QUiNN

List of modules:

  • solvers
    • quinn
      • QUiNNBase
        • QUiNNBase.nens
        • QUiNNBase.nnmodel
        • QUiNNBase.__init__()
        • QUiNNBase.print_params()
        • QUiNNBase.predict_sample()
        • QUiNNBase.predict_ens()
        • QUiNNBase.predict()
        • QUiNNBase.predict_mom_sample()
        • QUiNNBase.predict_plot()
        • QUiNNBase.plot_1d_fits()
    • nn_ens
      • NN_Ens
        • NN_Ens.dfrac
        • NN_Ens.learners
        • NN_Ens.nens
        • NN_Ens.verbose
        • NN_Ens.__init__()
        • NN_Ens.print_params()
        • NN_Ens.fit()
        • NN_Ens.predict_sample()
        • NN_Ens.predict_ens()
        • NN_Ens.predict_ens_fromsamples()
    • nn_mcmc
      • NN_MCMC
        • NN_MCMC.cmode
        • NN_MCMC.lpinfo
        • NN_MCMC.pdim
        • NN_MCMC.samples
        • NN_MCMC.verbose
        • NN_MCMC.__init__()
        • NN_MCMC.logpost()
        • NN_MCMC.logpostgrad()
        • NN_MCMC.fit()
        • NN_MCMC.get_best_model()
        • NN_MCMC.predict_MAP()
        • NN_MCMC.predict_sample()
        • NN_MCMC.predict_ens()
    • nn_vi
      • NN_VI
        • NN_VI.best_model
        • NN_VI.bmodel
        • NN_VI.device
        • NN_VI.trained
        • NN_VI.verbose
        • NN_VI.__init__()
        • NN_VI.fit()
        • NN_VI.predict_sample()
    • nn_laplace
      • NN_Laplace
        • NN_Laplace.cov_mats
        • NN_Laplace.cov_scale
        • NN_Laplace.datanoise
        • NN_Laplace.la_type
        • NN_Laplace.means
        • NN_Laplace.nparams
        • NN_Laplace.priorsigma
        • NN_Laplace.__init__()
        • NN_Laplace.fit()
        • NN_Laplace.la_calc()
        • NN_Laplace.predict_sample()
        • NN_Laplace.predict_ens()
    • nn_swag
      • NN_SWAG
        • NN_SWAG.c
        • NN_SWAG.cov_diags
        • NN_SWAG.cov_type
        • NN_SWAG.d_mats
        • NN_SWAG.datanoise
        • NN_SWAG.k
        • NN_SWAG.lr_swag
        • NN_SWAG.means
        • NN_SWAG.n_steps
        • NN_SWAG.nparams
        • NN_SWAG.priorsigma
        • NN_SWAG.__init__()
        • NN_SWAG.fit()
        • NN_SWAG.swag_calc()
        • NN_SWAG.predict_sample()
        • NN_SWAG.predict_ens()
    • nn_rms
      • NN_RMS
        • NN_RMS.datanoise
        • NN_RMS.nparams
        • NN_RMS.priorsigma
        • NN_RMS.__init__()
        • NN_RMS.fit()
  • nns
    • nns
      • Gaussian
        • Gaussian.__init__()
        • Gaussian.forward()
      • Sine
        • Sine.__init__()
        • Sine.forward()
      • Polynomial
        • Polynomial.order
        • Polynomial.coefs
        • Polynomial.__init__()
        • Polynomial.forward()
      • Polynomial3
        • Polynomial3.a
        • Polynomial3.b
        • Polynomial3.c
        • Polynomial3.d
        • Polynomial3.__init__()
        • Polynomial3.forward()
      • Constant
        • Constant.constant
        • Constant.__init__()
        • Constant.forward()
      • SiLU
        • SiLU.__init__()
        • SiLU.forward()
      • Expon
        • Expon.__init__()
        • Expon.forward()
      • TwoLayerNet
        • TwoLayerNet.linear1
        • TwoLayerNet.linear2
        • TwoLayerNet.cubic
        • TwoLayerNet.__init__()
        • TwoLayerNet.forward()
      • MLP_simple
        • MLP_simple.biasorno
        • MLP_simple.hls
        • MLP_simple.indim
        • MLP_simple.outdim
        • MLP_simple.model
        • MLP_simple.__init__()
        • MLP_simple.forward()
    • nnbase
      • MLPBase
        • MLPBase.best_model
        • MLPBase.device
        • MLPBase.history
        • MLPBase.indim
        • MLPBase.outdim
        • MLPBase.trained
        • MLPBase.__init__()
        • MLPBase.forward()
        • MLPBase.predict()
        • MLPBase.numpar()
        • MLPBase.fit()
        • MLPBase.printParams()
        • MLPBase.printParamNames()
        • MLPBase.predict_plot()
        • MLPBase.plot_1d_fits()
    • nnwrap
      • NNWrap
        • NNWrap.indices
        • NNWrap.nnmodel
        • NNWrap.__init__()
        • NNWrap.reinitialize_instance()
        • NNWrap.__call__()
        • NNWrap.predict()
        • NNWrap.p_flatten()
        • NNWrap.p_unflatten()
        • NNWrap.calc_loss()
        • NNWrap.calc_lossgrad()
        • NNWrap.calc_hess_full()
        • NNWrap.calc_hess_diag()
      • SNet
        • SNet.nnmodel
        • SNet.__init__()
        • SNet.forward()
      • nnwrapper()
      • nn_surrogate()
      • nn_surrogate_multi()
      • nn_p()
    • mlp
      • MLP
        • MLP.hls
        • MLP.biasorno
        • MLP.bnorm
        • MLP.bnlearn
        • MLP.dropout
        • MLP.final_transform
        • MLP.nlayers
        • MLP.nnmodel
        • MLP.__init__()
        • MLP.forward()
    • rnet
      • RNet
        • RNet.activ
        • RNet.bias_post
        • RNet.bias_pre
        • RNet.biasorno
        • RNet.final_layer
        • RNet.indim
        • RNet.init_factor
        • RNet.layer_post
        • RNet.layer_pre
        • RNet.mlp
        • RNet.nlayers
        • RNet.outdim
        • RNet.rdim
        • RNet.step_size
        • RNet.sum_dim
        • RNet.weight_post
        • RNet.weight_pre
        • RNet.wp_function
        • RNet.__init__()
        • RNet.forward()
      • LayerFcn
        • LayerFcn.npar
        • LayerFcn.__init__()
        • LayerFcn.__call__()
      • Const
        • Const.npar
        • Const.__init__()
        • Const.__call__()
      • Lin
        • Lin.npar
        • Lin.__init__()
        • Lin.__call__()
      • Quad
        • Quad.npar
        • Quad.__init__()
        • Quad.__call__()
      • Cubic
        • Cubic.npar
        • Cubic.__init__()
        • Cubic.__call__()
      • Poly
        • Poly.npar
        • Poly.__init__()
        • Poly.__call__()
      • NonPar
        • NonPar.npar
        • NonPar.__init__()
        • NonPar.__call__()
    • losses
      • LogLoss
        • LogLoss.forward()
      • PeriodicLoss
        • PeriodicLoss.model
        • PeriodicLoss.lam
        • PeriodicLoss.bdry1
        • PeriodicLoss.bdry2
        • PeriodicLoss.__init__()
        • PeriodicLoss.forward()
      • GradLoss
        • GradLoss.lam
        • GradLoss.nnmodel
        • GradLoss.__init__()
        • GradLoss.forward()
      • NegLogPost
        • NegLogPost.nnmodel
        • NegLogPost.priorparams
        • NegLogPost.sigma
        • NegLogPost.fulldatasize
        • NegLogPost.pi
        • NegLogPost.__init__()
        • NegLogPost.forward()
      • NegLogPrior
        • NegLogPrior.anchor
        • NegLogPrior.sigma
        • NegLogPrior.pi
        • NegLogPrior.__init__()
        • NegLogPrior.forward()
      • CustomLoss
        • CustomLoss.model
        • CustomLoss.lam1
        • CustomLoss.lam2
        • CustomLoss.__init__()
        • CustomLoss.forward()
    • tchutils
      • tch()
      • npy()
      • print_nnparams()
      • flatten_params()
      • recover_flattened()
  • mcmc
    • mcmc
      • MCMCBase
        • MCMCBase.logPost
        • MCMCBase.logPostGrad
        • MCMCBase.postInfo
        • MCMCBase.__init__()
        • MCMCBase.setLogPost()
        • MCMCBase.run()
        • MCMCBase.sampler()
    • admcmc
      • AMCMC
        • AMCMC.cov_ini
        • AMCMC.gamma
        • AMCMC._propcov
        • AMCMC.t0
        • AMCMC.tadapt
        • AMCMC.__init__()
        • AMCMC.sampler()
    • hmc
      • HMC
        • HMC.epsilon
        • HMC.L
        • HMC.__init__()
        • HMC.sampler()
    • mala
      • MALA
        • MALA.epsilon
        • MALA.__init__()
        • MALA.sampler()
  • utils
    • maps
      • scale01ToDom()
      • scaleDomTo01()
      • scaleTo01()
      • standardize()
      • XMap
        • XMap.__init__()
        • XMap.forw()
        • XMap.inv()
      • Expon
        • Expon.inv()
      • Logar
        • Logar.inv()
      • ComposeMap
        • ComposeMap.__init__()
        • ComposeMap.inv()
      • LinearScaler
        • LinearScaler.__init__()
        • LinearScaler.inv()
      • Standardizer
        • Standardizer.__init__()
      • Normalizer
        • Normalizer.__init__()
      • Domainizer
        • Domainizer.__init__()
      • Affine
        • Affine.__init__()
        • Affine.inv()
    • plotting
      • myrc()
      • saveplot()
      • set_colors()
      • lighten_color()
      • plot_dm()
      • plot_xrv()
      • parallel_coordinates()
      • plot_yx()
      • plot_sens()
      • plot_jsens()
      • plot_tri()
      • plot_pdf1d()
      • plot_pdf2d()
      • plot_pdfs()
      • plot_ens()
      • plot_vars()
      • plot_shade()
      • plot_1d_anchored_single()
      • plot_1d_anchored()
      • plot_2d_anchored_single()
      • plot_2d_anchored()
      • plot_fcn_1d_slice()
      • plot_fcn_2d_slice()
      • plot_uc_sample()
      • plot_uc_exact()
      • plot_samples_pdfs()
      • plot_1d()
      • plot_2d()
      • plot_parity()
      • plot_cov()
      • plot_cov_tri()
      • plot_sensmat()
      • plot_joy()
    • stats
      • get_stats()
      • get_domain()
      • intersect_domain()
      • diam()
    • xutils
      • idt()
      • savepk()
      • loadpk()
      • cartes_list()
      • read_textlist()
      • sample_sphere()
      • get_opt_bw()
      • get_pdf()
      • strarr()
      • project()
      • pick_basis()
      • safe_cholesky()
  • func
    • func
      • blundell()
      • Sine()
      • Summation()
      • Sine10()
      • Ackley()
      • x5()
  • rvar
    • rvs
      • RV
        • RV.__init__()
        • RV.sample()
        • RV.log_prob()
      • MVN
        • MVN.sample()
        • MVN.log_prob()
      • Gaussian_1d
        • Gaussian_1d.mu
        • Gaussian_1d.rho
        • Gaussian_1d.logsigma
        • Gaussian_1d.normal
        • Gaussian_1d.__init__()
        • Gaussian_1d.sample()
        • Gaussian_1d.log_prob()
      • GMM2_1d
        • GMM2_1d.pi
        • GMM2_1d.sigma1
        • GMM2_1d.sigma2
        • GMM2_1d.normal1
        • GMM2_1d.normal2
        • GMM2_1d.__init__()
        • GMM2_1d.log_prob()
  • vi
    • bnet
      • BNet
        • BNet.device
        • BNet.log_prior
        • BNet.log_variational_posterior
        • BNet.nnmodel
        • BNet.nparams
        • BNet.param_names
        • BNet.param_priors
        • BNet.params
        • BNet.rparams
        • BNet.__init__()
        • BNet.del_attr()
        • BNet.set_attr()
        • BNet.forward()
        • BNet.sample_elbo()
        • BNet.viloss()
  • ens
    • learner
      • Learner
        • Learner.nnmodel
        • Learner.best_model
        • Learner.trained
        • Learner.verbose
        • Learner.__init__()
        • Learner.print_params()
        • Learner.init_params()
        • Learner.fit()
        • Learner.predict()

Examples:

  • Fit Examples
    • ex_fit.py — 1-D Function Approximation
    • ex_fit_2d.py — 2-D Fit with Periodic Loss
    • ex_ufit.py — All UQ Solvers
  • Linear Regression
    • ex_lreg_mcmc.py — Linear Regression via MCMC
  • Loss Visualization
    • ex_loss.py — Loss Landscape Visualisation

Theory:

  • NN Architectures
    • Base Class (MLPBase)
    • Multilayer Perceptron (MLP)
      • Constructor Parameters
      • Architecture Diagram
    • Residual Network (RNet)
      • Weight Parameterization
      • Constructor Parameters
    • PyTorch Functions
    • Numpy Wrapper (NNWrap)
    • Summary
  • UQ4NN Solvers
    • MCMC (NN_MCMC)
      • Adaptive Metropolis (AMCMC)
      • Hamiltonian Monte Carlo (HMC)
      • Metropolis-Adjusted Langevin Algorithm (MALA)
    • Deep Ensemble (NN_Ens)
    • Randomized MAP Sampling (NN_RMS)
    • Variational Inference (NN_VI)
      • Variational Family
      • Scale Mixture Prior
      • ELBO Loss
    • Laplace Approximation (NN_Laplace)
    • SWAG (NN_SWAG)
    • Summary of Solvers
    • References
  • Loss Functions
    • Negative Log-Likelihood
    • Gaussian Prior
    • Negative Log-Posterior
  • NN Training
    • Training Objective
    • Loss Functions
      • Mean Squared Error (loss_fn='mse')
      • Negative Log-Posterior (loss_fn='logpost')
      • Log-Loss (loss_fn='logloss')
      • Custom Loss (loss_xy)
    • Optimizers
    • Learning Rate Schedules
    • Mini-Batch Training
    • Early Stopping
    • Arguments
    • Return Value

Misc:

  • Index
  • References
  • Test Suite
    • Test Modules
    • Coverage Summary
  • Class Inheritance Diagrams
    • UQ Solvers
    • Neural-Network Architectures
    • MCMC Samplers
    • Random Variables
    • Input/Output Maps
QUiNN
  • Overview: module code

All modules for which code is available

  • quinn.ens.learner
  • quinn.func.funcs
  • quinn.mcmc.admcmc
  • quinn.mcmc.hmc
  • quinn.mcmc.mala
  • quinn.mcmc.mcmc
  • quinn.nns.losses
  • quinn.nns.mlp
  • quinn.nns.nnbase
  • quinn.nns.nns
  • quinn.nns.nnwrap
  • quinn.nns.rnet
  • quinn.nns.tchutils
  • quinn.rvar.rvs
  • quinn.solvers.nn_ens
  • quinn.solvers.nn_laplace
  • quinn.solvers.nn_mcmc
  • quinn.solvers.nn_rms
  • quinn.solvers.nn_swag
  • quinn.solvers.nn_vi
  • quinn.solvers.quinn
  • quinn.utils.maps
  • quinn.utils.plotting
  • quinn.utils.stats
  • quinn.utils.xutils
  • quinn.vi.bnet

© Copyright TBD.

Built with Sphinx using a theme provided by Read the Docs.