Issues

• Verify evals on Papers with Code

  HumanMLLM/HumanOmniV2#18 · Filter by author NielsRogge opened on Sep 14, 2026

• A question about the frame and FPS settings in your paper.

  HumanMLLM/HumanOmniV2#17 · Filter by author wwxs123W opened on Dec 23, 2025
  1

• Why not show your model's results on OmniBench?

  HumanMLLM/HumanOmniV2#15 · Filter by author gengenkai opened on Sep 10, 2025
  1

• Request for Preprocessed qa_think.json Files for Daily-Omni and WorldSense

  HumanMLLM/HumanOmniV2#14 · Filter by author Ruize-Ma opened on Aug 17, 2025

• Any code for running a demo？

  HumanMLLM/HumanOmniV2#13 · Filter by author King-king424 opened on Jul 31, 2025

• How to reproduce Qwen2.5-Omni-7B-Thinker's 64.20% Result on IntentBench

  HumanMLLM/HumanOmniV2#12 · Filter by author shouzhoupeng opened on Jul 22, 2025
  5

• data_config/rollout_stage1.json？

  HumanMLLM/HumanOmniV2#11 · Filter by author shiyongde opened on Jul 18, 2025
  5

• Why don't we need to set the user part of the labels to -100 in sft.py?

  HumanMLLM/HumanOmniV2#10 · Filter by author wyclike opened on Jul 16, 2025
  1

• Working with VLLM

  HumanMLLM/HumanOmniV2#6 · Filter by author qq31415926 opened on Jul 12, 2025
  1

• table2

  HumanMLLM/HumanOmniV2#4 · Filter by author lijain opened on Jul 10, 2025
  1

• How much GPU resources are needed to reproduce this work？

  HumanMLLM/HumanOmniV2#3 · Filter by author vhzy opened on Jul 8, 2025
  1

• model is broken with vllm

  HumanMLLM/HumanOmniV2#2 · Filter by author AlexiaJM opened on Jul 2, 2025
  3