All pull requests

Pull requests

• fix(trainer): validate cache schema and table names

  size/M
  #882·Karman580 opened Oct 19, 2026·

  ·
  1

  1

• fix(trainer): Validate CustomTrainer func_args before submitting

  size/L
  #880·avanishkasar opened Oct 19, 2026·1 task done·

  ·
  1

  1

• fix(trainer): validate status in TrainerClient.wait_for_job_status

  size/S
  #878·G26karthik opened Oct 19, 2026·1 task done·

  ·
  1

  2

• fix(docs): fix outdated .types imports in docs that cause ImportErrors

  size/S
  #876·Akshat-Raj opened Oct 19, 2026Contributor·1 task done·

  ·
  1

  2

• fix(spark): validate requested job statuses

  ok-to-testsize/M
  #874·Karman580 opened Oct 18, 2026·1 task·

  ·
  1

  4

  

• fix(trainer): read the TrainJob runtime from its snapshot

  size/L
  #872·Ahmad-Faraj opened Oct 18, 2026·1 task·

  ·
  1

  2

• feat(trainer): make list_jobs() send a constant number of API requests

  size/L
  #871·DEVJHAWAR11 opened Oct 18, 2026·1 task·

  ·
  1

  1

• fix(trainer): reject conflicting resource aliases

  size/S
  #866·notshriyansh opened Oct 13, 2026·3 of 4 tasks·

  ·
  2

• fix(optimizer): preserve Katib integer search spaces

  size/L
  #865·SahilKumar75 opened Oct 13, 2026Contributor·1 task done·

  ·
  1

  2

• chore: Add package init docstrings

  size/M
  #864·agarwal-ridhima opened Oct 13, 2026·1 task·

  ·
  1

  2

• feat(hub): allow passing async_runner to ModelRegistryClient

  size/M
  #862·myselfkunal opened Oct 12, 2026·1 task done·

  ·
  1

  2

• fix(trainer): multiply decimal-form device counts across nodes

  size/M
  #861·aaa-aashna opened Oct 12, 2026·

  ·
  1

  3

• feat: add opt-in NullHandler for root kubeflow logger

  size/M
  #859·AnKiT-GaRG2 opened Oct 11, 2026·1 of 2 tasks·

  ·
  3

• fix(spark): propagate java_options to DriverSpec and ExecutorSpec

  size/L
  #856·rushirathod22 opened Oct 11, 2026·

  ·
  1

  2

• feat(trainer): support separate resource requests and limits

  size/L
  #854·RobuRishabh opened Oct 11, 2026·4 tasks done·

  ·
  1

  2

• fix(spark): reject a FuncJob that needs arguments but has none

  ok-to-testsize/S
  #853·singhharsh1708 opened Oct 10, 2026·1 task·

  ·
  1

  10

  

• fix(spark): split followed pod logs into lines

  ok-to-testsize/L
  #851·singhharsh1708 opened Oct 10, 2026·1 task·

  ·
  1

  7

  

• fix(trainer): build container paths with POSIX separators

  lgtmok-to-testsize/M
  #849·singhharsh1708 opened Oct 10, 2026·1 task·

  ·
  1

  7

  

• feat(spark): check control plane version ConfigMap (#221)

  ok-to-testsize/L
  #846·Neha-Nimsatkar opened Oct 9, 2026Contributor·2 tasks done·

  ·
  1

  10

  

• fix(trainer): retry transient API errors when reading pod logs

  size/L
  #844·aniket-3001 opened Oct 9, 2026·1 task·

  ·
  1

  2

• fix: keep events that only set eventTime in get_job_events

  size/M
  #842·Ahmad-Faraj opened Oct 9, 2026·1 task·

  ·
  1

  2

• chore(ci): raise actionlint shellcheck severity to warning

  size/L
  #840·TensorDevLJ opened Oct 9, 2026·1 of 2 tasks·

  ·
  1

  2

• chore(ci): migrate E2E tests to cncf-ubuntu-16-64-x86 runner

  lgtmsize/XS
  #839·tariq-hasan opened Oct 9, 2026Member·1 task·

  ·
  1

  

• fix(trainer): isolate runtime metadata between local jobs

  size/M
  #838·SahilKumar75 opened Oct 8, 2026Contributor·1 task·

  ·
  2

• fix(optimizer): pass hyperparameters to the training function with their type

  size/L
  #836·abhishekKokadwar opened Oct 8, 2026Member·1 task done·

  ·
  1

  2