lenetOnMnistExportImportToJSONWithAdamOptimizerState

This examples demonstrates model, model weights, and optimizer weights export and import back:

• Model is exported in Keras-style JSON format; weights are exported in custom (txt) format.

• Model is trained on Mnist dataset.

• It saves all the data to the project root directory.

• The first Sequential model is created via JSON configuration, weights, and optimizer state loading.

• After loading model is trained again with the same optimizer with frozen Conv2D layers. Only weights in Dense layers can be updated.

• The second Sequential model is created via JSON configuration and weights loading.

• After loading model is trained again with the same optimizer with frozen Conv2D layers. Only weights in Dense layers can be updated.

• Results of two training (with restored optimizer state and without) could be compared via accuracy comparison.