TensorFlow

1. Tensors

   Tensor shapes, dtypes, indexing, and ragged and sparse tensors.

2. Variables

   Stateful tf.Variable objects, placement, and assignment.

3. Automatic differentiation

   GradientTape, computing gradients, and higher-order derivatives.

4. Graphs and tf.function

   Tracing Python into graphs, performance, and retracing behavior.

5. Better performance with tf.function

   Polymorphism, control flow, and avoiding common tf.function pitfalls.

6. tf.data input pipelines

   Datasets, map, batch, shuffle, and prefetch for input pipelines.

7. tf.data performance

   Prefetching, parallel map, caching, and pipeline optimization.

8. Keras in TensorFlow

   Building and training models with the Keras API inside TensorFlow.

9. SavedModel format

   Saving and loading complete models for serving.

10. Checkpoints

    Saving and restoring variable state during training.

11. Distributed training

    tf.distribute strategies across GPUs, TPUs, and machines.

12. GPU and device placement

    Using GPUs, memory growth, and device placement.

13. Mixed precision

    Float16 and bfloat16 policies for faster training.

14. Profiler

    Analyzing performance bottlenecks with the TensorFlow Profiler.

15. TensorFlow Serving and deployment

    Serving models in production with TF Serving.

Each section links to the official page. Titles and summaries by Extra Credit; the docs belong to their authors.