tensor – Tensor operations in PyTensor

PyTensor’s strength is in expressing symbolic calculations involving tensors.

PyTensor tries to emulate the numpy interface as much as possible in the tensor module. This means that once TensorVariables are created, it should be possibly to define symbolic expressions using calls that look just like numpy calls, such as pt.exp(x).transpose(0, 1)[:, None]

• Basic Tensor Functionality
• random – Random number functionality
• tensor.utils – Tensor Utils
• tensor.elemwise – Tensor Elemwise
• tensor.extra_ops – Tensor Extra Ops
• tensor.io – Tensor IO Ops
• tensor.slinalg – Linear Algebra Ops Using Scipy
• tensor.nlinalg – Linear Algebra Ops Using Numpy
• tensor.fft – Fast Fourier Transforms
• tensor.conv – Tensor Convolutions
• tensor.rewriting.math – Tensor Rewrites for Math Operations
• tensor.rewriting.basic – Tensor Rewrites
• vectorize()
• tensor.optimize – Symbolic Optimization Routines