[Lesson 6] Saving and Loading Models

- Very deep and complex models can take days or even weeks to train. we can save our models after we finish training them.

 

- Tensorflow provides a direct way to take my saved models and deplay them on different platform.

 

 

- TensorFlow makes this all possible through its saved model format.

 

- A SavedModel contains all the information about the model including the weight values, the model's architecture and even the optimizers configuration.

 

- Since the SavedModel contains all the information about the model, it does not require the original code to run it.

 

- This means you can execute it directly in your application or servers using one of the deployment libraries such as TensorFlow serving for models to be served over HTTP.

 

<Colab Notebook>

- To access the Colab Notebook, login to your Google account and click on the link below:

Saving and Loading Models

Google Colaboratory

 

<소스 코드>

github.com/HoYoungChun/TensorFlow_study/blob/master/10_Saving%20and%20Loading%20Models.py

HoYoungChun/TensorFlow_study