All pull requests

Pull requests

• [CI] Measure warm-cache CI performance

  #12714·RAMitchell opened Oct 13, 2026Member·

  ·
  1

• [python-package] Report the actual split thresholds in get_split_value_histogram

  #12702·vsaraikin opened Oct 11, 2026·

  ·
  1

  1

• [python-package] Escape feature names in get_split_value_histogram

  #12697·aalquwayfili opened Oct 11, 2026·

  ·
  1

• Honor shuffle in stratified cross-validation

  #12691·aswanth-07 opened Oct 11, 2026Contributor·

  ·
  1

• [python-package] Preserve infinite integer thresholds in tree dataframes

  #12690·j45856021-dev opened Oct 11, 2026·

  ·
  1

• Use context for the tensor class.

  #12680·trivialfis opened Oct 9, 2026Member·

  ·
  1

• [jvm-packages] Fix model save/load on Spark 4.x and num_class on classifier load

  #12676·anthony-khong opened Oct 8, 2026·

  ·
  1

  1

• Avoid copying pandas input when dropping the qid column.

  #12673·bhaskargurram-ai opened Oct 7, 2026·

  ·
  1

  2

• Fix indicator branch labels, missing metadata, and Graphviz colors

  #12672·parlorsky opened Oct 7, 2026·

  ·
  5

• Register SYCL hinge objective kernels through kernel dispatch

  #12662·avolkov-intel opened Oct 6, 2026·

  ·
  6

• [python-package] Do not reset the global NumPy random state in xgboost.cv

  #12651·SulimanAbdulrazzaq opened Oct 3, 2026·

  ·
  4

• Preserve AFT metric parameters across booster reloads

  #12649·goutamadwant opened Oct 3, 2026·

  ·
  1

  1

• [R] Fix verbose early stopping best iteration output

  #12646·Mohit-25-tech opened Oct 2, 2026·

  ·
  1

  1

• [R] Use isTRUE() for robust comparison in early stopping callback

  #12644·jemus42 opened Oct 1, 2026·

  ·
  1

• [pyspark] Fix empty-partition handling in sparse DMatrix construction

  #12640·DiegoGCh opened Sep 30, 2026·

  ·
  1

  1

• Add type hints to test_quantile_dmatrix.py and enable mypy check.

  #12632·ryota717 opened Sep 29, 2026·

  ·
  1

• predictor: parallelize small batches over trees

  #12625·TCoderr opened Sep 26, 2026Contributor·

  ·
  5

• predictor: use blocked prediction for sparse data with few features

  #12623·TCoderr opened Sep 25, 2026Contributor·

  ·
  1

• [R] Warn about ignored arguments in predict.xgboost

  #12616·Eliaaazzz opened Sep 25, 2026·

  ·
  1

  5

• [Python] Add type hints to config and tree regularization tests

  #12613·ankitdhandharia opened Sep 24, 2026·

  ·
  1

• predictor: accumulate leaf values after complete array traversal

  #12611·TCoderr opened Sep 24, 2026Contributor·

  ·
  1

• predictor: reuse array tree layouts across row blocks

  #12609·TCoderr opened Sep 24, 2026Contributor·

  ·
  3

• Guard max_num_features against 0 and negative values in plot_importance

  #12606·Nityahapani opened Sep 24, 2026Contributor·

  ·
  1

• Fix typos in comments and docstrings

  #12605·Nityahapani opened Sep 24, 2026Contributor·

  ·
  1

• Fix XGBClassifier.predict threshold for binary:logitraw

  #12599·Nityahapani opened Sep 23, 2026Contributor·

  ·
  7