utils.benchmark.known_integrand_benchmarks module

class KnownIntegrandBenchmarker[source]

Bases: utils.benchmark.benchmarker.Benchmarker

Benchmark by comparing to a known integrand with an exact integral value

benchmark_method(d, integrand, integrator_config=None, integrand_params=None, n_batch=100000, keep_history=False, device=device(type='cpu'))[source]

Integrate an known integrand and compare with the theoretical result

Parameters

  • d (int) – number of dimensions

  • integrand (constructor for utils.integrands.abstract.KnownIntegrand) – integrand class to be tested. Expects a constuctor for that class, i.e. a callable that returns an instance.

  • integrator_config (dictwrapper.NestedMapping, None) – configuration to be passed to Integrator. If None, use the default.

  • n_batch (int) – batch size used for the benchmarking (after training)

  • integrand_params (dict) – dictionary of parameters provided to integrand through integrand(**integrand_params).

  • device (torch.device) – torch device on which to train and run the Integrator.

class KnownIntegrandGridBenchmarker(n_repeat=1)[source]

Bases: utils.benchmark.benchmarker.GridBenchmarker, utils.benchmark.known_integrand_benchmarks.KnownIntegrandBenchmarker

Benchmark against a known integrand by sampling configurations from a grid

Parameters

n (int) – Optional: How often the grid is sampled.

class KnownIntegrandRandomHPBenchmarker(n_samples=5, n_repeat=1)[source]

Bases: utils.benchmark.benchmarker.RandomHyperparameterBenchmarker, utils.benchmark.known_integrand_benchmarks.KnownIntegrandBenchmarker

Benchmark against a known integrand by sampling integrator hyperparameters randomly

Parameters

  • n_samples (int) – Number of random integrator configurations to draw

  • n_repeat (int) – Optional: How often the grid is sampled.