Commits

Branch selector

 main

User selector

All users

All time

Commit History

Commits on Apr 23, 2026

• Relax numba pin from >=0.62.0 to >=0.61.2 (#702)

  Show description for 73ecf57
  Maxusmusti

  authored
  73ecf57

  Copy full SHA for 73ecf57

Commits on Apr 20, 2026

• build(deps): Bump aws-actions/configure-aws-credentials (#689)

  Show description for 2c93507
  dependabot[bot]

  authored
  2c93507

  Copy full SHA for 2c93507

• build(deps): Bump astral-sh/setup-uv from 4 to 7 (#692)

  Show description for 80d35c6
  dependabot[bot]

  authored
  80d35c6

  Copy full SHA for 80d35c6

• Add on-demand full-state checkpointing for OpenShift AI / KubeFlow preemption (#686)

  Show description for e2cad6c
  

  RobotSail

  and

  claude

  authored
  e2cad6c

  Copy full SHA for e2cad6c

Commits on Mar 31, 2026

• Fix trust_remote_code and gradient checkpointing for custom models (#696)

  Show description for 091c21e
  

  3 peopleauthored
  091c21e

  Copy full SHA for 091c21e

Commits on Mar 24, 2026

• Fix Gemma 3 SFT training by detecting dual-registered VLM configs (#695)

  Show description for 75425ca
  RobotSail

  authored
  75425ca

  Copy full SHA for 75425ca

Commits on Mar 6, 2026

• add support for qwen3.5 vl model (#693)

  Show description for 0c6614a
  

  RobotSail

  and

  claude

  authored
  0c6614a

  Copy full SHA for 0c6614a

Commits on Feb 26, 2026

• Adds Validation Adds validation loss + exposes it in the API (#685)

  Show description for 1f02ea6
  RobotSail

  authored
  1f02ea6

  Copy full SHA for 1f02ea6

Commits on Feb 25, 2026

• Add backwards compatibility for transformers v4.57 (#684)

  Show description for 320b794
  Maxusmusti

  authored
  320b794

  Copy full SHA for 320b794

Commits on Feb 11, 2026

• fix _no_split_modules subscript error for transformers v5 (#683)

  Maxusmusti

  authored
  c517712

  Copy full SHA for c517712

Commits on Feb 5, 2026

• Add MLflow support and expose logging configuration in TrainingArgs (#680)

  Show description for 0c47c97
  

  RobotSail

  and

  claude

  authored
  0c47c97

  Copy full SHA for 0c47c97

Commits on Feb 3, 2026

• Fix HYBRID_SHARD failure when world_size < available GPUs (#682)

  Show description for aa9d705
  rtj1

  authored
  aa9d705

  Copy full SHA for aa9d705

Commits on Jan 30, 2026

• Add transformers v5 compatibility (#681)

  Maxusmusti

  authored
  c0d93d2

  Copy full SHA for c0d93d2

Commits on Jan 27, 2026

• adds integration for tox-uv and updates workflows to use tox-uv (#676)

  Show description for f170a49
  

  RobotSail

  and

  claude

  authored
  f170a49

  Copy full SHA for f170a49

Commits on Jan 8, 2026

• Exposes API for processing pretraining data (#672)

  Show description for 574f946
  RobotSail

  authored
  574f946

  Copy full SHA for 574f946

Commits on Dec 18, 2025

• fixes bug where process isn't completed by the time the process gets read (#675)

  RobotSail

  authored
  638a753

  Copy full SHA for 638a753

• Expose AdamW optimizer parameters in training API (#674)

  Show description for c495035
  

  RobotSail

  and

  claude

  authored
  c495035

  Copy full SHA for c495035

Commits on Dec 5, 2025

• Handle granite 4 as MoE models in training (#669)

  Show description for 3d05302
  mtake

  authored
  3d05302

  Copy full SHA for 3d05302

Commits on Nov 18, 2025

• fixes stray invalid variable access bug (#668)

  RobotSail

  authored
  781c36f

  Copy full SHA for 781c36f

• bumps hynek build dep (#670)

  RobotSail

  authored
  529c2f7

  Copy full SHA for 529c2f7

Commits on Oct 14, 2025

• fix(torchrun): Omit empty arguments and correct nproc_per_node type (#661)

  Show description for 637afae
  

  szaher

  and

  RobotSail

  authored
  637afae

  Copy full SHA for 637afae

Commits on Oct 13, 2025

• Update requirements.txt

  Show description for 2c8d676
  RobotSail

  committed
  2c8d676

  Copy full SHA for 2c8d676

• bump min constraint of llvmlite

  RobotSail

  committed
  386c791

  Copy full SHA for 386c791

• update constraints too

  RobotSail

  committed
  410b0f7

  Copy full SHA for 410b0f7

• set the minimum version of numba to the first release which effectively uncaps numpy

  Show description for 198e13e
  RobotSail

  committed
  198e13e

  Copy full SHA for 198e13e

• removes version cap on numpy

  RobotSail

  committed
  59c1646

  Copy full SHA for 59c1646

Commits on Oct 8, 2025

• add mamba-ssm[causal-conv1d] to cuda requirements (#663)

  RobotSail

  authored
  04fe3f5

  Copy full SHA for 04fe3f5

Commits on Sep 23, 2025

• Update requirements-cuda.txt to increase liger-kernel minimum (#659)

  Show description for cb644b9
  Maxusmusti

  authored
  cb644b9

  Copy full SHA for cb644b9

Commits on Sep 17, 2025

• Add kernels>0.9.0 to CUDA requirements (#658)

  Show description for 536ebfb
  Maxusmusti

  authored
  536ebfb

  Copy full SHA for 536ebfb

• Update numpy from <2.0 to <2.3 (#656)

  Show description for 8f0c6ed
  Maxusmusti

  authored
  8f0c6ed

  Copy full SHA for 8f0c6ed

• Adding GPT OSS Support (#646)

  Show description for 4934f4c
  

  3 peopleauthored
  4934f4c

  Copy full SHA for 4934f4c

Commits on Sep 5, 2025

• Merge pull request #648 from instructlab/update-constraints

  Show description for 2a02acc
  thisisatharva-rh

  authored
  2a02acc

  Copy full SHA for 2a02acc

Commits on Sep 1, 2025

• chore: update constraints-dev.txt (2025-09-01)

  Show description for d476e88
  

  mergify[bot]

  authored and

  github-actions[bot]

  committed
  d476e88

  Copy full SHA for d476e88

Commits on Aug 25, 2025

• Merge pull request #647 from instructlab/update-constraints

  Show description for 5722168
  thisisatharva-rh

  authored
  5722168

  Copy full SHA for 5722168

• chore: update constraints-dev.txt (2025-08-25)

  Show description for f8dc7aa
  

  mergify[bot]

  authored and

  github-actions[bot]

  committed
  f8dc7aa

  Copy full SHA for f8dc7aa