Welcome to mlgw_bns’s documentation!

The purpose of the mlgw_bns package is to speed up the generation of gravitational waveforms for binary neutron star mergers, by training a machine learning model on effective-one-body waveforms.

The code can be found on the github page.

Contact

• Contact

Explanation

• Higher order modes
  • Spherical harmonics
  • The time-domain 22 wave
  • Frequency-domain waves

Usage guides

• Installation
  • From pip
  • From the repo
  • Testing and dev functionality
• Overview
  • Making a new model
• Profiling the model evaluation
• Validating the waveforms’ accuracy
• Optimizing hyperparameters

Package reference

• Model for waveform generation
  • Model
  • ParametersWithExtrinsic
  • compute_polarizations()
• Management of a training and validation dataset
  • WaveformParameters
  • ParameterSet
  • Dataset
  • ParameterGenerator
  • WaveformGenerator
• Multibanding
  • seglen_from_freq()
  • reduced_frequency_array()
  • high_frequency_grid()
  • low_frequency_grid()
• Downsampling and interpolation
  • DownsamplingTraining
  • GreedyDownsamplingTraining
• Data management
  • SavableData
  • Residuals
  • DownsamplingIndices
  • FDWaveforms
  • PrincipalComponentData
  • phase_unwrapping()
• Principal Component Analysis
  • PrincipalComponentTraining
  • PrincipalComponentAnalysisModel
• Neural Network
  • Hyperparameters
  • NeuralNetwork
  • SklearnNetwork
  • retrieve_best_trials_list()
  • best_trial_under_n()
• Hyperparameter Optimization
  • HyperparameterOptimization
• Model Validation
  • ValidateModel
• Fixed Dataset Training
  • IndexedWaveformParameters
  • FixedParameterGenerator
  • FixedWaveformGenerator
  • make_fixed_generation_pair()

Indices and tables

• Index

• Module Index