Tried to convert 'y' to a tensor and failed. Error: None values not supported.

[BrikerMan]

System information

• Have I written custom code (as opposed to using a stock example script provided in TensorFlow): Yes
• OS Platform and Distribution (e.g., Linux Ubuntu 16.04): Colab
• TensorFlow installed from (source or binary): binary
• TensorFlow version (use command below): 2.0.0rc1

Describe the current behavior

Crash when fit. It works just fine in 1.14.0, 2.0.0b0 & 2.0.0b1.

Describe the expected behavior

Not to crash when fit.

Code to reproduce the issue

import numpy as np
import tensorflow as tf
from tensorflow.python.keras.layers import Dense
from tensorflow.python.keras.models import Sequential
from tensorflow.python.keras.optimizers import Adam

model = Sequential([
    Dense(24, input_dim=4, activation='relu'),
    Dense(24, activation='relu'),
    Dense(2, activation='linear')
])
model.compile(loss='mse', optimizer=Adam(lr=0.001))

x = np.array([[-0.08623559, -0.79897248,  0.03606475,  1.09068178]])
y = np.array([[ 1.0449973,  -0.14471795]])
model.fit(x, y)

Here is the colab link

Other info / logs

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
/usr/local/lib/python3.6/dist-packages/tensorflow_core/python/framework/op_def_library.py in _apply_op_helper(self, op_type_name, name, **keywords)
    526                 as_ref=input_arg.is_ref,
--> 527                 preferred_dtype=default_dtype)
    528           except TypeError as err:

22 frames
ValueError: None values not supported.

During handling of the above exception, another exception occurred:

ValueError                                Traceback (most recent call last)
ValueError: None values not supported.

During handling of the above exception, another exception occurred:

ValueError                                Traceback (most recent call last)
/usr/local/lib/python3.6/dist-packages/tensorflow_core/python/framework/op_def_library.py in _apply_op_helper(self, op_type_name, name, **keywords)
    543               raise ValueError(
    544                   "Tried to convert '%s' to a tensor and failed. Error: %s" %
--> 545                   (input_name, err))
    546             prefix = ("Input '%s' of '%s' Op has type %s that does not match" %
    547                       (input_name, op_type_name, observed))

ValueError: Tried to convert 'y' to a tensor and failed. Error: None values not supported.

[gadagashwini-zz]

It works fine with TF 1.15.0.rc0.
I could reproduce the issue with Tensorflow 2.0.0.rc1. Please see the colab gist. Thanks!

[bermeitinger-b]

This also happens in 2.0.0.rc0.
A "workaround" is not using the class for the optimizer but using the name as string, of course, then you can't adjust the learning rate.

Instead of:
model.compile(optimizer=keras.optimizers.SGD(), loss='mse')
you would write:
model.compile(optimizer='sgd', loss='mse')

[ymodak]

Changing your imports can help resolve this issue. Tested in '2.0.0-rc1'

import numpy as np
import tensorflow as tf
from tensorflow.keras.layers import Dense
from tensorflow.keras.models import Sequential
from tensorflow.keras.optimizers import Adam

model = Sequential([
    Dense(24, input_dim=4, activation='relu'),
    Dense(24, activation='relu'),
    Dense(2, activation='linear')
])
model.compile(loss='mse', optimizer=Adam(lr=0.001))

x = np.array([[-0.08623559, -0.79897248,  0.03606475,  1.09068178]])
y = np.array([[ 1.0449973,  -0.14471795]])
model.fit(x, y)

output:

Train on 1 samples
1/1 [==============================] - 0s 239ms/sample - loss: 0.5793
<tensorflow.python.keras.callbacks.History at 0x7fa54c4f6630>

[tensorflow-bot]

Are you satisfied with the resolution of your issue?
Yes
No

[Times125]

I have this problem ,too. my code works well in 2.0b1, not work in tf 2.0rc1

[fengwang]

Changing your imports can help resolve this issue. Tested in '2.0.0-rc1'

what is the difference between tensorflow.python.keras and tensorflow.keras? Why this trick works?

[LiutongZhou]

from tensorflow.keras.optimizers import Adam
in place of
from tensorflow.python.keras.optimizers import Adam
helps

[dsryu0822]

from tensorflow.keras.optimizers import Adam
in place of
from tensorflow.python.keras.optimizers import Adam
helps

It works for me

[jawadSajid]

Worked for me!

from tensorflow.keras.optimizers import Adam
in place of
from tensorflow.python.keras.optimizers import Adam
helps