Deep Learning Toolbox Converter for TensorFlow Models

Import and export TensorFlow models within MATLAB

The converter for TensorFlow models enables you to import a pretrained TensorFlow model and weights and export a MATLAB dlnetwork as a TensorFlow model.

Import:

To import a TensorFlow network in MATLAB, please refer to importNetworkFromTensorFlow.

Note: importNetworkFromTensorFlow accepts TensorFlow models in the SavedModel format. Use the following commands before import to save a TensorFlow model in the SavedModel format.

For TensorFlow versions 2.15 and earlier, save the model with the following Python command: :

model.save("modelName")

For TensorFlow versions 2.16 and later, save the model with the following Python commands.

First, install the Keras 2 (tf_keras) package:

pip install tf_keras

Note: if you are using a Jupyter notebook, restart the notebook kernel before running the code.

Now, precede any import tensorflow statement with the following lines of code:

import os
os.environ["TF_USE_LEGACY_KERAS"] = "1"
import tensorflow

Next, Import Keras 2 (tf_keras) as keras. Do not import Keras 3 (keras) directly. Keras 3 specific features are unsupported:

import tf_keras as keras
from tf_keras import layers

To verify if you are using Keras 2, the following command should print `False`

is_keras3 = hasattr(keras, "version")) 
print(is_keras3) 

Finally, once you verify the Keras version, save the model:

model.save("modelName")

The TensorFlow Model Importer has been tested up to 2.15. SavedModels saved with versions 2.16 and higher may also work without issues. When using tf_keras for versions 2.16 and higher, Keras 3 specific features such as Keras Ops will not be supported.

Export:

To export a TensorFlow dlnetwork from MATLAB, please refer to exportNetworkToTensorFlow.

Export supports:

• TensorFlow v2.0 or later, no support for Keras 3.0
• Python version 3.0 or later