** WARNING ** : outdated materials, this material is pretty much outdated since the release of TF2.0+

TensorFlow Tutorials/Examples

I have more materials here, but will wait until release of TF 2.0 and re-work these tutorials to make sure there are not using any depricated API.

These tutorials are all for TensorFlow 1.8. (See what is new in TF 1.9: TensorFlow 1.9.0 Release)

• Tutorial 1: Basic variables, placeholders, matrix operations and session in TF

• Tutorial 2: Batch gradient decent linear regression in TensorFlow

• Tutorial 3: Logical/Control operations

• Tutorial 4: Logistic regression using TensorFlow

• Tutorial 5: Fully connected neural net "from scratch" (using TF's matrix operations) for binary classification

• Tutorial 6: MNIST image classification using multilayer fully connected NN built using tf.layers API and data processing using tf.data API

• Tutorial 7: More on tf.data.Dataset API and building high performance IO and data processing pipeline

• Tutorial 8: Premade estimators and tf.feature_column API for classification of iris dataset.

• Tutorial 9: Writing custom estimators with tf.estimator API, CSV IO example with dataset API and tf.summary for tensorboard.

• Tutorial 10: Embedding layer to vectorize sequence of indicies (text vectorization) using tf.nn.embedding_lookup.

---

Extras

These are mainly some confusing concepts in TF/Keras that I tried to clarify by code snippets.

• XTutorial 1: Demonstrates that TimeDistributed(Dense(n)) layer in Keras is identical to applying Dense(n) layer.