Help Center> ModelArts> FAQs> Notebook> Constraints> Is the Keras Engine Supported?

Is the Keras Engine Supported?

Notebook instances in DevEnviron support the Keras engine. The Keras engine is not supported in job training and model deployment (inference).

Keras is an advanced neural network API written in Python. It is capable of running on top of TensorFlow, CNTK, or Theano. Notebook instances in DevEnviron support tf.keras.

How Do I View Keras Versions?

  1. On the ModelArts console, create a notebook instance.

    Select the Multi-Engine 1.0 (Python3, Recommended) work environment to create an instance.

  2. Click the notebook instance name to go to the Jupyter Notebook page. On the right of the page, click New and select TensorFlow-1.8 or TensorFlow-1.13.1 to create a file.
  3. In the notebook instance, run !pip list to view Keras versions.
    Figure 1 Viewing Keras versions