Changelog

Navigation

Changelog#

v0.26.1#

⚠️ Breaking Changes#

  • Make PyMC and Pyro optional dependencies (#1835): pip install sbi no longer installs pymc or pyro-ppl. Users who need Pyro or PyMC MCMC samplers should install extras:

    • pip install "sbi[pyro]" for Pyro samplers (hmc_pyro, nuts_pyro)

    • pip install "sbi[pymc]" for PyMC samplers (slice_pymc, hmc_pymc, nuts_pymc)

    • pip install "sbi[all]" for both

    • Using a Pyro/PyMC method without the dependency installed raises a clear ImportError with install instructions.

🐛 Bug Fixes#

  • Fix TARP z-scoring bug (#1832): Reference points are now z-scored alongside thetas and posterior_samples when z_score_theta=True, fixing incorrect distance calculations that masked bias detection.

  • Fix broken biased_toy_gaussian test helper: Rewrote to create actual location bias (posterior mean shifted from truth) instead of the previous NaN-producing formula.

📖 Documentation#

  • Streamlined README installation section, recommend uv as default.

  • Added optional dependency install instructions to README, installation guide, and relevant tutorials.

🔧 Improvements#

  • Change default num_bins in TARP: run_tarp and _run_tarp now default to num_bins=None (auto-scales to num_sims // 10) instead of the hardcoded 30, improving KS test power for larger sample sizes.

🔒 Security#

  • Added SLSA build provenance attestations to the release workflow.

  • Added Dependabot for automated GitHub Actions version updates.

  • Added SECURITY.md with vulnerability disclosure policy.

  • Fixed release uploads with --clobber flag.

v0.26.0#

✨ Highlights#

🚀 New Inference Methods & Posteriors#

  • Add AmortizedVIPosterior for amortized variational inference by @janfb in https://github.com/sbi-dev/sbi/pull/1751

  • Diffusion model guidance including PriorGuide by @manuelgloeckler in https://github.com/sbi-dev/sbi/pull/1482

  • Add iid support for NPE via MCMC, VI, IS by @satwiksps in https://github.com/sbi-dev/sbi/pull/1810

🧠 Neural Network Architectures#

  • Port MDN from pyknos, improve numerical stability by @janfb in https://github.com/sbi-dev/sbi/pull/1724

  • Add EDM-style noise schedules for VEScoreEstimator by @janfb in https://github.com/sbi-dev/sbi/pull/1754

  • Add 1D time-series support and automatic input projection to TransformerEmbedding by @satwiksps in https://github.com/sbi-dev/sbi/pull/1703

  • Optional LayerNorm and GELU activation for FC embedding by @renecotyfanboy in https://github.com/sbi-dev/sbi/pull/1809

  • Switch build_mlp_classifier to LayerNorm by default by @rsvr76 in https://github.com/sbi-dev/sbi/pull/1806

🔧 Code Quality & Refactoring#

  • Config dataclasses for all net builders to fix kwargs chaining by @janfb in https://github.com/sbi-dev/sbi/pull/1795

  • Config classes for vectorfield kwargs by @janfb in https://github.com/sbi-dev/sbi/pull/1777

  • Refactor and modularize inference training method by @abelaba in https://github.com/sbi-dev/sbi/pull/1651

  • Training dataclasses by @abelaba in https://github.com/sbi-dev/sbi/pull/1668

  • Refactor Plotting functions by @abelaba in https://github.com/sbi-dev/sbi/pull/1631

⭐ New Features#

  • Add max_sampling_time support to rejection samplers by @satwiksps in https://github.com/sbi-dev/sbi/pull/1705

  • Add return_partial_on_timeout option to rejection samplers by @janfb in https://github.com/sbi-dev/sbi/pull/1720

  • FMPE with higher-dim conditions by @gmoss13 in https://github.com/sbi-dev/sbi/pull/1704

  • LC2ST MLP GPU support (closes #1160) by @Dev-Sudarshan in https://github.com/sbi-dev/sbi/pull/1715

  • Pairplot support for mixed samples by @janfb in https://github.com/sbi-dev/sbi/pull/1808

  • Add protocols for sample proposal and accept/reject by @janfb in https://github.com/sbi-dev/sbi/pull/1722

  • Add Tracker protocol and training tracking guide by @janfb in https://github.com/sbi-dev/sbi/pull/1741

  • Use Literal for strict typing of z-scoring options by @satwiksps in https://github.com/sbi-dev/sbi/pull/1744

  • Support shape broadcasting and condition sample_dim in density estimators by @satwiksps in https://github.com/sbi-dev/sbi/pull/1791

  • Decouple hidden features and layers for mixed type nets by @janfb in https://github.com/sbi-dev/sbi/pull/1798

  • Refactoring noise_schedule and time schedule into base class by @janfb in https://github.com/sbi-dev/sbi/pull/1736

🐛 Bug Fixes#

  • fix: disable transform, add warning for empirical prior by @janfb in https://github.com/sbi-dev/sbi/pull/1669

  • MNPE fix for categorical variables by @dgedon in https://github.com/sbi-dev/sbi/pull/1671

  • Fix: Make NPSE picklable by @michaeldeistler in https://github.com/sbi-dev/sbi/pull/1679

  • Fix for z-scoring in vectorfield nets by @michaeldeistler in https://github.com/sbi-dev/sbi/pull/1681

  • fix: autoregressive sampling bug in categorical MADE by @janfb in https://github.com/sbi-dev/sbi/pull/1684

  • Remove gate_activation to allow pickling by @michaeldeistler in https://github.com/sbi-dev/sbi/pull/1694

  • assert finite x_o by @gmoss13 in https://github.com/sbi-dev/sbi/pull/1701

  • fix: ImproperEmpirical is inefficient by @gmoss13 in https://github.com/sbi-dev/sbi/pull/1700

  • Fix device mismatch in NPSE marginal mean/std computation by @satwiksps in https://github.com/sbi-dev/sbi/pull/1707

  • Fix misleading assertion message in NPE train by @satwiksps in https://github.com/sbi-dev/sbi/pull/1706

  • fix: pin PyMC version below 5.20.1 to avoid TypeError (closes #1397) by @Dev-Sudarshan in https://github.com/sbi-dev/sbi/pull/1697

  • Fix memory bloat in multi-round inference by @satwiksps in https://github.com/sbi-dev/sbi/pull/1749

  • Raise informative RuntimeError when training with empty simulations by @satwiksps in https://github.com/sbi-dev/sbi/pull/1750

  • Re-enable time-dependent z-scoring for Flow Matching by @satwiksps in https://github.com/sbi-dev/sbi/pull/1752

  • Fix: Improve error message for x_o shape mismatch by @XBastille in https://github.com/sbi-dev/sbi/pull/1759

  • Fix: forward dropout_probability to CategoricalMADE in mixed density estimator by @coschroeder in https://github.com/sbi-dev/sbi/pull/1789

  • Fix Gaussian Linear task in mini SBI-BM by @jsvetter in https://github.com/sbi-dev/sbi/pull/1782

  • Fix: Add fallback to move_distribution_to_device for complex distributions by @XBastille in https://github.com/sbi-dev/sbi/pull/1785

  • fix: adjust leading dimension handling in log prob batched by @janfb in https://github.com/sbi-dev/sbi/pull/1799

  • fix: Check instance of prior so ValueError is raised after .append_simulations was executed (#1793) by @Jocho-Smith in https://github.com/sbi-dev/sbi/pull/1803

  • fix_mcmc_iid_for_npe by @manuelgloeckler in https://github.com/sbi-dev/sbi/pull/1813

  • fix: quick error fix for broadcasting refactor by @manuelgloeckler in https://github.com/sbi-dev/sbi/pull/1824

  • fix: rejection posterior should use theta_transform by @janfb in https://github.com/sbi-dev/sbi/pull/1827

  • Fix ImproperEmpirical: remove Empirical inheritance, constant log_prob, fix to() by @patelshivani2283-lab in sbi-dev/sbi#1822

🛠️ Maintenance & Improvements#

📖 Documentation & Website#

  • Add how-to guide for embedding time-series by @satwiksps in https://github.com/sbi-dev/sbi/pull/1695

  • Improve docs of MNPE and EnsemblePosterior by @michaeldeistler in https://github.com/sbi-dev/sbi/pull/1677

  • Add SNPE-B to implemented methods by @michaeldeistler in https://github.com/sbi-dev/sbi/pull/1691

  • Fixups for the API of implemented methods by @michaeldeistler in https://github.com/sbi-dev/sbi/pull/1692

  • Fixups by @michaeldeistler in https://github.com/sbi-dev/sbi/pull/1693

  • Update Potentially Misleading Error Message in simulate_for_sbi by @vagechirkov in https://github.com/sbi-dev/sbi/pull/1689

  • Update README with tutorial paper reference by @jahma in https://github.com/sbi-dev/sbi/pull/1712

  • Fix Discord server link in README by @janfb in https://github.com/sbi-dev/sbi/pull/1729

  • How-to guide on Pyro with SBI by @touronc in https://github.com/sbi-dev/sbi/pull/1740

  • Add how-to-guide for hyperparameter optimization with Optuna by @janfb in https://github.com/sbi-dev/sbi/pull/1742

  • Remove arviz dependency, move arviz plotting to how-to-guide by @janfb in https://github.com/sbi-dev/sbi/pull/1743

  • Navigation subheadings for tutorials and API by @dgedon in https://github.com/sbi-dev/sbi/pull/1760

  • Add code examples to trainer class docstrings by @XBastille in https://github.com/sbi-dev/sbi/pull/1763

  • Add type hints and standardize docstrings in torchutils by @khaledeslam20 in https://github.com/sbi-dev/sbi/pull/1815

  • refactor: fix docstrings and signatures of sample* methods by @janfb in https://github.com/sbi-dev/sbi/pull/1719

  • New abstraction level guide by @dgedon in https://github.com/sbi-dev/sbi/pull/1758

  • Unify vector field tutorials, fix legend in pairplot by @janfb in https://github.com/sbi-dev/sbi/pull/1830

🧪 Testing & CI/CD#

  • Refactor z-scoring tests to fix shapes and enable MNLE support by @satwiksps in https://github.com/sbi-dev/sbi/pull/1711

  • fix: skip GPU tests on MPS devices instead of failing by @XBastille in https://github.com/sbi-dev/sbi/pull/1769

  • fix: remove xfail; increase mcmc warmup by @janfb in https://github.com/sbi-dev/sbi/pull/1774

  • Trigger workflow when PR is marked ready for review by @satwiksps in https://github.com/sbi-dev/sbi/pull/1753

  • Refactor embedding net API tests with shared helper by @Dev-Sudarshan in https://github.com/sbi-dev/sbi/pull/1794

  • Fix failing test: Make prior more “within train” by @manuelgloeckler in https://github.com/sbi-dev/sbi/pull/1800

🏗️ Infrastructure & Dependencies#

  • fix release workflow, add sign workflow by @janfb in https://github.com/sbi-dev/sbi/pull/1663

  • Fix release workflow by @janfb in https://github.com/sbi-dev/sbi/pull/1664

  • An Enhancement Proposal Workflow by @janfb in https://github.com/sbi-dev/sbi/pull/1674

  • build: Move notebook dependency to ‘notebook’ extra by @matthewfeickert in https://github.com/sbi-dev/sbi/pull/1714

  • fix: add dependencies for tutorial tests; fix pymc errors by @janfb in https://github.com/sbi-dev/sbi/pull/1716

  • chore: update Codecov settings for coverage flags by @janfb in https://github.com/sbi-dev/sbi/pull/1726

  • Remove pyknos dependency by @janfb in https://github.com/sbi-dev/sbi/pull/1767

  • Remove PyTorch upper version restriction by @manuelgloeckler in https://github.com/sbi-dev/sbi/pull/1768

  • refactor: format pyro guide by @janfb in https://github.com/sbi-dev/sbi/pull/1746

  • fix ruff formatting in main by @manuelgloeckler in https://github.com/sbi-dev/sbi/pull/1792

  • Fix GH workflow vulnerabilities: SHA-pin all actions, fix template injections, restrict permissions, remove legacy dependency pins (fixes CVE-2023-24816) by @janfb in https://github.com/sbi-dev/sbi/pull/1831

🎉 New Contributors#

  • @vagechirkov made their first contribution in https://github.com/sbi-dev/sbi/pull/1689

  • @satwiksps made their first contribution in https://github.com/sbi-dev/sbi/pull/1695

  • @jahma made their first contribution in https://github.com/sbi-dev/sbi/pull/1712

  • @Dev-Sudarshan made their first contribution in https://github.com/sbi-dev/sbi/pull/1697

  • @touronc made their first contribution in https://github.com/sbi-dev/sbi/pull/1740

  • @XBastille made their first contribution in https://github.com/sbi-dev/sbi/pull/1759

  • @coschroeder made their first contribution in https://github.com/sbi-dev/sbi/pull/1789

  • @jsvetter made their first contribution in https://github.com/sbi-dev/sbi/pull/1782

  • @renecotyfanboy made their first contribution in https://github.com/sbi-dev/sbi/pull/1809

  • @rsvr76 made their first contribution in https://github.com/sbi-dev/sbi/pull/1806

  • @khaledeslam20 made their first contribution in https://github.com/sbi-dev/sbi/pull/1815

  • @Jocho-Smith made their first contribution in https://github.com/sbi-dev/sbi/pull/1803

Full Changelog: https://github.com/sbi-dev/sbi/compare/v0.25.0…v0.26.0

v0.25.0#

✨ Highlights#

🚀 New Inference Methods#

  • MNPE class mixed parameter (similar to MNLE) by @dgedon in https://github.com/sbi-dev/sbi/pull/1362

  • Implementation of SNPE-B (#199) by @etouron1 in https://github.com/sbi-dev/sbi/pull/1471

🧠 Neural Network Architectures & Embedding Networks#

  • Add transformer embedding net by @NicolasRR in https://github.com/sbi-dev/sbi/pull/1494

  • Add embedding net that uses 1D causal convolutions (#1459) by @Aranka-S in https://github.com/sbi-dev/sbi/pull/1499

  • Add LRU-backed embedding networks by @famura in https://github.com/sbi-dev/sbi/pull/1512

  • Add ResNet as embedding model by @StefanWahl in https://github.com/sbi-dev/sbi/pull/1472

  • Spectral convolution embedding net by @L-in-da in https://github.com/sbi-dev/sbi/pull/1503

⭐ Major Features & Capabilities#

  • Unify flow matching and score-based models by @StarostinV in https://github.com/sbi-dev/sbi/pull/1497

  • Model misspecification detection based on MMD by @coschroeder in https://github.com/sbi-dev/sbi/pull/1502

  • Marginal estimator log-prob based test for misspecification by @swag2198 in https://github.com/sbi-dev/sbi/pull/1522

  • Adding interface for unconditional flow training by @plcrodrigues in https://github.com/sbi-dev/sbi/pull/1470

  • Support using trained estimators in Pyro models by @sethaxen in https://github.com/sbi-dev/sbi/pull/1491

  • Add util to generate mcmc samples from user defined potential (#1405) by @hayden-johnson in https://github.com/sbi-dev/sbi/pull/1483

  • Logit transform by @anastasiakrouglova in https://github.com/sbi-dev/sbi/pull/1485

  • Log-prob for iid data for score estimators by @Kartik-Sama in https://github.com/sbi-dev/sbi/pull/1508

📚 Documentation & Tutorials#

  • Tutorial on new features for score-based methods #1392 by @touronc in https://github.com/sbi-dev/sbi/pull/1489

  • Docs: Introduce Readthedocs website by @michaeldeistler in https://github.com/sbi-dev/sbi/pull/1463

🐛 Bug Fixes#

  • z_score correct order in Zuko by @anastasiakrouglova in https://github.com/sbi-dev/sbi/pull/1492

  • Minor fix when moving thetas from GPU to CPU by @famura in https://github.com/sbi-dev/sbi/pull/1515

  • Minor fix while using unconditional density estimator and LRU embedding by @ARna06 in https://github.com/sbi-dev/sbi/pull/1556

  • fix: replace ‘in’ operator with ‘==’ for proper classifier comparison by @abelaba in https://github.com/sbi-dev/sbi/pull/1550

  • flowmatching condition shape fix by @gmoss13 in https://github.com/sbi-dev/sbi/pull/1584

  • patch for torch bug in tarp, run torch.histogram with cpu-only tensor by @psteinb in https://github.com/sbi-dev/sbi/pull/1596

  • fix failing tarp test by @janfb in https://github.com/sbi-dev/sbi/pull/1628

  • fix: cap max_sampling_batch_size to prevent excessive memory by @janfb in https://github.com/sbi-dev/sbi/pull/1624

  • 1561 computation of denoising posterior precision matrix in jac method score fn iid by @manuelgloeckler in https://github.com/sbi-dev/sbi/pull/1636

  • fix xfail test, fix deprecation warnings by @janfb in https://github.com/sbi-dev/sbi/pull/1642

  • fix: iid-score device handling by @janfb in https://github.com/sbi-dev/sbi/pull/1650

  • fix: fmpe singularity on sde sampling by @manuelgloeckler in https://github.com/sbi-dev/sbi/pull/1661

🛠️ Maintenance & Improvements#

🔧 Code Quality & Refactoring#

  • Refactoring flow and score matching classes and nets by @manuelgloeckler in https://github.com/sbi-dev/sbi/pull/1544

  • Rename inference trainer classes by @abelaba in https://github.com/sbi-dev/sbi/pull/1605

  • Rename VectorFieldInference to VectorFieldTrainer by @abelaba in https://github.com/sbi-dev/sbi/pull/1614

  • Refactor build_posterior to Eliminate Duplication Across Trainers by @abelaba in https://github.com/sbi-dev/sbi/pull/1610

  • Refactor build posterior method arguments to use Literals by @abelaba in https://github.com/sbi-dev/sbi/pull/1606

  • Refactor build_posterior Posterior Configuration Using Dataclasses by @abelaba in https://github.com/sbi-dev/sbi/pull/1619

  • Use TypeAlias and consistent naming for sbi types by @janfb in https://github.com/sbi-dev/sbi/pull/1637

  • Add protocol for estimator builder by @abelaba in https://github.com/sbi-dev/sbi/pull/1633

  • Improve abc implementation by @janfb in https://github.com/sbi-dev/sbi/pull/1615

  • Refactor RatioEstimator to subclass ConditionalEstimator @abelaba in https://github.com/sbi-dev/sbi/pull/1652

🏷️ Type Hints & API Improvements#

  • fix: add enum for flow options to fix type hints. by @janfb in https://github.com/sbi-dev/sbi/pull/1562

  • fix LC2ST kwarg typing by @janfb in https://github.com/sbi-dev/sbi/pull/1565

  • fix: Update RatioEstimator classifier argument to use a Protocol by @abelaba in https://github.com/sbi-dev/sbi/pull/1582

  • Update append_simulations return type to Self by @abelaba in https://github.com/sbi-dev/sbi/pull/1622

  • Deprecation Warnings for build_posterior stringly typed parameters by @abelaba in https://github.com/sbi-dev/sbi/pull/1627

🧪 Testing & CI/CD#

  • Testmon by @manuelgloeckler in https://github.com/sbi-dev/sbi/pull/1452

  • disable testmon for now by @manuelgloeckler in https://github.com/sbi-dev/sbi/pull/1467

  • chore: Use pytest-split plugin in ci workflow by @schroedk in https://github.com/sbi-dev/sbi/pull/1465

  • tests: refactor “not slow” tests to be not so slow by @janfb in https://github.com/sbi-dev/sbi/pull/1495

  • Test for known pytorch distribution transform issue by @dgedon in https://github.com/sbi-dev/sbi/pull/1504

  • xfail scan test on python 3.13 by @manuelgloeckler in https://github.com/sbi-dev/sbi/pull/1533

  • Gpu test for VectorFieldPosterior by @jorobledo in https://github.com/sbi-dev/sbi/pull/1542

  • set vector field iid-tests xfail by @janfb in https://github.com/sbi-dev/sbi/pull/1554

  • Changed the xfail condition for LRU tests with mode=’scan’ by @famura in https://github.com/sbi-dev/sbi/pull/1552

  • Fix/lru test by @Matthijspals in https://github.com/sbi-dev/sbi/pull/1568

  • refactor sbc funcs and tests by @janfb in https://github.com/sbi-dev/sbi/pull/1578

  • chore: remove testmon, add codecov test analytics by @janfb in https://github.com/sbi-dev/sbi/pull/1592

  • chore: reorder setup steps for Python and uv in CI/CD workflows by @janfb in https://github.com/sbi-dev/sbi/pull/1601

  • Fix/lc2st numpy type fixes by @janfb in https://github.com/sbi-dev/sbi/pull/1613

  • Fix failing CI on main. by @janfb in https://github.com/sbi-dev/sbi/pull/1618

  • Fix slow vector field tests by @janfb in https://github.com/sbi-dev/sbi/pull/1657

  • Add tests for sensitivity analysis by @janfb in https://github.com/sbi-dev/sbi/pull/1662

📖 Documentation & Website#

  • Fix tests for new docs by @michaeldeistler in https://github.com/sbi-dev/sbi/pull/1475

  • Prevent notebook execution upon doc build by @michaeldeistler in https://github.com/sbi-dev/sbi/pull/1477

  • Fix broken links on website by @michaeldeistler in https://github.com/sbi-dev/sbi/pull/1480

  • fix: Update documentation folder path by @abelaba in https://github.com/sbi-dev/sbi/pull/1510

  • fix path to contribute.md by @psteinb in https://github.com/sbi-dev/sbi/pull/1507

  • Add utils to docs by @sethaxen in https://github.com/sbi-dev/sbi/pull/1520

  • Update new Readthedocs website by @michaeldeistler in https://github.com/sbi-dev/sbi/pull/1519

  • fix broken links on new website by @michaeldeistler in https://github.com/sbi-dev/sbi/pull/1538

  • Fixups for new website landing page by @michaeldeistler in https://github.com/sbi-dev/sbi/pull/1539

  • Fix: add tutorial page to mkdocs website by @michaeldeistler in https://github.com/sbi-dev/sbi/pull/1540

  • Fix broken links in some tutorials by @michaeldeistler in https://github.com/sbi-dev/sbi/pull/1541

  • Add examples to documentation by @jorobledo in https://github.com/sbi-dev/sbi/pull/1548

  • docs: Add importance_sampling_parameters to build_posterior docstring. by @abelaba in https://github.com/sbi-dev/sbi/pull/1558

  • Add missing arguments to LikelihoodEstimator and RatioEstimator docstrings. by @abelaba in https://github.com/sbi-dev/sbi/pull/1571

  • Fixups for new RTD website by @michaeldeistler in https://github.com/sbi-dev/sbi/pull/1573

  • Tutorial with a more representative training loop by @michaeldeistler in https://github.com/sbi-dev/sbi/pull/1574

  • Fixups for rendering of HH tutorial notebook by @michaeldeistler in https://github.com/sbi-dev/sbi/pull/1575

  • fix for colors in Hodgkin-Huxley notebook by @michaeldeistler in https://github.com/sbi-dev/sbi/pull/1576

  • Add citations to how-to guide by @michaeldeistler in https://github.com/sbi-dev/sbi/pull/1579

  • Clarify fullscreen view of applications-explorer by @michaeldeistler in https://github.com/sbi-dev/sbi/pull/1580

  • Fixups for the documentation by @michaeldeistler in https://github.com/sbi-dev/sbi/pull/1585

  • fixed misrendered bullet list, tested locally by @psteinb in https://github.com/sbi-dev/sbi/pull/1594

  • Improvements to L-C2ST tutorial by @michaeldeistler in https://github.com/sbi-dev/sbi/pull/1588

  • docs: Change colortheme in light mode by @michaeldeistler in https://github.com/sbi-dev/sbi/pull/1638

  • Posterior parameters doc by @abelaba in https://github.com/sbi-dev/sbi/pull/1644

  • fix contributing links by @janfb in https://github.com/sbi-dev/sbi/pull/1647

  • docs: add posterior parameters dataclass how to guide by @abelaba in https://github.com/sbi-dev/sbi/pull/1654

🏗️ Infrastructure & Dependencies#

  • Add uv support by @abelaba in https://github.com/sbi-dev/sbi/pull/1518

  • switch to numfocus code of conduct by @janfb in https://github.com/sbi-dev/sbi/pull/1560

  • Update readme with new JOSS citation by @janfb in https://github.com/sbi-dev/sbi/pull/1564

  • update numfocus code of conduct by @janfb in https://github.com/sbi-dev/sbi/pull/1602

  • Added Apache License reference comments to source files + CI bash script check by @nMaax in https://github.com/sbi-dev/sbi/pull/1599

👥 User Experience & Warnings#

  • Change xfail to skipif as outcome is not consistent by @gmoss13 in https://github.com/sbi-dev/sbi/pull/1487

  • Add warning when using append_simulations with exclude_invalid_x=True by @abelaba in https://github.com/sbi-dev/sbi/pull/1486

  • Batch sampling slow without warning by @dgedon in https://github.com/sbi-dev/sbi/pull/1490

  • Clarify pbar annotation in sample_batched for DirectPosterior by @StefanWahl in https://github.com/sbi-dev/sbi/pull/1493

🎮 GPU Support & Device Handling#

  • Prior to(device) by @jorobledo in https://github.com/sbi-dev/sbi/pull/1505

  • posterior.to(device) by @jorobledo in https://github.com/sbi-dev/sbi/pull/1527

🔧 Miscellaneous Improvements#

  • ref: update tests, add types and docs to marginal trainer by @janfb in https://github.com/sbi-dev/sbi/pull/1516

  • integrate sbi application eplorer by @lappalainenj in https://github.com/sbi-dev/sbi/pull/1567

  • fix: update notebook references by @emmanuel-ferdman in https://github.com/sbi-dev/sbi/pull/1563

  • Update sbiutils.py to use one-dimensional batch by @vivienr in https://github.com/sbi-dev/sbi/pull/1577

  • fix: remove empty list default argument by @abelaba in https://github.com/sbi-dev/sbi/pull/1608

  • fix: throw exception on unsupported activation function by @emmanuel-ferdman in https://github.com/sbi-dev/sbi/pull/1609

  • fix: resolve logger warnings by @emmanuel-ferdman in https://github.com/sbi-dev/sbi/pull/1598

  • fix paths by @manuelgloeckler in https://github.com/sbi-dev/sbi/pull/1641

🎉 New Contributors#

  • @abelaba made their first contribution in https://github.com/sbi-dev/sbi/pull/1486

  • @dgedon made their first contribution in https://github.com/sbi-dev/sbi/pull/1490

  • @StefanWahl made their first contribution in https://github.com/sbi-dev/sbi/pull/1493

  • @touronc made their first contribution in https://github.com/sbi-dev/sbi/pull/1489

  • @jorobledo made their first contribution in https://github.com/sbi-dev/sbi/pull/1505

  • @hayden-johnson made their first contribution in https://github.com/sbi-dev/sbi/pull/1483

  • @sethaxen made their first contribution in https://github.com/sbi-dev/sbi/pull/1491

  • @etouron1 made their first contribution in https://github.com/sbi-dev/sbi/pull/1471

  • @Aranka-S made their first contribution in https://github.com/sbi-dev/sbi/pull/1499

  • @swag2198 made their first contribution in https://github.com/sbi-dev/sbi/pull/1522

  • @StarostinV made their first contribution in https://github.com/sbi-dev/sbi/pull/1497

  • @L-in-da made their first contribution in https://github.com/sbi-dev/sbi/pull/1503

  • @NicolasRR made their first contribution in https://github.com/sbi-dev/sbi/pull/1494

  • @vivienr made their first contribution in https://github.com/sbi-dev/sbi/pull/1577

Full Changelog: https://github.com/sbi-dev/sbi/compare/v0.24.0…v0.25.0

v0.24.0#

✨ Highlights#

  • feat: add CategoricalMADE by @jnsbck in https://github.com/sbi-dev/sbi/pull/1269 (Major New Feature)

  • tests: mini-sbibm by @manuelgloeckler in https://github.com/sbi-dev/sbi/pull/1335 (Major New Feature)

  • feat: Score-based iid sampling by @manuelgloeckler in https://github.com/sbi-dev/sbi/pull/1381 (Major New Feature)

  • Drop python3.9 support, fix ci by @janfb in https://github.com/sbi-dev/sbi/pull/1412 (Python Version Support Change)

  • additional features for NPSE by @gmoss13 in https://github.com/sbi-dev/sbi/pull/1370 (Enhancement)

🐛 Bug Fixes#

  • #1350 leakage correction breaks consistency of log prob vs log prob batched by @manuelgloeckler in https://github.com/sbi-dev/sbi/pull/1355

  • fix #1343 device handling in mog_log_prob by @janfb in https://github.com/sbi-dev/sbi/pull/1356

  • Fix failing tutorials, change MNLE default for log_transform to False by @janfb in https://github.com/sbi-dev/sbi/pull/1367

  • Fix conditional posterior shape and device bugs. by @janfb in https://github.com/sbi-dev/sbi/pull/1373

  • fix: type fix in VI subclasses. xfail pymc tests. by @janfb in https://github.com/sbi-dev/sbi/pull/1390

  • Temporary Wrappers to fix MADE by @gmoss13 in https://github.com/sbi-dev/sbi/pull/1398

  • Fix mnle tests, MCMC init pbar, sir batch size. by @janfb in https://github.com/sbi-dev/sbi/pull/1410

  • fix mnle tests by @janfb in https://github.com/sbi-dev/sbi/pull/1415

  • fix: protocol and refactor for custom potential by @janfb in https://github.com/sbi-dev/sbi/pull/1409

  • fix docs workflow by @janfb in https://github.com/sbi-dev/sbi/pull/1419

  • fix: gpu-handling for CategoricalMADE by @janfb in https://github.com/sbi-dev/sbi/pull/1448

🛠️ Maintenance & Improvements#

  • increase tutorial test timeout by @janfb in https://github.com/sbi-dev/sbi/pull/1360

  • add nan check to _loss method. by @janfb in https://github.com/sbi-dev/sbi/pull/1361

  • Update and pin pre commit and ruff to recent version. by @janfb in https://github.com/sbi-dev/sbi/pull/1358

  • add nan handling in diagnostics by @janfb in https://github.com/sbi-dev/sbi/pull/1359

  • Improve tests to detect circular imports and resolve all of them by @manuelgloeckler in https://github.com/sbi-dev/sbi/pull/1357

  • docs: add details about zuko density estimators by @janfb in https://github.com/sbi-dev/sbi/pull/1387

  • fix: duplication of ‘large number’ in warning by @turnmanh in https://github.com/sbi-dev/sbi/pull/1391

  • improve PR and issue templates by @janfb in https://github.com/sbi-dev/sbi/pull/1399

  • perf: speed up CI with uv by @janfb in https://github.com/sbi-dev/sbi/pull/1400

  • tests: add pytest testmon plugin to speed up CI by @janfb in https://github.com/sbi-dev/sbi/pull/1402

  • small fixes to score methods by @janfb in https://github.com/sbi-dev/sbi/pull/1404

  • docs: Corrected python version in installation document (Revised) by @VijaySamant4368 in https://github.com/sbi-dev/sbi/pull/1423

  • docs: Markdown formatting compliant to Markdown Linter (solves #1434) by @nMaax in https://github.com/sbi-dev/sbi/pull/1443

🎉 New Contributors#

  • @VijaySamant4368 made their first contribution in https://github.com/sbi-dev/sbi/pull/1423

  • @nMaax made their first contribution in https://github.com/sbi-dev/sbi/pull/1443

Full Changelog: https://github.com/sbi-dev/sbi/compare/v0.23.3…v0.24.0

v0.23.3#

Highlights 🤩#

What’s Changed 🚧#

New Contributors 🎉#

Full Changelog: https://github.com/sbi-dev/sbi/compare/v0.23.2…v0.23.3

v0.23.2#

Bug Fixes#

  • fixup for failing hmc test by @michaeldeistler (#1247)

  • fix: make RestrictedPrior a distribution to enable log_prob @janfb (#1257)

  • fix: npe iid handling by @janfb (#1262)

  • fix: tutorials test error handling, fix bugs in tutorials by @janfb (#1264)

  • fix #1260: include points in plotting limits by @janfb (#1265)

  • fix: conditioned potential error handling by @janfb, @michaeldeistler (#1275, #1289)

  • fix: Allow 1D pytorch distributions by @michaeldeistler (#1286)

Documentation#

  • Rename SNPE to NPE in the README by @michaeldeistler (#1248)

  • update pickling FAQ by @michaeldeistler (#1255)

  • Adding example for custom DataLoader to tutorial 18 by @psteinb (#1256)

  • docs: add readme intro to docs landing page by @janfb (#1272)

  • Change sampling method for LC2ST to sample_batched() by @JuliaLinhart (#1279)

Maintenance#

  • Refactor simulate_for_sbi location by @samadpls (#1253)

  • build: devcontainer update by @janfb (#1252)

  • fix: docker notebook python version by @janfb (#1258)

  • refactor: remove outputs except plots from tutorials. by @janfb (#1266)

  • build: automatic nb stripping and pypi upload by @janfb (#1267)

  • refactor: remove deprecated x_shape where not needed by @janfb (#1271)

  • more explicit error message for CNN shapes by @Ankush7890 (#1281)

v0.23.1#

  • fix: include score folder by adding __init__.py (#1245 #1246)

v0.23.0#

Announcements#

  • Re-licensing: license change from AGPLv3 to Apache-2.0 (see #997 for details)

  • sbi is now affiliated with NumFOCUS 🎉

  • New contributors 🎉: @anastasiakrouglova, @theogruner, @felixp8, @Matthijspals, @jsvetter, @pfuhr, @turnmanh, @fariedabuzaid, @augustes, @zinastef, @Baschdl, @danielmk, @lisahaxel, @janko-petkovic, @samadpls, @ThomasGesseyJonesPX, @schroedk

Major Changes#

  • internal renaming of all inference classes from, e.g., SNPE to NPE (i.e., we removed the S prefix). The functionality of the classes remains the same. The NPE class handles both the amortized and sequential versions of neural posterior estimation. An alias for SNPE (and other sequential methods) still exists for backwards compatibility (#1238) (@michaeldeistler).

  • change sbi default parameters: training_batch_size=200, num_chains=20 (#1221) (@janfb)

  • change imports of posterior_nn, likelihood_nn, and classifier_nn. They should now be imported from sbi.neural_nets, not from sbi.utils (#994) (@famura)

  • big refactoring of plotting utilities, new tutorial (#1084) (@Matthijspals)

  • improved tutorials and website documentation (#1012, #1051, #1073) (@augustes, @zinaStef, @lisahaxel, @psteinb)

  • improved website structure and contribution guides (#1019) (@tomMoral, @janfb)

  • drop support for python3.8 and torch1.12 (#1233)

  • refactor folder structure and naming of neural_nets (#1237) (@michaeldeistler)

New Features#

  • full flexibility over the training loop (#983) (@michaeldeistler)

  • unified density estimator classes (#952, #965, #979, #1151) (@michaeldeistler, @gmoss13, @tomMoral, @manualgloeckler)

  • vectorized sampling and log_prob for (S)NPE given batches of x (#1153) (@manuelgloeckler, @michaeldeistler)

  • batched sampling for vectorized MCMC samplers (#1176, #1210) (@gmoss13, @janfb)

  • support @zuko as a backend for normalizing flows (#1088, #1116) (@anastasiakrouglova)

  • local c2st metric (#1109) (@JuliaLinhart)

  • tarp coverage metric (#1106) (@psteinb)

  • add interface for @PyMC samplers (#1053) (@famura, @felixp8)

  • flow matching density estimators (#1049) (@turnmanh, @fariedabuzaid, @janfb)

  • score matching density estimators (#1015) (@rdgao, @jsvetter, @pfuhr, @manuelgloeckler, @michaeldeistler, @janfb)

  • ABC methods for trial-based data using statistical distances (#1104) (@theogruner)

  • support Apple MPS as gpu device (#912) (@janfb)

  • dev container for using sbi in codespaces on GitHub (#1070) (@turnmanh)

  • enable importance sampling for likelihood-based estimators (#1183) (@manuelgloeckler)

  • refactoring and unified shape handling for RatioEstimator (#1097) (@bkmi)

  • faster sbc and tarp calibration checks via batched sampling (#1196) (@janfb)

  • batched sampling and embedding net support for MNLE (#1203) (@janfb)

  • adapt MNLE to new densitye stimator abstraction (#1089) (@coschroeder)

  • better plotting options for coverage plots (#1039, #1212) (@janfb)

  • allow for potential_fn to be a Callable (#943) (@michaeldeistler)

Bug Fixes#

  • bugfix for embedding net tutorial (#1159) (@deismic)

  • Fixup for process_x in EnsemblePosterior (#1148) (@deismic)

  • fixed notebook by changing MCMC parameters (#1058) (@zinaStef)

  • fix: add NeuralPosteriorEnsemble to utils.init (#1002) (@jnsbck)

  • fix: print_false_positive_rate (#976) (@danielmk)

  • fix: make VIPosterior pickable (#951) (@manuelgloeckler)

  • fix: bug in importance sampled posterior (#1081) (@max-dax)

  • fix: embedding device and warning handling (#1186) (@janfb)

  • fix: c2st with constant features (#1204) (@janfb)

  • fix: erroneous warnings about different devices (#1225, @ThomasGesseyJonesPX)

  • fix: type annotation in class ConditionedPotential (#1222) (@schroedk)

Maintenance and other changes#

  • add pre-commit hooks (#955) (@janfb)

  • add ruff to replace isort, black, flake (#960, #978, #1113) (@janfb)

  • switch to pyproject.toml for package specification (#941) (@janfb)

  • Split the GitHub workflow in CI and CD (#1063) (@famura)

  • split linting process from the CI/CD workflow (#1164) (@tomMoral)

  • Switch to the newest pyright and fix all typing errors (#1045, #1108) (@Baschdl)

  • introduce two docs versions: latest pointing to latest release at https://sbi-dev.github.io/sbi/latest/ and dev pointing to the latest version on main https://sbi-dev.github.io/sbi/dev/

v0.22.0#

API change#

  • We have moved sbi to an new github organization: https://github.com/sbi-dev/sbi

  • We have changed the website of the sbi docs: https://sbi-dev.github.io/sbi/.

  • sbi.analysis.pairplot: upper was replaced by offdiag and will be deprecated in a future release.

Features and enhancements#

  • size-invariant embedding nets for amortized inference with iid-data (@janfb, #808)

  • option for new using MAF with rational quadratic splines (thanks to @ImahnShekhzadeh, #819)

  • improved docstring for process_prior (thanks to @musoke, #813)

  • extended tutorial for SBI with iid data (@janfb, #857)

  • new tutorial for SBI with experimental conditions and mixed data (@janfb, #829)

  • New options for pairplot:

    • upper is now called offdiag to match other kwargs.

    • alternating colors for samples and points

    • option to add a legend and pass kwargs for the legend.

Bug fixes#

  • fixed memory leak in in append_simulations (thanks to @VictorSven, #803)

  • bug fix for CNRE (thanks to @bkmi, #815)

  • bug fix for iid-inference with posterior ensembles (@janfb, #826)

  • bug fix for simulation-based calibration with VI posteriors (@janfb, #834, #838)

  • bug fix for BoxUniform device handling (@janfb, #854, #856)

  • bug fix for MAP estimates with independent priors (@janfb, #867)

  • bug fix for tutorial on SBC (@michaeldeistler, #891)

  • fix spurious seeding for simulate_for_sbi (@jan-matthis, #876)

  • bump python version of github action tests to 3.9.13 (@michaeldeistler, #888, #900)

v0.21.0#

v0.20.0#

Major changes and bug fixes#

  • implementation of “Truncated proposals for scalable and hassle-free sbi” (#754)

  • sample-based expected coverage tests (#754)

  • permutation invariant embedding to allow iid data in SNPE (thanks @coschroeder, #751)

  • convolutional neural network embedding (thanks @coschroeder, #745, #751, #769)

  • disallow invalid simulations when using SNLE, SNRE, or atomic SNPE-C (#768)

Enhancements#

  • add tutorial on all available methods (#754)

  • allow seeding of simulate_for_sbi on multiple workers (#762)

  • expose enable_transforms in sampler interface (#756)

  • bugfix for building the transformation of transformed distributions (#756)

v0.19.2#

  • Rely on new version of pyknos with bugfix for APT with MDNs (#734)

  • bugfix: atomic SNPE-C now allows any kind of proposal (#732)

  • bugfix for SNPE with implicit prior on GPU (#730)

  • SNPE-A has force_first_round_loss=True as default (#729)

v0.19.1#

  • bug fix for ArviZ integration (#727)

v0.19.0#

Major changes and bug fixes#

  • new option to sample posterior using importance sampling (#692)

  • new option to use arviz for posterior plotting and MCMC diagnostics (#546, #607, thanks to @sethaxen)

  • fixes for using the VIPosterior with MultipleIndependent prior, a51e93b

  • bug fix for sir (sequential importance reweighting) for MCMC initialization (#692)

  • bug fix for SNPE-A 565082c

  • bug fix for validation loader batch size (#674, thanks to @bkmi)

  • small bug fixes for pairplot and MCMC kwargs

Enhancements#

  • improved and new tutorials:

    • Tutorial for simulation-based calibration (SBC) (#629, thanks to @psteinb)

    • Tutorial for sampling the conditional posterior (#667)

  • new option to use first-round loss in all rounds

  • simulated data is now stored as Dataset to reduce memory load and add flexibility with large data sets (#685, thanks to @tbmiller-astro)

  • refactoring of summary write for better training logs with tensorboard (#704)

  • new option to find peaks of 1D posterior marginals without gradients (#707, #708, thanks to @Ziaeemehr)

  • new option to not use parameter transforms in DirectPosterior for more flexibility with custom priors (#714)

v0.18.0#

Breaking changes#

  • Posteriors saved under sbi v0.17.2 or older can not be loaded under sbi v0.18.0 or newer.

  • sample_with can no longer be passed to .sample(). Instead, the user has to rerun .build_posterior(sample_with=...). (#573)

  • the posterior no longer has the the method .sample_conditional(). Using this feature now requires using the sampler interface (see tutorial here) (#573)

  • retrain_from_scratch_each_round is now called retrain_from_scratch (#598, thanks to @jnsbck)

  • API changes that had been introduced in sbi v0.14.0 and v0.15.0 are not enforced. Using the interface prior to those changes leads to an error (#645)

  • prior passed to SNPE / SNLE / SNRE must be a PyTorch distribution (#655), see FAQ-7 for how to pass use custom prior.

Major changes and bug fixes#

  • new sampler interface (#573)

  • posterior quality assurance with simulation-based calibration (SBC) (#501)

  • added Sequential Neural Variational Inference (SNVI) (Glöckler et al. 2022) (#609, thanks to @manuelgloeckler)

  • bugfix for SNPE-C with mixture density networks (#573)

  • bugfix for sampling-importance resampling (SIR) as init_strategy for MCMC (#646)

  • new density estimator for neural likelihood estimation with mixed data types (MNLE, #638)

  • MCMC can now be parallelized across CPUs (#648)

  • improved device check to remove several GPU issues (#610, thanks to @LouisRouillard)

Enhancements#

  • pairplot takes ax and fig (#557)

  • bugfix for rejection sampling (#561)

  • remove warninig when using multiple transforms with NSF in single dimension (#537)

  • Sampling-importance-resampling (SIR) is now the default init_strategy for MCMC (#605)

  • change mp_context to allow for multi-chain pyro samplers (#608, thanks to @sethaxen)

  • tutorial on posterior predictive checks (#592, thanks to @LouisRouillard)

  • add FAQ entry for using a custom prior (#595, thanks to @jnsbck)

  • add methods to plot tensorboard data (#593, thanks to @lappalainenj)

  • add option to pass the support for custom priors (#602)

  • plotting method for 1D marginals (#600, thanks to @guymoss)

  • fix GPU issues for conditional_pairplot and ActiveSubspace (#613)

  • MCMC can be performed in unconstrained space also when using a MultipleIndependent distribution as prior (#619)

  • added z-scoring option for structured data (#597, thanks to @rdgao)

  • refactor c2st; change its default classifier to random forest (#503, thanks to @psteinb)

  • MCMC init_strategy is now called proposal instead of prior (#602)

  • inference objects can be serialized with pickle (#617)

  • preconfigured fully connected embedding net (#644, thanks to @JuliaLinhart #624)

  • posterior ensembles (#612, thanks to @jnsbck)

  • remove gradients before returning the posterior (#631, thanks to @tomMoral)

  • reduce batchsize of rejection sampling if few samples are left (#631, thanks to @tomMoral)

  • tutorial for how to use SBC (#629, thanks to @psteinb)

  • tutorial for how to use SBI with trial-based data and mixed data types (#638)

  • allow to use a RestrictedPrior as prior for SNPE (#642)

  • optional pre-configured embedding nets (#568, #644, thanks to @JuliaLinhart)

v0.17.2#

Minor changes#

  • bug fix for transforms in KDE (#552)

v0.17.1#

Minor changes#

  • improve kwarg handling for rejection abc and smcabc

  • typo and link fixes (#549, thanks to @pitmonticone)

  • tutorial notebook on crafting summary statistics with sbi (#511, thanks to @ybernaerts)

  • small fixes and improved documenentation for device handling (#544, thanks to @milagorecki)

v0.17.0#

Major changes#

  • New API for specifying sampling methods (#487). Old syntax:

posterior = inference.build_posterior(sample_with_mcmc=True)

New syntax:

posterior = inference.build_posterior(sample_with="mcmc")  ## or "rejection"

  • Rejection sampling for likelihood(-ratio)-based posteriors (#487)

  • MCMC in unconstrained and z-scored space (#510)

  • Prior is now allowed to lie on GPU. The prior has to be on the same device as the one passed for training (#519).

  • Rejection-ABC and SMC-ABC now return the accepted particles / parameters by default, or a KDE fit on those particles (kde=True) (#525).

  • Fast analytical sampling, evaluation and conditioning for DirectPosterior trained with MDNs (thanks @jnsbck #458).

Minor changes#

  • scatter allowed for diagonal entries in pairplot (#510)

  • Changes to default hyperparameters for SNPE_A (thanks @famura, #496, #497)

  • bugfix for within_prior checks (#506)

v0.16.0#

Major changes#

  • Implementation of SNPE-A (thanks @famura and @theogruner, #474, #478, #480, #482)

  • Option to do inference over iid observations with SNLE and SNRE (#484, #488)

Minor changes#

  • Fixed unused argument num_bins when using nsf as density estimator (#465)

  • Fixes to adapt to the new support handling in torch v1.8.0 (#469)

  • More scalars for monitoring training progress (thanks @psteinb #471)

  • Fixed bug in minimal.py (thanks @psteinb, #485)

  • Depend on pyknos v0.14.2

v0.15.1#

  • add option to pass torch.data.DataLoader kwargs to all inference methods (thanks @narendramukherjee, #445)

  • fix bug due to release of torch v1.8.0 (#451)

  • expose leakage_correction parameters for log_prob correction in unnormalized posteriors (thanks @famura, #454)

v0.15.0#

Major changes#

  • Active subspaces for sensitivity analysis (#394, tutorial)

  • Method to compute the maximum-a-posteriori estimate from the posterior (#412)

API changes#

  • pairplot(), conditional_pairplot(), and conditional_corrcoeff() should now be imported from sbi.analysis instead of sbi.utils (#394).

  • Changed fig_size to figsize in pairplot (#394).

  • moved user_input_checks to sbi.utils (#430).

Minor changes#

  • Depend on new joblib=1.0.0 and fix progress bar updates for multiprocessing (#421).

  • Fix for embedding nets with SNRE (thanks @adittmann, #425).

  • Is it now optional to pass a prior distribution when using SNPE (#426).

  • Support loading of posteriors saved after sbi v0.15.0 (#427, thanks @psteinb).

  • Neural network training can be resumed (#431).

  • Allow using NSF to estimate 1D distributions (#438).

  • Fix type checks in input checks (thanks @psteinb, #439).

  • Bugfix for GPU training with SNRE_A (thanks @glouppe, #442).

v0.14.3#

  • Fixup for conditional correlation matrix (thanks @JBeckUniTb, #404)

  • z-score data using only the training data (#411)

v0.14.2#

  • Small fix for SMC-ABC with semi-automatic summary statistics (#402)

v0.14.1#

  • Support for training and sampling on GPU including fixes from nflows (#331)

  • Bug fix for SNPE with neural spline flow and MCMC (#398)

  • Small fix for SMC-ABC particles covariance

  • Small fix for rejection-classifier (#396)

v0.14.0#

  • New flexible interface API (#378). This is going to be a breaking change for users of the flexible interface and you will have to change your code. Old syntax:

from sbi.inference import SNPE, prepare_for_sbi

simulator, prior = prepare_for_sbi(simulator, prior)
inference = SNPE(simulator, prior)

## Simulate, train, and build posterior.
posterior = inference(num_simulation=1000)

New syntax:

from sbi.inference import SNPE, prepare_for_sbi, simulate_for_sbi

simulator, prior = prepare_for_sbi(simulator, prior)
inference = SNPE(prior)

theta, x = simulate_for_sbi(simulator, proposal=prior, num_simulations=1000)
density_estimator = inference.append_simulations(theta, x).train()
posterior = inference.build_posterior(density_estimator)  ## MCMC kwargs go here.

More information can be found here here.

  • Fixed typo in docs for infer (thanks @glouppe, #370)

  • New RestrictionEstimator to learn regions of bad simulation outputs (#390)

  • Improvements for and new ABC methods (#395)

    • Linear regression adjustment as in Beaumont et al. 2002 for both MCABC and SMCABC

    • Semi-automatic summary statistics as in Fearnhead & Prangle 2012 for both MCABC and SMCABC

    • Small fixes to perturbation kernel covariance estimation in SMCABC.

v0.13.2#

  • Fix bug in SNRE (#363)

  • Fix warnings for multi-D x (#361)

  • Small improvements to MCMC, verbosity and continuing of chains (#347, #348)

v0.13.1#

  • Make logging of vectorized numpy slice sampler slightly less verbose and address NumPy future warning (#347)

  • Allow continuation of MCMC chains (#348)

v0.13.0#

  • Conditional distributions and correlations for analysing the posterior (#321)

  • Moved rarely used arguments from pairplot into kwargs (#321)

  • Sampling from conditional posterior (#327)

  • Allow inference with multi-dimensional x when appropriate embedding is passed (#335)

  • Fixes a bug with clamp_and_warn not overriding num_atoms for SNRE and the warning message itself (#338)

  • Compatibility with Pyro 1.4.0 (#339)

  • Speed up posterior rejection sampling by introducing batch size (#340, #343)

  • Allow vectorized evaluation of numpy potentials (#341)

  • Adds vectorized version of numpy slice sampler which allows parallel log prob evaluations across all chains (#344)

v0.12.2#

  • Bug fix for zero simulations in later rounds (#318)

  • Bug fix for sbi.utils.sbiutils.Standardize; mean and std are now registered in state dict (thanks @plcrodrigues, #325)

  • Tutorials on embedding_net and presimulated data (thanks @plcrodrigues, #314, #318)

  • FAQ entry for pickling error

v0.12.1#

  • Bug fix for broken NSF (#310, thanks @tvwenger).

v0.12.0#

  • Add FAQ (#293)

  • Fix bug in embedding_net when output dimension does not equal input dimension (#299)

  • Expose arguments of functions used to build custom networks (#299)

  • Implement non-atomic APT (#301)

  • Depend on pyknos 0.12 and nflows 0.12

  • Improve documentation (#302, #305, thanks to @agramfort)

  • Fix bug for 1D uniform priors (#307).

v0.11.2#

  • Fixed pickling of SNRE by moving StandardizeInputs (#291)

  • Added check to ensure correct round number when presimulated data is provided

  • Subclassed Posterior depending on inference algorithm (#282, #285)

  • Pinned pyro to v1.3.1 as a temporary workaround (see #288)

  • Detaching weights for MCMC SIR init immediately to save memory (#292)

v0.11.1#

  • Bug fix for log_prob() in SNRE (#280)

v0.11.0#

  • Changed the API to do multi-round inference (#273)

  • Allow to continue inference (#273)

v0.10.2#

  • Added missing type imports (#275)

  • Made compatible for Python 3.6 (#275)

v0.10.1#

  • Added mcmc_parameters to init methods of inference methods (#270)

  • Fixed detaching of log_weights when using sir MCMC init (#270)

  • Fixed logging for SMC-ABC

v0.10.0#

  • Added option to pass external data (#264)

  • Added setters for MCMC parameters (#267)

  • Added check for density_estimator argument (#263)

  • Fixed NeuralPosterior pickling error (#265)

  • Added code coverage reporting (#269)

v0.9.0#

  • Added ABC methods (#250)

  • Added multiple chains for MCMC and new init strategy (#247)

  • Added options for z-scoring for all inference methods (#256)

  • Simplified swapping out neural networks (#256)

  • Improved tutorials

  • Fixed device keyword argument (#253)

  • Removed need for passing x-shapes (#259)

v0.8.0#

  • First public version

Navigation