Issues

• [serge] integration failure triage - 2026-11-12

  huggingface/transformers#50683 · Filter by author github-actions opened on Nov 12, 2026
  

• Trainer: after resume_from_checkpoint=<older>, a crash while re-saving checkpoint-N can silently mix two runs

  huggingface/transformers#50677 · Filter by author asonnino opened on Nov 12, 2026

• Multimodal pipelines crash on string assistant prefill with return_full_text=True

  bug
  huggingface/transformers#50676 · Filter by author BlazeSR10hawk opened on Nov 12, 2026

• Additional tokenizer normalization regressions in v5

  huggingface/transformers#50674 · Filter by author lotem-factorial opened on Nov 12, 2026

• [Bug] Concurrent chat templating silently loses assistant generation masks due to shared tracker state

  bug
  huggingface/transformers#50669 · Filter by author BlazeSR10hawk opened on Nov 12, 2026
  2

• [serve] OpenAI-valid frequency_penalty values return 500 instead of a response or 400

  bug
  huggingface/transformers#50657 · Filter by author fruitstash opened on Nov 12, 2026

• JetMoE: attention router logits are counted twice in the auxiliary load-balancing loss

  huggingface/transformers#50656 · Filter by author taking-lying-flat opened on Nov 12, 2026

• Chat response parsing: dropped closing captures, premature streaming $ matches, and JSON string corruption

  bug
  huggingface/transformers#50647 · Filter by author nico-martin opened on Nov 12, 2026
  4

• Add a PIL backend for Gemma4UnifiedImageProcessor (Gemma 4 12B)

  Feature request
  huggingface/transformers#50639 · Filter by author jjbuck opened on Nov 12, 2026
  1

• [Bug] AutoProcessor.from_pretrained(..., backend="pil") fails when the video processor has no PIL backend

  bug
  huggingface/transformers#50638 · Filter by author jjbuck opened on Nov 12, 2026
  1

• LayoutLMv2/v3: relative position bias tables get no gradient since #30988 (torch.no_grad around the bias lookup)

  huggingface/transformers#50636 · Filter by author 14H034160212 opened on Nov 12, 2026
  2

• Migrate GPT-J to the attention interface (SDPA / flex attention support)

  Feature request
  huggingface/transformers#50627 · Filter by author xquantize opened on Nov 11, 2026
  2

• [Voxtral/pipelines] Assistant prefill fails because continue_final_message is dropped across processor and pipeline boundaries

  bug
  huggingface/transformers#50623 · Filter by author BlazeSR10hawk opened on Nov 11, 2026

• Mistral Large 4: LeRoPE alpha is silently frozen during full-parameter training

  huggingface/transformers#50622 · Filter by author taking-lying-flat opened on Nov 11, 2026
  2

• SmolVLM: built-in video chat template drops non-empty string message content

  huggingface/transformers#50613 · Filter by author arjhinety opened on Nov 11, 2026
  3

• TimesFM 2.5: training loss feeds the point-forecast channel to the quantile loss

  bug
  huggingface/transformers#50612 · Filter by author m1o0 opened on Nov 11, 2026
  

• finegrained_fp8 models are silently mis-dequantized when loaded on CPU if a quantized weight has partial blocks

  bug
  huggingface/transformers#50603 · Filter by author YancyLii opened on Nov 11, 2026
  4

• Default configs for 11 models can't build or run their own model (CohereCompass, DiffusionGemma, DeepseekOcr2, Granite4Vision, HunYuanVL, Moshi, PPOCRV6MediumDet, Qwen4Exp, Qwen3ASR, Molmo2, Step3p7)

  huggingface/transformers#50601 · Filter by author hameerabbasi opened on Nov 11, 2026
  5

• Universal speculative decoding with a pruned assistant LM head can map logits to the wrong target tokens

  huggingface/transformers#50598 · Filter by author Incheonkirin opened on Nov 11, 2026
  1

• FineGrainedFP8 dequantize with dtype=float32 leaves non-quantized weights in bfloat16

  huggingface/transformers#50594 · Filter by author qgallouedec opened on Nov 10, 2026

• center_crop pads the odd leftover pixel on opposite sides in the torchvision vs numpy/PIL backends

  huggingface/transformers#50584 · Filter by author SID-6921 opened on Nov 10, 2026
  4

• v5: full fine-tuning a bf16 checkpoint with Trainer silently runs pure bf16 AdamW, and most updates round away

  huggingface/transformers#50581 · Filter by author Andreas-Hum opened on Nov 10, 2026
  2

• Docs: dead links in model_doc pages (12 URLs return 404)

  huggingface/transformers#50568 · Filter by author Hisernberg opened on Nov 10, 2026
  1

• FSDP checkpoint resumption deadlocks during collective optimizer state sync with PEFT LoRA in Trainer

  bug
  huggingface/transformers#50549 · Filter by author utkarshxdev opened on Nov 10, 2026
  2

• MoE load-balancing loss adds no gradient under reentrant gradient checkpointing

  huggingface/transformers#50545 · Filter by author hivaze opened on Nov 10, 2026
  6