Welcome to TensorWaves!

TensorWaves originates from pycompwa, which did its computations through a function tree that is similar to TensorFlow graphs and SymPy expression trees. The decision to use Python only has been further inspired by TensorFlowAnalysis, AmpliTF and zfit.

TensorWaves is an optimizer package that can ‘fit’ mathematical models (expressed with sympy) to data distributions using a variety of computational back-ends, optimizers, and estimator functions. In addition, the tensorwaves.data module helps one to generate data distributions for those mathematical expressions.

The package is developed in parallel with AmpForm, which implements physics models for amplitude analysis, but its mechanisms for creating computational backend functions can in principle be used independently.

General examples

Amplitude analysis

• Installation
  • Quick installation
  • Editable installation
• General examples
  • Overview
    • Optimize parameters
    • Computational backends
    • Generate and transform data
  • Advanced examples
    • Core ideas illustrated
    • Chi-squared estimator
    • Binned fit
    • Unbinned fit
    • Constant sub-expressions
    • Speed up lambdifying
• Amplitude analysis
  • Step 1: Formulate model
    • 1.1 Generate transitions
    • 1.2 Build amplitude model
  • Step 2: Generate data
    • 2.1 Generate phase space sample
    • 2.2 Generate intensity-based sample
    • 2.3 Visualize kinematic variables
    • 2.4 Export data sets
  • Step 3: Perform fit
    • 3.1 Prepare parametrized function
    • 3.2 Define estimator
    • 3.3 Optimize fit parameters
    • 3.4 Export and import
    • 3.5 Analyze fit result
  • Advanced examples
    • Analytic continuation