Jupyter notebooks to help you started with tensorflow 2.0. The repository includes notebooks to start with simple NN, understand autograph, learn how to train models, CNN to analyze images with tf.data and RNN to generate text!
$> pip install tensorflow
-
If your are new into Tensorflow2.0 or you love to use some old fashion Tensorflow then you might want to start with this notebook: Introduction - Simple Neural Network.ipynb With tensorflow2.0 it is higly recommend to use the tf.keras module if possible. This notebook will give you a first example with the Fashion MNIST Dataset.
-
Now that you are familar with tf.keras you can check the difference between eager mode and grah mode in this notebook: Graph Mode vs Eager Mode.ipynb.
-
Training a model has never been so easy. Let's check how you can do it with this notebook: Train a model.ipynb
-
Learn how to use the Subclassing API to built your model: Subclassing API
-
Learn how to create a custom Layer with low level operations: Create your own layer
-
Learn how to create a CNN model to analayse the QuickDraw dataset. You will use autograph, the model subclassing API and tf.data: Quick Draw CNN.ipynb
-
Learn how to generate text with an RNN. You will use a simple custom keras layer and autograph: RNN - Text Generator.ipynb