Issues

• Use dtype instead of deprecated torch_dtype for transformers >= 4.56

  EvolvingLMMs-Lab/LLaVA-OneVision-2#197 · Filter by author xyf5432 opened on Aug 28, 2026

• Verify evals on Papers with Code

  EvolvingLMMs-Lab/LLaVA-OneVision-2#193 · Filter by author NielsRogge opened on Jul 22, 2026

• Discrepancy between paper's 4-stage training curriculum and the training scripts/data configs in the repo

  EvolvingLMMs-Lab/LLaVA-OneVision-2#187 · Filter by author jd730 opened on Jul 9, 2026

• Where can I find the codec-aligned video preprocessing code?

  EvolvingLMMs-Lab/LLaVA-OneVision-2#184 · Filter by author JerryPW opened on Jun 5, 2026
  2

• Question about codec preprocessing internals and source availability

  EvolvingLMMs-Lab/LLaVA-OneVision-2#182 · Filter by author pfeibai opened on Jun 1, 2026
  2

• [Data Release] Missing split_180s.jsonl, split_gt10min.jsonl and corresponding video tars in HuggingFace dataset

  EvolvingLMMs-Lab/LLaVA-OneVision-2#176 · Filter by author liaoyifei1998 opened on May 28, 2026
  1

• Is there a Thinking/RL variant of LLaVA-OneVision-2?

  EvolvingLMMs-Lab/LLaVA-OneVision-2#172 · Filter by author pengzhansun opened on May 28, 2026

• Release LLaVA-OneVision-2 artifacts (models, dataset) on Hugging Face

  EvolvingLMMs-Lab/LLaVA-OneVision-2#171 · Filter by author NielsRogge opened on May 27, 2026

• Question about fine-tuning LLaVA-OneVision-2

  EvolvingLMMs-Lab/LLaVA-OneVision-2#163 · Filter by author wnn2000 opened on May 25, 2026

• Are there plans to release long video data?

  EvolvingLMMs-Lab/LLaVA-OneVision-2#146 · Filter by author BoweiPu opened on May 18, 2026
  1

• Missing checkpoint and merge script for LLaVA-OneVision-2-stage-0 (initial unaligned model)

  EvolvingLMMs-Lab/LLaVA-OneVision-2#145 · Filter by author foamliu opened on May 18, 2026

• Hyper-parameters used for reproducibility

  EvolvingLMMs-Lab/LLaVA-OneVision-2#109 · Filter by author dmgcsilva opened on Apr 24, 2026
  1

• [Compatibility] LLaVA-OneVision-1.5-8B-Instruct cannot run on transformers>=5.0.0

  EvolvingLMMs-Lab/LLaVA-OneVision-2#108 · Filter by author zhonghuajiang opened on Apr 17, 2026
  1

• The datasets library does not support loading this llava-next-780k-webdataset.

  EvolvingLMMs-Lab/LLaVA-OneVision-2#107 · Filter by author qq1243196045 opened on Apr 9, 2026

• 环境配置

  EvolvingLMMs-Lab/LLaVA-OneVision-2#106 · Filter by author sihany077 opened on Apr 9, 2026

• How were OCRBenchV2, MMStar and LiveXivVQA evaluated with lmms-eval 0.2.0? (Table 2 reproduction)

  EvolvingLMMs-Lab/LLaVA-OneVision-2#105 · Filter by author yjb6 opened on Apr 1, 2026

• 关于mid-train的caption

  EvolvingLMMs-Lab/LLaVA-OneVision-2#97 · Filter by author xnsfqcmm opened on Jan 19, 2026

• "No space left on device" error during offline data packing due to excessive file count

  EvolvingLMMs-Lab/LLaVA-OneVision-2#96 · Filter by author cskang0205 opened on Jan 16, 2026

• Fail to reproduce results in Figure 4

  EvolvingLMMs-Lab/LLaVA-OneVision-2#91 · Filter by author PhoenixGS opened on Jan 5, 2026
  8

• Cannot use apply_chat_template because this processor does not have a chat template.

  EvolvingLMMs-Lab/LLaVA-OneVision-2#90 · Filter by author zitai-wang opened on Jan 5, 2026
  1

• Instruction prompt used in the mid stage

  EvolvingLMMs-Lab/LLaVA-OneVision-2#89 · Filter by author trolommonm opened on Jan 2, 2026

• How to change attention backend

  EvolvingLMMs-Lab/LLaVA-OneVision-2#85 · Filter by author peterwisu opened on Dec 23, 2025

• Maximum sequence length

  EvolvingLMMs-Lab/LLaVA-OneVision-2#84 · Filter by author trolommonm opened on Dec 19, 2025
  3

• Question on training script for quick start

  EvolvingLMMs-Lab/LLaVA-OneVision-2#83 · Filter by author peterwisu opened on Dec 19, 2025
  1

• Video Input Support

  EvolvingLMMs-Lab/LLaVA-OneVision-2#82 · Filter by author CharlesGong12 opened on Dec 17, 2025
  7