# Version 1.6.0

We are happy to announce the AutoGluon 1.6.0 release!

AutoGluon 1.6.0 introduces new features and improvements to both the tabular and time series modules. This release adds several new tabular foundation models (TabPFN-3, TabDPT-Turbo, TabPFN-2.6, TabICLv2, and Nori) and the Toto-2 pretrained forecasting model, along with new presets that deliver strong accuracy and efficiency gains. It also includes new forecasting metrics, updated model memory estimation, refreshed dependency support (including transformers v5 and the latest PyTorch), and numerous bug fixes and quality-of-life improvements.

This release contains [184 commits from 19 contributors](https://github.com/autogluon/autogluon/graphs/contributors?from=12%2F19%2F2025&to=<RELEASE_DATE>&type=c)! See the full commit change-log here: https://github.com/autogluon/autogluon/compare/1.5.0...1.6.0

Join the community: [![](https://img.shields.io/discord/1043248669505368144?logo=discord&style=flat)](https://discord.gg/wjUmjqAc2N)
Get the latest updates: [![Twitter](https://img.shields.io/twitter/follow/autogluon?style=social)](https://twitter.com/autogluon)

This release supports Python versions 3.10, 3.11, 3.12, and 3.13. Loading models trained on older versions of AutoGluon is not supported. Please re-train models using AutoGluon 1.6.0.

## Spotlight

### Tabular

- New `extreme` preset, the recommended preset when a GPU is available: a 67% win-rate vs AutoGluon 1.5 `extreme` with 27x faster training and 4x faster inference, and an 84% win-rate vs `best` (4h) with 65x faster training. Led by the new [Nori](https://github.com/Synthefy/synthefy-nori), [TabICLv2](https://arxiv.org/abs/2602.11139), and [TabDPT-Turbo](https://openreview.net/pdf?id=Y00pwFyrHR) foundation models alongside a [TabPrep](https://arxiv.org/abs/2606.02384) LightGBM, and every model in it is free for commercial use. [@Innixma](https://github.com/Innixma) ([#5812](https://github.com/autogluon/autogluon/pull/5812))
- New `noncommercial` preset, setting a new state-of-the-art on TabArena: `extreme` plus [TabPFN-3](https://arxiv.org/abs/2605.13986), outperforming [TabFM](https://github.com/google-research/tabfm) in a quarter of the training time, with 20x faster inference and 10x fewer foundation model parameters. Free for research and internal experimentation; commercial use of TabPFN-3 requires a license from [Prior Labs](https://docs.priorlabs.ai/models#tabpfn-model-license). [@Innixma](https://github.com/Innixma) ([#5812](https://github.com/autogluon/autogluon/pull/5812))

**TabArena-Lite, 51/51 tasks**, against the published leaderboard:

| method | elo | rank | improvability (%) | train s | infer s |
|---|---:|---:|---:|---:|---:|
| AutoGluon 1.6 `noncommercial` | **1821** | 6.2 | 5.8 | 159 | 7.0 |
| TabFM (default) | 1791 | 7.0 | 5.1 | 613 | 138.4 |
| AutoGluon 1.6 `extreme` | **1745** | 8.5 | 6.8 | 137 | 6.2 |
| AutoGluon 1.5 `extreme` (4h) | 1638 | 12.9 | 7.9 | 3758 | 27.4 |
| TabPFN-3 (default) | 1632 | 13.2 | 9.1 | 26 | 2.4 |
| AutoGluon 1.4 `best` (4h) | 1516 | 20.0 | 12.1 | 8869 | 33.8 |
| RealMLP (tuned + ensemble) | 1480 | 22.5 | 13.0 | 31982 | 33.8 |

- New `validation_structure` fit argument for non-IID data: declare a grouped or temporal structure and every internal validation split respects it across all models, keeping each group's rows on one side of the split and validating strictly forward in time, with optional forward-chaining for temporal data. [@Innixma](https://github.com/Innixma) ([#5756](https://github.com/autogluon/autogluon/pull/5756), [#5796](https://github.com/autogluon/autogluon/pull/5796))
- New [TabPrep](https://arxiv.org/abs/2606.02384) feature engineering: the paper's feature generators (group-by aggregates, random-subset compression, arithmetic and categorical interactions, out-of-fold target encoding) are now built in, and models can declare their own preprocessing pipelines via `ag.model_specific_feature_generator_kwargs`, applied per model rather than globally. This is what powers the TabPrep LightGBM in the new presets. [@atschalz](https://github.com/atschalz) [@Innixma](https://github.com/Innixma) [@LennartPurucker](https://github.com/LennartPurucker) ([#5618](https://github.com/autogluon/autogluon/pull/5618), [#5588](https://github.com/autogluon/autogluon/pull/5588), [#5665](https://github.com/autogluon/autogluon/pull/5665), [#5278](https://github.com/autogluon/autogluon/pull/5278))


--------

## General

### Dependencies
- Update torch to `>=2.10,<2.14`. [@shchur](https://github.com/shchur) ([#5724](https://github.com/autogluon/autogluon/pull/5724))
- Update transformers to `>=5.3,<5.15`. [@shchur](https://github.com/shchur) ([#5735](https://github.com/autogluon/autogluon/pull/5735))
- Update scikit-learn to `>=1.4.0,<1.10.0`. [@shchur](https://github.com/shchur) ([#5731](https://github.com/autogluon/autogluon/pull/5731), [#5769](https://github.com/autogluon/autogluon/pull/5769))
- Update ray to `>=2.55.0,<2.57`. [@shchur](https://github.com/shchur) ([#5553](https://github.com/autogluon/autogluon/pull/5553), [#5707](https://github.com/autogluon/autogluon/pull/5707))
- Update lightgbm to `>=4.0,<4.8`. [@Innixma](https://github.com/Innixma) ([#5781](https://github.com/autogluon/autogluon/pull/5781))
- Update xgboost to `>=2.0,<3.4`. [@Innixma](https://github.com/Innixma) ([#5783](https://github.com/autogluon/autogluon/pull/5783), [#5784](https://github.com/autogluon/autogluon/pull/5784))
- Update tabpfn to `>=8.0,<8.3`. [@Innixma](https://github.com/Innixma) ([#5788](https://github.com/autogluon/autogluon/pull/5788))
- Update tabicl to `>=2.0,<2.2`. [@Innixma](https://github.com/Innixma) ([#5788](https://github.com/autogluon/autogluon/pull/5788))
- Update numpy to `>=1.25.0,<2.6.0`. [@celestinoxp](https://github.com/celestinoxp) ([#5615](https://github.com/autogluon/autogluon/pull/5615), [#5514](https://github.com/autogluon/autogluon/pull/5514))
- Update onnx to `>=1.21.0,<1.23.0`. [@shchur](https://github.com/shchur) ([#5747](https://github.com/autogluon/autogluon/pull/5747))
- Update nltk to `>=3.10,<3.11`. [@shchur](https://github.com/shchur) ([#5747](https://github.com/autogluon/autogluon/pull/5747))
- Update pyarrow to `>=23.0.1,<25.0.0`. [@shchur](https://github.com/shchur) ([#5707](https://github.com/autogluon/autogluon/pull/5707))
- Update Pillow to `>=10.0.1,<13`. [@shchur](https://github.com/shchur) ([#5707](https://github.com/autogluon/autogluon/pull/5707))
- Update skl2onnx to `>=1.20.0,<1.21.0`. [@shchur](https://github.com/shchur) ([#5707](https://github.com/autogluon/autogluon/pull/5707))
- Update typing-extensions to `>=4.14.0,<5`. [@celestinoxp](https://github.com/celestinoxp) ([#5589](https://github.com/autogluon/autogluon/pull/5589))
- Minor dependency fixes and improvements (raise lower bounds, fix fastai dependency, fix macOS + torch compat issue). [@shchur](https://github.com/shchur) ([#5734](https://github.com/autogluon/autogluon/pull/5734), [#5528](https://github.com/autogluon/autogluon/pull/5528), [#5790](https://github.com/autogluon/autogluon/pull/5790), [#5792](https://github.com/autogluon/autogluon/pull/5792))
- Bump transitive dependencies in `uv.lock` (aiohttp, cryptography, msgpack, pyasn1, pydantic-settings) to satisfy Dependabot. [@shchur](https://github.com/shchur) [@dependabot](https://github.com/dependabot) ([#5759](https://github.com/autogluon/autogluon/pull/5759), [#5742](https://github.com/autogluon/autogluon/pull/5742), [#5745](https://github.com/autogluon/autogluon/pull/5745), [#5743](https://github.com/autogluon/autogluon/pull/5743), [#5800](https://github.com/autogluon/autogluon/pull/5800))

### Fixes and Improvements
- PEP 621 + uv workspace for the full monorepo. [@Innixma](https://github.com/Innixma) ([#5688](https://github.com/autogluon/autogluon/pull/5688), [#5536](https://github.com/autogluon/autogluon/pull/5536))
- Lazy-import networkx, requests, tqdm to speed up `import TabularPredictor` (~25% faster). [@Innixma](https://github.com/Innixma) ([#5677](https://github.com/autogluon/autogluon/pull/5677))
- Remove unsupported "lite" mode. [@Innixma](https://github.com/Innixma) ([#5679](https://github.com/autogluon/autogluon/pull/5679))
- Various code & setup improvement. [@Innixma](https://github.com/Innixma) ([#5687](https://github.com/autogluon/autogluon/pull/5687))
- Fix remote filesystem collision on artifact directory creation. [@Innixma](https://github.com/Innixma) ([#5653](https://github.com/autogluon/autogluon/pull/5653))
- Fix memory hog: remove `pickle.dumps` for DataFrame size estimation. [@celestinoxp](https://github.com/celestinoxp) ([#5563](https://github.com/autogluon/autogluon/pull/5563))
- Fix `get_bool_true_val` failing on category dtype. [@shchur](https://github.com/shchur) ([#5595](https://github.com/autogluon/autogluon/pull/5595))
- Minor fixes and improvements. [@celestinoxp](https://github.com/celestinoxp) [@shchur](https://github.com/shchur) ([#5562](https://github.com/autogluon/autogluon/pull/5562), [#5561](https://github.com/autogluon/autogluon/pull/5561), [#5560](https://github.com/autogluon/autogluon/pull/5560), [#5557](https://github.com/autogluon/autogluon/pull/5557), [#5479](https://github.com/autogluon/autogluon/pull/5479))

--------

## Tabular

<!-- TODO: module intro paragraph -->

### New Models
- Add Nori. [@minkyu-choi07](https://github.com/minkyu-choi07) ([#5705](https://github.com/autogluon/autogluon/pull/5705))
- Add TabPFN-3. [@Innixma](https://github.com/Innixma) ([#5754](https://github.com/autogluon/autogluon/pull/5754))
- Add TabDPT-Turbo. [@Innixma](https://github.com/Innixma) ([#5754](https://github.com/autogluon/autogluon/pull/5754))
- Add TabPFN-2.6. [@Innixma](https://github.com/Innixma) ([#5647](https://github.com/autogluon/autogluon/pull/5647))
- Add TabICLv2. [@Innixma](https://github.com/Innixma) ([#5601](https://github.com/autogluon/autogluon/pull/5601))

### New Features
- Remove fasttext model. [@Innixma](https://github.com/Innixma) ([#5582](https://github.com/autogluon/autogluon/pull/5582))
- Remove old TabPFNv2 code. [@Innixma](https://github.com/Innixma) ([#5524](https://github.com/autogluon/autogluon/pull/5524))
- Add `core_kwargs` and `aux_kwargs` fit params. [@Innixma](https://github.com/Innixma) ([#5646](https://github.com/autogluon/autogluon/pull/5646))
- Express the auto-selected validation method as size curves. [@Innixma](https://github.com/Innixma) ([#5789](https://github.com/autogluon/autogluon/pull/5789))
- Report constraint misses as skips, and add `min_features` / `min_cells` / `max_cells`. [@Innixma](https://github.com/Innixma) ([#5798](https://github.com/autogluon/autogluon/pull/5798))
- Remove the feature cap on TabPFN-2.6 and TabPFN-3. [@Innixma](https://github.com/Innixma) ([#5797](https://github.com/autogluon/autogluon/pull/5797))
- Lift Nori and TabDPT-Turbo feature caps, clamp Nori's GPU estimate. [@Innixma](https://github.com/Innixma) ([#5801](https://github.com/autogluon/autogluon/pull/5801))
- Apply TabICL's pinned checkpoint, add a batch tier, lower the row cap. [@Innixma](https://github.com/Innixma) ([#5802](https://github.com/autogluon/autogluon/pull/5802))
- Add support for quantile problem type in TabPFNv2. [@colesussmeier](https://github.com/colesussmeier) ([#5571](https://github.com/autogluon/autogluon/pull/5571))
- Support loading Mitra from a local checkpoint path. [@shchur](https://github.com/shchur) ([#5765](https://github.com/autogluon/autogluon/pull/5765))
- Reduce feature importance mem usage by 25x. [@Innixma](https://github.com/Innixma) ([#5645](https://github.com/autogluon/autogluon/pull/5645))
- GPU memory estimation plumbing + calibrated CPU/GPU memory estimates for foundation models. [@Innixma](https://github.com/Innixma) ([#5757](https://github.com/autogluon/autogluon/pull/5757))
- Simplify the memory-estimate API: derive capability tags, default the instance wrappers. [@Innixma](https://github.com/Innixma) ([#5768](https://github.com/autogluon/autogluon/pull/5768))
- Account for GPU memory when budgeting parallel bagging folds. [@Innixma](https://github.com/Innixma) ([#5791](https://github.com/autogluon/autogluon/pull/5791))
- [experimental] Prototype GPU support for `fit_strategy='parallel'` (behind `AG_PARALLEL_GPU`). [@Innixma](https://github.com/Innixma) ([#5702](https://github.com/autogluon/autogluon/pull/5702))
- Support xgboost v3.3: shared fit/predict category space for `enable_categorical`. [@Innixma](https://github.com/Innixma) ([#5784](https://github.com/autogluon/autogluon/pull/5784))
- TabDPT: default to sequential fold fitting. [@Innixma](https://github.com/Innixma) ([#5785](https://github.com/autogluon/autogluon/pull/5785))
- Minor changes for expert/experimental usage. [@LennartPurucker](https://github.com/LennartPurucker) ([#5629](https://github.com/autogluon/autogluon/pull/5629))

### API Changes and Deprecations
- Make params immutable after construction (deprecation, raises in 1.7). [@Innixma](https://github.com/Innixma) ([#5776](https://github.com/autogluon/autogluon/pull/5776))
- `AuxiliaryParams` typed schema; make `params_aux` immutable after construction (deprecation, raises in 1.7). [@Innixma](https://github.com/Innixma) ([#5775](https://github.com/autogluon/autogluon/pull/5775))
- Declare supported problem types via `_supported_problem_types` class attribute. [@Innixma](https://github.com/Innixma) ([#5771](https://github.com/autogluon/autogluon/pull/5771))
- Declare auxiliary-param overrides via `_default_auxiliary_params_extra` class attribute. [@Innixma](https://github.com/Innixma) ([#5772](https://github.com/autogluon/autogluon/pull/5772))
- Declare ensemble-args overrides and fit-resource defaults as class attributes. [@Innixma](https://github.com/Innixma) ([#5777](https://github.com/autogluon/autogluon/pull/5777))
- Validate `_default_auxiliary_params_extra` keys in `verify_model`. [@Innixma](https://github.com/Innixma) ([#5773](https://github.com/autogluon/autogluon/pull/5773))

### Fixes and Improvements
- `__init__.py` is missing `TabPFNV2Model`. [@zhaozhiqiang200207](https://github.com/zhaozhiqiang200207) ([#5511](https://github.com/autogluon/autogluon/pull/5511))
- Fix flash attention check for TabDPT. [@colesussmeier](https://github.com/colesussmeier) ([#5516](https://github.com/autogluon/autogluon/pull/5516))
- Refactor TabularNeuralNetTorch to use `state_dict` for persistence. [@celestinoxp](https://github.com/celestinoxp) ([#5485](https://github.com/autogluon/autogluon/pull/5485))
- Share the categorical label-encoding preprocess step across models. [@Innixma](https://github.com/Innixma) ([#5778](https://github.com/autogluon/autogluon/pull/5778))
- Share the `_fit` device-resolution block on `AbstractTorchModel`. [@Innixma](https://github.com/Innixma) ([#5780](https://github.com/autogluon/autogluon/pull/5780))
- Fix TabPrep running the feature generator twice. [@Innixma](https://github.com/Innixma) ([#5650](https://github.com/autogluon/autogluon/pull/5650))
- Fix LightGBM feature name collision. [@Innixma](https://github.com/Innixma) ([#5643](https://github.com/autogluon/autogluon/pull/5643))
- Disable text and datetime feature types in unrelated preprocessors. [@Innixma](https://github.com/Innixma) ([#5644](https://github.com/autogluon/autogluon/pull/5644))
- Small fixes to AutoGluon code from TabArena benchmarks. [@LennartPurucker](https://github.com/LennartPurucker) ([#5668](https://github.com/autogluon/autogluon/pull/5668))
- Workaround for not enough samples per class. [@LennartPurucker](https://github.com/LennartPurucker) ([#5613](https://github.com/autogluon/autogluon/pull/5613))
- Add tests and a sanity check for custom split. [@LennartPurucker](https://github.com/LennartPurucker) ([#5633](https://github.com/autogluon/autogluon/pull/5633))
- Allow fit helper to use large test datasets. [@LennartPurucker](https://github.com/LennartPurucker) ([#5548](https://github.com/autogluon/autogluon/pull/5548))
- Validate invalid `feature_generator` inputs early. [@bluehyena](https://github.com/bluehyena) ([#5635](https://github.com/autogluon/autogluon/pull/5635))
- Silence spurious warnings on degenerate (constant / all-NaN) columns during fit. [@Innixma](https://github.com/Innixma) ([#5701](https://github.com/autogluon/autogluon/pull/5701))
- Fix IndexError in GroupByFeatureGenerator on all-NaN categorical columns. [@Innixma](https://github.com/Innixma) ([#5700](https://github.com/autogluon/autogluon/pull/5700))
- Fix typos and bugs in feature generators. [@celestinoxp](https://github.com/celestinoxp) ([#5577](https://github.com/autogluon/autogluon/pull/5577))
- Fix mutable default argument in DatetimeFeatureGenerator. [@celestinoxp](https://github.com/celestinoxp) ([#5558](https://github.com/autogluon/autogluon/pull/5558))
- Fix incorrect NaN comparisons in feature importance logic. [@celestinoxp](https://github.com/celestinoxp) ([#5555](https://github.com/autogluon/autogluon/pull/5555))
- Fix feature importance p-value calculation and optimize batch size. [@celestinoxp](https://github.com/celestinoxp) ([#5584](https://github.com/autogluon/autogluon/pull/5584))
- Refactor greedy ensemble selection to use `np.isclose` for floating point comparisons. [@celestinoxp](https://github.com/celestinoxp) ([#5556](https://github.com/autogluon/autogluon/pull/5556))
- EnsembleSelection hygiene: no caller mutation, predict `problem_type` override. [@Innixma](https://github.com/Innixma) ([#5720](https://github.com/autogluon/autogluon/pull/5720))
- Fix OOF alignment for the weighted ensemble; size curves override `auto_stack` and gain dynamic stacking. [@Innixma](https://github.com/Innixma) ([#5804](https://github.com/autogluon/autogluon/pull/5804))
- Resolve the structure holdout before feature generation. [@Innixma](https://github.com/Innixma) ([#5799](https://github.com/autogluon/autogluon/pull/5799))
- Disable `use_child_oof` when `custom_splits` are given. [@Innixma](https://github.com/Innixma) ([#5795](https://github.com/autogluon/autogluon/pull/5795))
- Do not give up the requested fold count over a rare stratification value. [@Innixma](https://github.com/Innixma) ([#5806](https://github.com/autogluon/autogluon/pull/5806))
- Stop losing missing categorical values in the TabPFN wrappers. [@Innixma](https://github.com/Innixma) ([#5809](https://github.com/autogluon/autogluon/pull/5809))
- Keep missing values and lift the class cap for TabDPT. [@Innixma](https://github.com/Innixma) ([#5810](https://github.com/autogluon/autogluon/pull/5810))
- Optimize `augment_rare_classes` with single concat. [@celestinoxp](https://github.com/celestinoxp) ([#5564](https://github.com/autogluon/autogluon/pull/5564))
- Optimize category counting in TabularTorchDataset using numpy. [@celestinoxp](https://github.com/celestinoxp) ([#5566](https://github.com/autogluon/autogluon/pull/5566))
- Support `np.ndarray` in cv_splitter. [@Innixma](https://github.com/Innixma) ([#5538](https://github.com/autogluon/autogluon/pull/5538))
- Support `pd.SparseArray` in simulation_utils.py. [@Innixma](https://github.com/Innixma) ([#5602](https://github.com/autogluon/autogluon/pull/5602))
- Add check that models can predict single samples. [@Innixma](https://github.com/Innixma) ([#5542](https://github.com/autogluon/autogluon/pull/5542))
- Verify unmodified data outer context. [@Innixma](https://github.com/Innixma) ([#5600](https://github.com/autogluon/autogluon/pull/5600))
- Safe rmtree for dystack on Windows. [@celestinoxp](https://github.com/celestinoxp) ([#5616](https://github.com/autogluon/autogluon/pull/5616))
- Fix misleading `augment_method` in `distill()` log. [@petervanwylen](https://github.com/petervanwylen) ([#5656](https://github.com/autogluon/autogluon/pull/5656))
- Delete redundant debug logging. [@Innixma](https://github.com/Innixma) ([#5525](https://github.com/autogluon/autogluon/pull/5525))
- Add test_augment_rare_classes.py. [@Innixma](https://github.com/Innixma) ([#5596](https://github.com/autogluon/autogluon/pull/5596))
- Add test_pandas_utils.py. [@Innixma](https://github.com/Innixma) ([#5597](https://github.com/autogluon/autogluon/pull/5597))

--------

## TimeSeries

<!-- TODO: module intro paragraph -->

### Highlights
- Add Toto-2 pretrained forecasting model. [@shchur](https://github.com/shchur) ([#5740](https://github.com/autogluon/autogluon/pull/5740), [#5760](https://github.com/autogluon/autogluon/pull/5760))
- New forecasting metrics: `MAEB` and `WAPEB` (bias-penalized), `BIAS` (to check for forecast bias), and `MQL` (mean quantile loss). [@shchur](https://github.com/shchur) ([#5782](https://github.com/autogluon/autogluon/pull/5782), [#5779](https://github.com/autogluon/autogluon/pull/5779))
- New `TimeSeriesPredictor.export_model` method to export a trained model to a standalone checkpoint that loads without AutoGluon. [@shchur](https://github.com/shchur) ([#5767](https://github.com/autogluon/autogluon/pull/5767))
- New presets, achieving a 65% win rate vs. AutoGluon v1.5. [@shchur](https://github.com/shchur) ([#5786](https://github.com/autogluon/autogluon/pull/5786), [#5811](https://github.com/autogluon/autogluon/pull/5811))

### New Features
- [experimental] Add `TimeSeriesPredictor.update()` for ensemble re-selection on fresh data. [@shchur](https://github.com/shchur) ([#5803](https://github.com/autogluon/autogluon/pull/5803))

### API Changes and Deprecations
- Prediction caching in `TimeSeriesPredictor` is now disabled by default and deprecated; it will be removed entirely in a future release. [@shchur](https://github.com/shchur) ([#5761](https://github.com/autogluon/autogluon/pull/5761))

### Fixes and Improvements
- Use new Chronos-2 preprocess API for fine-tuning. [@shchur](https://github.com/shchur) ([#5739](https://github.com/autogluon/autogluon/pull/5739))
- Respect user-provided device for Chronos models. [@shchur](https://github.com/shchur) ([#5764](https://github.com/autogluon/autogluon/pull/5764))
- Align timeseries `eval_metric` format with tabular. [@Mateusz-Switala](https://github.com/Mateusz-Switala) ([#5661](https://github.com/autogluon/autogluon/pull/5661))
- Fix Chronos2 covariate unit tests for chronos-forecasting 2.3.1. [@shchur](https://github.com/shchur) ([#5706](https://github.com/autogluon/autogluon/pull/5706))
- Silence noisy model loading logs for Chronos and TFT. [@shchur](https://github.com/shchur) ([#5805](https://github.com/autogluon/autogluon/pull/5805), [#5807](https://github.com/autogluon/autogluon/pull/5807))

--------

## Multimodal

### Fixes and Improvements
- Add missing image backbones to `hf_model_list.yaml`. [@celestinoxp](https://github.com/celestinoxp) ([#5591](https://github.com/autogluon/autogluon/pull/5591))
- Update `semantic_segmentation.py`. [@eyh002](https://github.com/eyh002) ([#5762](https://github.com/autogluon/autogluon/pull/5762))
- Fix multimodal tests. [@shchur](https://github.com/shchur) ([#5718](https://github.com/autogluon/autogluon/pull/5718))

--------

## Documentation and CI
- Documentation fixes and improvements. [@Innixma](https://github.com/Innixma) [@shchur](https://github.com/shchur) [@biefan](https://github.com/biefan) [@jwu7778](https://github.com/jwu7778) [@celestinoxp](https://github.com/celestinoxp) [@MaximAliev](https://github.com/MaximAliev) [@felixkosmalla](https://github.com/felixkosmalla) ([#5534](https://github.com/autogluon/autogluon/pull/5534), [#5667](https://github.com/autogluon/autogluon/pull/5667), [#5686](https://github.com/autogluon/autogluon/pull/5686), [#5612](https://github.com/autogluon/autogluon/pull/5612), [#5621](https://github.com/autogluon/autogluon/pull/5621), [#5174](https://github.com/autogluon/autogluon/pull/5174), [#5574](https://github.com/autogluon/autogluon/pull/5574), [#5575](https://github.com/autogluon/autogluon/pull/5575), [#4709](https://github.com/autogluon/autogluon/pull/4709), [#5554](https://github.com/autogluon/autogluon/pull/5554), [#5813](https://github.com/autogluon/autogluon/pull/5813), [#5814](https://github.com/autogluon/autogluon/pull/5814))
- CI, linting, and tooling fixes and improvements. [@shchur](https://github.com/shchur) [@celestinoxp](https://github.com/celestinoxp) [@Innixma](https://github.com/Innixma) [@prateekdesai04](https://github.com/prateekdesai04) [@github-actions](https://github.com/features/actions) ([#5531](https://github.com/autogluon/autogluon/pull/5531), [#5532](https://github.com/autogluon/autogluon/pull/5532), [#5530](https://github.com/autogluon/autogluon/pull/5530), [#5543](https://github.com/autogluon/autogluon/pull/5543), [#5559](https://github.com/autogluon/autogluon/pull/5559), [#5592](https://github.com/autogluon/autogluon/pull/5592), [#5585](https://github.com/autogluon/autogluon/pull/5585), [#5758](https://github.com/autogluon/autogluon/pull/5758), [#5572](https://github.com/autogluon/autogluon/pull/5572), [#5669](https://github.com/autogluon/autogluon/pull/5669), [#5672](https://github.com/autogluon/autogluon/pull/5672), [#5586](https://github.com/autogluon/autogluon/pull/5586), [#5674](https://github.com/autogluon/autogluon/pull/5674), [#5728](https://github.com/autogluon/autogluon/pull/5728), [#5736](https://github.com/autogluon/autogluon/pull/5736), [#5738](https://github.com/autogluon/autogluon/pull/5738), [#5507](https://github.com/autogluon/autogluon/pull/5507))
- Auto-update ruff pre-commit hook. [@github-actions](https://github.com/features/actions) ([#5607](https://github.com/autogluon/autogluon/pull/5607), [#5610](https://github.com/autogluon/autogluon/pull/5610), [#5614](https://github.com/autogluon/autogluon/pull/5614), [#5624](https://github.com/autogluon/autogluon/pull/5624), [#5626](https://github.com/autogluon/autogluon/pull/5626), [#5630](https://github.com/autogluon/autogluon/pull/5630), [#5636](https://github.com/autogluon/autogluon/pull/5636), [#5639](https://github.com/autogluon/autogluon/pull/5639), [#5651](https://github.com/autogluon/autogluon/pull/5651), [#5659](https://github.com/autogluon/autogluon/pull/5659), [#5670](https://github.com/autogluon/autogluon/pull/5670), [#5675](https://github.com/autogluon/autogluon/pull/5675), [#5696](https://github.com/autogluon/autogluon/pull/5696), [#5699](https://github.com/autogluon/autogluon/pull/5699), [#5703](https://github.com/autogluon/autogluon/pull/5703), [#5704](https://github.com/autogluon/autogluon/pull/5704), [#5719](https://github.com/autogluon/autogluon/pull/5719), [#5733](https://github.com/autogluon/autogluon/pull/5733), [#5755](https://github.com/autogluon/autogluon/pull/5755), [#5787](https://github.com/autogluon/autogluon/pull/5787))

--------

## Contributors

Full Contributor List (ordered by # of commits):

[@Innixma](https://github.com/Innixma) [@shchur](https://github.com/shchur) [@celestinoxp](https://github.com/celestinoxp) [@LennartPurucker](https://github.com/LennartPurucker) [@Atharva-Rajan-Kale](https://github.com/Atharva-Rajan-Kale) [@prateekdesai04](https://github.com/prateekdesai04) [@jalencato](https://github.com/jalencato) [@colesussmeier](https://github.com/colesussmeier) [@atschalz](https://github.com/atschalz) [@zhaozhiqiang200207](https://github.com/zhaozhiqiang200207) [@petervanwylen](https://github.com/petervanwylen) [@minkyu-choi07](https://github.com/minkyu-choi07) [@jwu7778](https://github.com/jwu7778) [@felixkosmalla](https://github.com/felixkosmalla) [@eyh002](https://github.com/eyh002) [@bluehyena](https://github.com/bluehyena) [@biefan](https://github.com/biefan) [@MaximAliev](https://github.com/MaximAliev) [@Mateusz-Switala](https://github.com/Mateusz-Switala)

### New Contributors

[@zhaozhiqiang200207](https://github.com/zhaozhiqiang200207) [@jwu7778](https://github.com/jwu7778) [@felixkosmalla](https://github.com/felixkosmalla) [@MaximAliev](https://github.com/MaximAliev) [@biefan](https://github.com/biefan) [@bluehyena](https://github.com/bluehyena) [@Mateusz-Switala](https://github.com/Mateusz-Switala) [@jalencato](https://github.com/jalencato) [@Atharva-Rajan-Kale](https://github.com/Atharva-Rajan-Kale) [@petervanwylen](https://github.com/petervanwylen) [@minkyu-choi07](https://github.com/minkyu-choi07) [@eyh002](https://github.com/eyh002)

--------

## Acknowledgements

AutoGluon's continued development is made possible by the support of Amazon and Prior Labs. The project was created at Amazon, which has supported it since the very first release. With this release, Prior Labs joins in supporting AutoGluon's core development, and the maintainer team now spans both organizations. We are grateful to both companies for investing in open source machine learning and enabling AutoGluon to keep pushing the state of the art.