Loading the keras model developed in python API on jetson environment

hi,
Can anyone suggest how to load the keras model developed in anaconda environment on windows to jetson environment for training the model.
Also please tell me how to define and train the keras model directly on jetson tx2.

Moving to TX2 forum for support coverage.

Hi,

You can install Keras (TensorFlow backend) on Jetson directly.
Check this tutorial for information: https://www.hackster.io/wilson-wang/jetson-tx2-tensorflow-opencv-keras-install-b74e40

To run a model cross platform, you will need to serialize it into a standalone file and de-serialize it on the target platform.
https://keras.io/getting-started/faq/#how-can-i-install-hdf5-or-h5py-to-save-my-models-in-keras

More, it’s not recommended to use Jetson for training.
Please train the model on a desktop environment and use Jetson just for inferencing.

Thanks.