Tutorial

Let’s start an interactive session (e.g. python or ipython) and import Theano.

>>> from theano import *

Many of symbols you will need to use are in the tensor subpackage of Theano. Let’s import that subpackage under a handy name. I like T (and many tutorials use this convention).

>>> import theano.tensor as T

If that worked you’re ready for the tutorial, otherwise check your installation (see Installing Theano).

Throughout the tutorial, bear in mind that there is a Glossary of terminology to help you out.

• NumPy refresher
  • Matrix conventions for machine learning
  • Broadcasting
• Baby steps - Adding two numbers together
  • Adding two scalars
  • Adding two matrices
• More examples
  • Logistic function
  • Computing more than one thing at the same time
  • Computing gradients
  • Setting a default value for an argument
  • Using shared variables
  • Using Random Numbers
    • Brief example
    • Seedings Streams
    • Sharing Streams between Functions
• Loading and Saving
  • The basics of pickling
  • Short-term serialization
  • Long-term serialization
• Graph Structures
  • Automatic Differentiation
  • Optimizations
• Using different compiling modes
  • Mode
  • Using DebugMode
  • ProfileMode
    • Creating a ProfileMode Instance
      • Compiling your Graph with ProfileMode
      • Retrieving Timing Information
• Using the GPU
  • Setting up CUDA
  • Making Theano use CUDA
  • Putting it all Together
  • Returning a handle to device-allocated data
  • Running the GPU at Full Speed
  • What can be accelerated on the GPU?
  • Tips for improving performance on GPU
• Some general Remarks
  • Limitations
  • A few tips
• Debugging Theano: FAQ and Troubleshooting
  • How do I print an intermediate value in a Function/Method?
  • How do I print a graph (before or after compilation)?
  • The function I compiled is too slow, what’s up?
  • How do I step through a compiled function with the WrapLinker?
  • How to use pdb ?