Tutorial

>>> import theano.tensor as T

• Python tutorial
• • Matrix conventions for machine learning

• Baby Steps - Algebra
  • Adding two Matrices
• More Examples
  • Computing More than one Thing at the Same Time
  • Using Shared Variables
  • Using Random Numbers
    • Seeding Streams
    • Copying Random State Between Theano Graphs
    • Other Implementations
• Derivatives in Theano
  • Computing the Jacobian
  • Jacobian times a Vector
    • L-operator
  • Final Pointers
• • IfElse vs Switch
• • Scan
• How Shape Information is Handled by Theano
  • Specifying Exact Shape
• Broadcasting

• • Compressed Sparse Format
  • Handling Sparse in Theano
    • Properties and Construction
• Using the GPU
  • • Testing Theano with GPU
    • What Can be Accelerated on the GPU
    • GPU Async Capabilities
    • Exercise
  • Learning to Program with PyCUDA
    • Exercise
• Using multiple GPUs
  • A simple graph on two GPUs
• Convolution arithmetic tutorial
  • Refresher: discrete convolutions
  • Convolution arithmetic
    • Zero padding, unit strides
    • • Half (same) padding
    • No zero padding, non-unit strides
  • Transposed convolution arithmetic
    • Transposed convolution
    • Zero padding, unit strides, transposed
    • • Half (same) padding, transposed
    • No zero padding, non-unit strides, transposed
  • Miscellaneous convolutions
    • Grouped Convolutions
  • Quick reference

• • Configuration
  • Mode
  • Optimizers
• Printing/Drawing Theano graphs
  • Picture Printing of Graphs
• Debugging Theano: FAQ and Troubleshooting
  • Interpreting Error Messages
  • “How do I Print an Intermediate Value in a Function?”
  • “The Function I Compiled is Too Slow, what’s up?”
  • “How do I Step through a Compiled Function?”
  • Dumping a Function to help debug
• Dealing with NaNs
  • Run in NanGuardMode, DebugMode, or MonitorMode
  • Algorithm Related
  • NaN Introduced by AllocEmpty

• Graph Structures
  • Graph Structures
    • Op
    • Variable
  • Graph Structures Extension
  • Optimizations
• • The Basics of Pickling
  • Robust Serialization
• Understanding Memory Aliasing for Speed and Correctness
  • Borrowing when Creating Shared Variables
  • • Retrieving
  • Borrowing when Constructing Function Objects
• • Basic Objects
  • Pickle
• • Convolution and Pooling
  • Parallel element wise ops with OpenMP