Keras: Latest version with lot's of deprecations

Created on 11 Jun 2019 · 7Comments · Source: keras-team/keras

W0611 16:49:22.343930 139742361401152 deprecation_wrapper.py:119] From anaconda3/envs/keras-tf-py36/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:517: The name tf.placeholder is deprecated. Please use tf.compat.v1.placeholder instead.

W0611 16:49:22.354705 139742361401152 deprecation_wrapper.py:119] From anaconda3/envs/keras-tf-py36/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:74: The name tf.get_default_graph is deprecated. Please use tf.compat.v1.get_default_graph instead.

W0611 16:49:22.356116 139742361401152 deprecation_wrapper.py:119] From anaconda3/envs/keras-tf-py36/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:4138: The name tf.random_uniform is deprecated. Please use tf.random.uniform instead.

W0611 16:49:22.377483 139742361401152 deprecation_wrapper.py:119] From anaconda3/envs/keras-tf-py36/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:3976: The name tf.nn.max_pool is deprecated. Please use tf.nn.max_pool2d instead.

W0611 16:49:22.378576 139742361401152 deprecation_wrapper.py:119] From anaconda3/envs/keras-tf-py36/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:133: The name tf.placeholder_with_default is deprecated. Please use tf.compat.v1.placeholder_with_default instead.

W0611 16:49:22.384115 139742361401152 deprecation.py:506] From anaconda3/envs/keras-tf-py36/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:3445: calling dropout (from tensorflow.python.ops.nn_ops) with keep_prob is deprecated and will be removed in a future version.
Instructions for updating:
Please use rate instead of keep_prob. Rate should be set to rate = 1 - keep_prob.
W0611 16:49:23.080871 139742361401152 deprecation_wrapper.py:119] From ./train.py:289: The name tf.trainable_variables is deprecated. Please use tf.compat.v1.trainable_variables instead.

Total number of trainable parameters: 1498733
W0611 16:49:23.093107 139742361401152 deprecation_wrapper.py:119] From anaconda3/envs/keras-tf-py36/lib/python3.6/site-packages/keras/optimizers.py:790: The name tf.train.Optimizer is deprecated. Please use tf.compat.v1.train.Optimizer instead.

tensorflow support

Source

christopher5106

Most helpful comment

I met with the same issue also when running

from keras import models
from keras import layers

I am using python=3.7.2, keras=2.2.4 and tensorflow=1.14.0.
After I modified the code to

from tensorflow.keras import models
from tensorflow.keras import layers

The warnings disappeared.

caozd999 on 17 Oct 2019

👍2 ❤1 🎉1

All 7 comments

@christopher5106 As part of performance improvements, removing similar API calls and also to rearrange APIs, it is required to deprecate some of them. But, the log clearly shows how to replace those deprecated APIs with new APIs.

Please fill the issue template here. Could you update them if they are relevant in your case, or leave them as N/A? Along with the template, please provide as many details as possible to find the root cause of the issue. Thanks!

jvishnuvardhan on 12 Jun 2019

Closing due to lack of recent activity. Please update the issue when new information becomes available, and we will reopen the issue. Thanks!

jvishnuvardhan on 22 Jun 2019

Can this please be opened again? I'm seeing these warnings with Keras==2.2.4 and tensorflow-gpu==1.14.0.

As far as I understand, there shouldn't be any more information needed. It's a matter of updating keras to conform to the new tensorflow way of doing things, right? Most of the warning messages give suggested replacements.

andyDoucette on 8 Aug 2019

What I see from these messages, is that Keras 2.2.4 is not compatible with latest versions of Tensorflow.

It has never been very compliant with Theano and CNTK also. To my point of view, it breaks its promise.

christopher5106 on 9 Aug 2019

There are two versions of Keras, tf.keras and keras.io