shape of x_valid: (74084, 207)
Epoch 1/1
y.shape: (1407577, 207)
X.shape[0]: 1407577
batch_size: 64
number_of_batches: 21993
counter: 0
batch_index: [  55578  885043  826530 1085983  596906  990879  150577  916018 1283648
  657662   22295  978107 1330422 1094582 1323311 1386819 1349709  361209
 1314015  112607 1363984  179845  464715  474078  337839 1097602 1173681
  910575 1113526   27510  647144 1081263  678677  567365  149537  936813
  202732  727397   21943  563727  923775  661254  465769  383644   99391
  918760 1152295   23548  511531  661704  590537  855781  654484 1008951
  690456  414501  998788   97076  334435   57625  822775  125039  857771
 1103476]
yielded X_batch shape: (64, 207) where counter: 1
yielded y_batch shape: (64, 1) where counter: 1
counter: 1
batch_index: [ 784817  248893  472646 1010253 1400872  522094  217448 1120737  844047
  837030 1298761  768817  728090  799502  951408 1356634  365863  997856
  752368  558098  732175  571739 1048726  536754 1315847 1032336  413975
  366115  550960  976257 1116407  237934  386406 1160910  113013  163694
  437774  372660 1164205  335589  223032 1364559  626345 1135905  396354
 1120329  567954 1252556  763024  351094  445536  734038 1344405  239830
  869035  728995  288278  238975  699285 1267507  131834   24448  233283
  623641]
Traceback (most recent call last):
  File "mlp.py", line 443, in <module>
yielded X_batch shape: (64, 207) where counter: 2
yielded y_batch shape: (64, 1) where counter: 2
    main()
  File "mlp.py", line 371, in main
    model1 = run_model1(X_train, y_train, X_valid, y_valid)
  File "mlp.py", line 255, in run_model1
    verbose=1)
  File "/path_to_project/venv/lib/python3.6/site-packages/keras/legacy/interfaces.py", line 91, in wrapper
    return func(*args, **kwargs)
  File "/path_to_project/venv/lib/python3.6/site-packages/keras/engine/training.py", line 1732, in fit_generator
    initial_epoch=initial_epoch)
  File "/path_to_project/venv/lib/python3.6/site-packages/keras/engine/training_generator.py", line 220, in fit_generator
    reset_metrics=False)
  File "/path_to_project/venv/lib/python3.6/site-packages/keras/engine/training.py", line 1514, in train_on_batch
    outputs = self.train_function(ins)
  File "/path_to_project/venv/lib/python3.6/site-packages/tensorflow_core/python/keras/backend.py", line 3727, in __call__
    outputs = self._graph_fn(*converted_inputs)
  File "/path_to_project/venv/lib/python3.6/site-packages/tensorflow_core/python/eager/function.py", line 1551, in __call__
    return self._call_impl(args, kwargs)
  File "/path_to_project/venv/lib/python3.6/site-packages/tensorflow_core/python/eager/function.py", line 1591, in _call_impl
    return self._call_flat(args, self.captured_inputs, cancellation_manager)
  File "/path_to_project/venv/lib/python3.6/site-packages/tensorflow_core/python/eager/function.py", line 1692, in _call_flat
    ctx, args, cancellation_manager=cancellation_manager))
  File "/path_to_project/venv/lib/python3.6/site-packages/tensorflow_core/python/eager/function.py", line 527, in call
    (len(args), len(list(self.signature.input_arg))))
ValueError: Arguments and signature arguments do not match. got: 37, expected: 39 

It fails in the second iteration of the loop.

The problem solved after replacing import keras by import tensorflow.keras and making the necessary modifications