All pull requests

Pull requests

• Add Intel generation logic for Intel notebooks

  new-feature#regresses#>3000#
  #408·JoshuaL3000 opened Oct 19, 2026·

  ·

• AMD: keep the git-main unsloth install in the Qwen3 (4B) GRPO tool-use notebook

  edge case#
  #405·NoahSabb opened Oct 18, 2026·

  ·
  1

• Correct the distillation config table: gemma-4 does not fit two T4s

  edge case#
  #404·danielhanchen opened Oct 17, 2026Member·

  ·
  4

• Bump the actions group with 2 updates

  dependenciesgithub_actions
  #401·dependabot[bot] opened Oct 15, 2026Bot·

  ·

• Gemma 4 Text notebooks: keep <bos> in the training text

  conflicts#important#>3000#
  #383·danielhanchen opened Oct 6, 2026Member·

  ·
  1

  7

• fix(diffusiongemma): use decoder_input_ids instead of canvas_ids in Sudoku training loop

  conflicts#superseded#
  #311·InfoSage05 opened Jul 28, 2026·

  ·
  1

• Fix Gemma 4 audio inference dtype mismatch

  refactor#
  #308·Auro-rium opened Jul 27, 2026Contributor·

  ·

• Add SmolLM3 (3B) Conversational finetuning notebook

  conflicts#new-feature#
  #306·gowtham-sai-yadav opened Jul 22, 2026·

  ·
  10

• Replace standalone Studio wording with Unsloth

  conflicts#refactor#
  #302·shimmyshimmer opened Jul 18, 2026Member·

  ·
  2

• AMD notebooks: install ROCm-compatible bitsandbytes on all 4-bit notebooks to fix 4-bit NaN

  refactor#>3000#
  #300·LeoBorcherding opened Jul 7, 2026Contributor·

  ·
  6

• notebooks: retarget AMD notebooks to AMD Dev Cloud

  conflicts#new-feature#refactor#regresses#>3000#
  #298·Etherll opened Jul 3, 2026Collaborator·

  ·
  15

• Fix LoRA reload path in Gemma4 and Qwen3.5 GRPO notebooks

  refactor#superseded#
  #291·danielhanchen opened Jun 13, 2026Member·

  ·
  6

• Whisper: pass processing_class instead of removed tokenizer kwarg

  edge case#
  #290·danielhanchen opened Jun 12, 2026Member·

  ·
  6

• fix(Qwen3-VL-8B): replace deprecated tokenizer arg in SFTTrainer with processing_class and set eos_token in SFTConfig

  #253·Saad1926Q opened Apr 24, 2026·1 task done·

  ·
  9

• Add image token budget example to Gemma 4 vision notebook

  conflicts#new-feature#refactor#>3000#
  #250·siddhantoon opened Apr 18, 2026·

  ·
  15

• Fix Gemma3N audio notebooks to use reentrant checkpointing

  superseded#
  #215·danielhanchen opened Mar 26, 2026Member·2 tasks done·

  ·
  2

• Feature: Port and Optimize Unsloth Notebooks for Kaggle Dual T4 GPUs (T4x2)

  defer#new-feature#refactor#regresses#>3000#
  #212·AAB20 opened Mar 17, 2026·

  ·
  8

• Fix angle-bracket tags hidden by GitHub notebook renderer

  conflicts#superseded#>3000#
  #204·danielhanchen opened Mar 8, 2026Member·5 tasks done·

  ·
  5

• Fix Qwen2.5 VL GRPO prompt templating for issue #4081

  refactor#
  #197·danielhanchen opened Feb 26, 2026Member·

  ·
  3

• Fix Gemma3N notebook loading by forcing eager attention

  superseded#
  #196·danielhanchen opened Feb 25, 2026Member·

  ·
  9

• Update Whisper notebook to use AutoModelForSpeechSeq2Seq

  refactor#
  #194·danielhanchen opened Feb 25, 2026Member·

  ·
  1

• Add news section to fine-tuning notebook

  defer#
  #191·adityatiwari12 opened Feb 17, 2026·

  ·
  1

• Gemma3N audio notebooks: bound test index and subset size

  edge case#
  #190·danielhanchen opened Feb 11, 2026Member·

  ·
  4

• Remove input_ids arg from dataset filter

  conflicts#superseded#
  #188·Datta0 opened Feb 10, 2026Collaborator·

  ·
  5

• Fix two notebooks for script conversion

  conflicts#refactor#
  #186·danielhanchen opened Feb 8, 2026Member·

  ·
  4