python ncf_main.py AttributeError: 'float' object has no attribute '_unconditional_loss'

[chongfaqin]

D:\anaconda3\python.exe D:/widedeep/dnnrecommend/recommendation/ncf_main.py
D:\anaconda3\lib\site-packages\h5py_init_.py:36: FutureWarning: Conversion of the second argument of issubdtype from float to np.floating is deprecated. In future, it will be treated as np.float64 == np.dtype(float).type.
from ._conv import register_converters as _register_converters
2018-06-15 17:12:47.470001: I T:\src\github\tensorflow\tensorflow\core\platform\cpu_feature_guard.cc:140] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2
I0615 17:12:47.473455 16368 tf_logging.py:116] Data preprocessing...
I0615 17:12:52.540076 16368 tf_logging.py:116] Creating Estimator from Keras model...
Traceback (most recent call last):
File "D:/widedeep/dnnrecommend/recommendation/ncf_main.py", line 335, in
absl_app.run(main)
File "D:\anaconda3\lib\site-packages\absl\app.py", line 274, in run
_run_main(main, argv)
File "D:\anaconda3\lib\site-packages\absl\app.py", line 238, in _run_main
sys.exit(main(argv))
File "D:/widedeep/dnnrecommend/recommendation/ncf_main.py", line 172, in main
run_ncf(FLAGS)
File "D:/widedeep/dnnrecommend/recommendation/ncf_main.py", line 194, in run_ncf
layers, FLAGS.batch_size, FLAGS.mf_regularization,mlp_regularization)
File "D:\widedeep\dnnrecommend\recommendation\official\recommendation\neumf_model.py", line 105, in init
mf_user_latent = tf.keras.layers.Flatten()(mf_embedding_user(user_input))
File "D:\anaconda3\lib\site-packages\tensorflow\python\keras_impl\keras\engine\base_layer.py", line 239, in call
output = super(Layer, self).call(inputs, **kwargs)
File "D:\anaconda3\lib\site-packages\tensorflow\python\layers\base.py", line 696, in call
self.build(input_shapes)
File "D:\anaconda3\lib\site-packages\tensorflow\python\keras_impl\keras\engine\base_layer.py", line 499, in wrapper
output_shape = fn(instance, input_shape)
File "D:\anaconda3\lib\site-packages\tensorflow\python\keras_impl\keras\layers\embeddings.py", line 122, in build
dtype=self.dtype)
File "D:\anaconda3\lib\site-packages\tensorflow\python\keras_impl\keras\engine\base_layer.py", line 201, in add_weight
trainable=trainable)
File "D:\anaconda3\lib\site-packages\tensorflow\python\layers\base.py", line 576, in add_variable
self.add_loss(regularization)
File "D:\anaconda3\lib\site-packages\tensorflow\python\layers\base.py", line 373, in add_loss
loss._unconditional_loss = True # pylint: disable=protected-access
AttributeError: 'float' object has no attribute '_unconditional_loss'

Process finished with exit code 1

[tensorflowbutler]

Thank you for your post. We noticed you have not filled out the following field in the issue template. Could you update them if they are relevant in your case, or leave them as N/A? Thanks.
What is the top-level directory of the model you are using
Have I written custom code
OS Platform and Distribution
TensorFlow installed from
TensorFlow version
Bazel version
CUDA/cuDNN version
GPU model and memory
Exact command to reproduce

[chongfaqin]

I am use pycharm，my dirctory is like imape.

OS Platform is win10
Tensorflow 1.7.0
anaconda 3
memory
python ncf_main.py
Thanks.

[robieta]

This is because your tensorflow version is too old; official models are only guaranteed for the tf-nightly. Unfortunately this also occurs in tf 1.8 (the latest anaconda tf build, which Google does not control) so you'll either need to build from source or use pip (since we do maintain up to date tf-nightly and tf-nightly-gpu packages).

[chongfaqin]

@robieta thanks，but tensorflow 1.8 still wrong.

[robieta]

Yes, that's what I was saying. 1.8 is still too old.

[chongfaqin]

Thanks ,1.9 is ok.