module 'keras.backend' has no attribute 'tensorflow_backend'

[Jagadeesh Sparrow]

I tried to run the following code, from an old script, after updating keras and tensorflow

print(K.tensorflow_backend._get_available_gpus())

Keras - 2.4.3

Tensorflow - 2.2.0

and I faced the following error : 

AttributeError: module 'keras.backend' has no attribute 'tensorflow_backend'

Is there any depreciation / update with the newer versions?

[Paige Bailey]

You should be able to use tf.config.list_physical_devices to get the number of visible GPUs.

https://www.tensorflow.org/api_docs/python/tf/config/list_physical_devices

--
You received this message because you are subscribed to the Google Groups "Keras-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to keras-users...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/keras-users/5987c9f4-0e2e-49e3-b27d-e3a4cfc8d9cco%40googlegroups.com.

--

• • • •

Paige Bailey    Product Manager (TensorFlow) @DynamicWebPaige webp...@google.com

[JB Data31]

My env is a classic virtual 3 procs VM.

This snippet

import tensorflow as tf
print(tf.config.list_physical_devices('GPU'))
print(tf.config.list_physical_devices('CPU'))

outputs

[]
[PhysicalDevice(name='/physical_device:CPU:0', device_type='CPU')]

When I execute TF computes all 3 procs warm (see screenshots).

According to my device list there's 1 proc. Is there a config entry that shows such a multi proc env ?

JB

To view this discussion on the web, visit https://groups.google.com/d/msgid/keras-users/CACMT%2BuJhp1tf8XBG0hOJ_b%3DVAu76vzseHW-zo62g9jDORk1WUg%40mail.gmail.com.