FLAML Files

This release improves reproducibility and correctness across AutoML and tuning, including consistent random seed handling for SGD, logistic regression, random forest, and XGBoost, per-fold resampling for class imbalance, corrected quantized float sampling, per-metric trial averages, and preservation of zero-valued anonymous tuning metrics. It strengthens data preparation and evaluation with stable categorical encoding, aligned sample weights for non-shuffled and group splits, and free-memory propagation, while adding compatibility for recent scikit-learn and statsmodels changes. It corrects Statsmodels future forecast horizons and adds a Simplified Chinese README. It also adds Windows ARM64 CI, hardens GitHub Actions permissions and action pinning, and refreshes documentation and dependencies.

Thank you to all contributors for the PRs, issues, reviewing and testing.

Welcome @immu4989, @arpitjain099, @arushiiii18, @danfiedler-msft, @MeiSiristhebest, @AmirF194, @shoemoney, @VenishPaneliya, @nithin-aashik-mcw, @betacatsling, @wangyusheng1985, and @JasonYeYuhe to the community!

What's Changed

• Bump fast-uri from 3.1.0 to 3.1.2 in /website by @dependabot[bot] in https://github.com/microsoft/FLAML/pull/1542
• Bump @babel/plugin-transform-modules-systemjs from 7.19.6 to 7.29.4 in /website by @dependabot[bot] in https://github.com/microsoft/FLAML/pull/1543
• ci(docs): always trigger docs workflow on PRs; move filtering into workflow by @thinkall in https://github.com/microsoft/FLAML/pull/1544
• fix: SGD results now reproducible by @immu4989 in https://github.com/microsoft/FLAML/pull/1541
• fix: LRL1 results now reproducible by @immu4989 in https://github.com/microsoft/FLAML/pull/1546
• fix: RandomForestEstimator honors random_seed config by @immu4989 in https://github.com/microsoft/FLAML/pull/1547
• Add explicit least-privilege permissions to CD workflow by @arpitjain099 in https://github.com/microsoft/FLAML/pull/1548
• Bump brace-expansion from 5.0.5 to 5.0.6 in /website by @dependabot[bot] in https://github.com/microsoft/FLAML/pull/1550
• fix: XGBoostSklearnEstimator honors random_seed config by @immu4989 in https://github.com/microsoft/FLAML/pull/1549
• fix: XGBoostEstimator honors random_seed config by @immu4989 in https://github.com/microsoft/FLAML/pull/1551
• fix: LRL2Classifier honors random_seed config by @immu4989 in https://github.com/microsoft/FLAML/pull/1552
• fix: prepare_data crashes with non shuffle split + sample_weight (#887, [#1553]) by @immu4989 in https://github.com/microsoft/FLAML/pull/1554
• fix: LRL1/LRL2 — migrate off sklearn 1.10 deprecations (penalty, n_jobs) by @immu4989 in https://github.com/microsoft/FLAML/pull/1556
• fix: FLOW2.reach() handles mixed-type conditional parameters (#903) by @immu4989 in https://github.com/microsoft/FLAML/pull/1557
• Bump shell-quote from 1.7.4 to 1.8.4 in /website by @dependabot[bot] in https://github.com/microsoft/FLAML/pull/1559
• docs: ensemble component predict pattern + regression test (#1136) by @immu4989 in https://github.com/microsoft/FLAML/pull/1558
• Bump joi from 17.7.0 to 17.13.4 in /website by @dependabot[bot] in https://github.com/microsoft/FLAML/pull/1560
• fix: DataTransformer pins categorical codes at fit time + warns on unseen values (#1101) by @immu4989 in https://github.com/microsoft/FLAML/pull/1561
• docs: production-deployment guide for trained FLAML models by @immu4989 in https://github.com/microsoft/FLAML/pull/1562
• build(deps): bump http-proxy-middleware from 2.0.9 to 2.0.10 in /website by @dependabot[bot] in https://github.com/microsoft/FLAML/pull/1565
• Avoid pandas resample crash in Windows Python 3.12 forecast tests by @thinkall with @Copilot in https://github.com/microsoft/FLAML/pull/1566
• build(deps): bump websocket-driver from 0.7.4 to 0.7.5 in /website by @dependabot[bot] in https://github.com/microsoft/FLAML/pull/1572
• docs: cross-link AutoML and zero-shot LightGBM notebooks by @arushiiii18 in https://github.com/microsoft/FLAML/pull/1571
• Fix MLflow autologged model reload fitted-state detection for scikit-learn>=1.8.0 by @thinkall with @Copilot in https://github.com/microsoft/FLAML/pull/1573
• build(deps): bump brace-expansion from 5.0.6 to 5.0.7 in /website by @dependabot[bot] in https://github.com/microsoft/FLAML/pull/1574
• build(deps): bump postcss from 8.5.10 to 8.5.25 in /website by @dependabot[bot] in https://github.com/microsoft/FLAML/pull/1579
• build(deps): bump fast-uri from 3.1.2 to 3.1.5 in /website by @dependabot[bot] in https://github.com/microsoft/FLAML/pull/1575
• build(deps): bump shell-quote from 1.8.4 to 1.10.0 in /website by @dependabot[bot] in https://github.com/microsoft/FLAML/pull/1577
• build(deps): bump brace-expansion from 5.0.7 to 5.0.9 in /website by @dependabot[bot] in https://github.com/microsoft/FLAML/pull/1578
• Pin GitHub Actions to full-length commit SHAs by @danfiedler-msft in https://github.com/microsoft/FLAML/pull/1582
• build(deps): bump nanoid from 3.3.16 to 3.3.18 in /website by @dependabot[bot] in https://github.com/microsoft/FLAML/pull/1581
• build(deps): bump the github-actions group with 8 updates by @dependabot[bot] in https://github.com/microsoft/FLAML/pull/1583
• docs: fix undefined variable and comment mismatch in time series forecast example by @MeiSiristhebest in https://github.com/microsoft/FLAML/pull/1580
• build(deps): bump hydra-core from 1.1.1 to 1.3.4 in /test/pipeline_tuning_example by @dependabot[bot] in https://github.com/microsoft/FLAML/pull/1585
• build(deps): bump browserslist from 4.21.4 to 4.28.8 in /website by @dependabot[bot] in https://github.com/microsoft/FLAML/pull/1590
• build(deps): bump fast-uri from 3.1.5 to 3.1.7 in /website by @dependabot[bot] in https://github.com/microsoft/FLAML/pull/1588
• fix: support statsmodels 0.15 in naive forecasters by @thinkall in https://github.com/microsoft/FLAML/pull/1587
• fix(examples): compute the discrete-label threshold from the training split only by @AmirF194 in https://github.com/microsoft/FLAML/pull/1584
• Fix misaligned sample_weight in group holdout split by @shoemoney in https://github.com/microsoft/FLAML/pull/1586
• Forward free_mem_ratio from compute_estimator to the evaluation by @VenishPaneliya in https://github.com/microsoft/FLAML/pull/1591
• Add Windows ARM64 CI support by @nithin-aashik-mcw in https://github.com/microsoft/FLAML/pull/1589
• docs: rerun time series forecasting notebook with saved results by @thinkall in https://github.com/microsoft/FLAML/pull/1592
• build(deps-dev): bump cryptography from 46.0.3 to 50.0.0 by @dependabot[bot] in https://github.com/microsoft/FLAML/pull/1593
• Remove unsupported cryptography pin from Windows ARM CI by @thinkall with @Copilot in https://github.com/microsoft/FLAML/pull/1594
• Compute trial averages from per-metric report counts by @betacatsling in https://github.com/microsoft/FLAML/pull/1601
• build(deps): bump joi from 17.13.4 to 17.13.7 in /website by @dependabot[bot] in https://github.com/microsoft/FLAML/pull/1597
• Honor unit quantization for float sampling by @betacatsling in https://github.com/microsoft/FLAML/pull/1599
• build(deps): bump colord from 2.9.3 to 2.10.0 in /website by @dependabot[bot] in https://github.com/microsoft/FLAML/pull/1595
• Docs: use consistent four-space indentation in Task-Oriented AutoML by @wangyusheng1985 in https://github.com/microsoft/FLAML/pull/1596
• feat: add optional resampler= kwarg to AutoML.fit for per-fold class-imbalance handling (#1200) by @immu4989 in https://github.com/microsoft/FLAML/pull/1568
• Bump version from 2.6.0 to 2.7.0 by @thinkall in https://github.com/microsoft/FLAML/pull/1607
• Fix Statsmodels future forecast horizon by @thinkall in https://github.com/microsoft/FLAML/pull/1608
• Preserve zero-valued anonymous tuning metrics by @betacatsling in https://github.com/microsoft/FLAML/pull/1603
• docs: add Simplified Chinese translation (README.zh.md) by @JasonYeYuhe in https://github.com/microsoft/FLAML/pull/1604

New Contributors

• @immu4989 made their first contribution in https://github.com/microsoft/FLAML/pull/1541
• @arpitjain099 made their first contribution in https://github.com/microsoft/FLAML/pull/1548
• @arushiiii18 made their first contribution in https://github.com/microsoft/FLAML/pull/1571
• @danfiedler-msft made their first contribution in https://github.com/microsoft/FLAML/pull/1582
• @MeiSiristhebest made their first contribution in https://github.com/microsoft/FLAML/pull/1580
• @AmirF194 made their first contribution in https://github.com/microsoft/FLAML/pull/1584
• @shoemoney made their first contribution in https://github.com/microsoft/FLAML/pull/1586
• @VenishPaneliya made their first contribution in https://github.com/microsoft/FLAML/pull/1591
• @nithin-aashik-mcw made their first contribution in https://github.com/microsoft/FLAML/pull/1589
• @betacatsling made their first contribution in https://github.com/microsoft/FLAML/pull/1601
• @wangyusheng1985 made their first contribution in https://github.com/microsoft/FLAML/pull/1596
• @JasonYeYuhe made their first contribution in https://github.com/microsoft/FLAML/pull/1604

Full Changelog: https://github.com/microsoft/FLAML/compare/v2.6.0...v2.7.0