Issues

• about training data of in-context editing in Figure 6

  VectorSpaceLab/OmniGen2#138 · Filter by author garychan22 opened on Nov 3, 2025
  1

• Question about num_images_per_prompt Setting in OmniContext Benchmark

  VectorSpaceLab/OmniGen2#137 · Filter by author Ryann-Ran opened on Oct 21, 2025

• OmniGen2在UniGenBench上的效果

  VectorSpaceLab/OmniGen2#136 · Filter by author CodeGoat24 opened on Sep 23, 2025

• [2img2img] Not possible to replace texture in 1.jpg with 2.jpg

  VectorSpaceLab/OmniGen2#135 · Filter by author AngelTs opened on Sep 10, 2025

• [Bug]: CUDA error: device-side assert triggered with long prompts in T2I script

  VectorSpaceLab/OmniGen2#134 · Filter by author Jaguar156 opened on Aug 31, 2025

• about video_icgen

  VectorSpaceLab/OmniGen2#133 · Filter by author LHRYANG opened on Aug 29, 2025

• question about noise_scheduler

  VectorSpaceLab/OmniGen2#132 · Filter by author hrz2000 opened on Aug 21, 2025

• Inquiry about training compute setup

  VectorSpaceLab/OmniGen2#131 · Filter by author Jaguar156 opened on Aug 19, 2025
  1

• edit_pf_one data still broken after Aug 13 update

  VectorSpaceLab/OmniGen2#130 · Filter by author jacklishufan opened on Aug 14, 2025

• 输入图是否过了vit？

  VectorSpaceLab/OmniGen2#129 · Filter by author hrz2000 opened on Aug 9, 2025
  5

• 数据集中多图输入的部分

  VectorSpaceLab/OmniGen2#128 · Filter by author hrz2000 opened on Aug 9, 2025

• Any plan for distilled version of model?

  VectorSpaceLab/OmniGen2#126 · Filter by author ChunChenLin opened on Aug 7, 2025

• For interleaved video data, what is the time interval between every two frames in video clips?

  VectorSpaceLab/OmniGen2#125 · Filter by author LHRYANG opened on Aug 6, 2025
  1

• Number of T2I pairs

  VectorSpaceLab/OmniGen2#124 · Filter by author dongbo1998 opened on Aug 2, 2025
  1

• 现有代码不支持生成文本的训练。

  VectorSpaceLab/OmniGen2#123 · Filter by author galactic123 opened on Jul 31, 2025

• getting error saying that get the hf login token for flux.1-dev repo as its is a gated repository, not for commercial purpose

  VectorSpaceLab/OmniGen2#121 · Filter by author hereiamravi opened on Jul 27, 2025

• How to finetune for Visual Understanding tasks?

  VectorSpaceLab/OmniGen2#120 · Filter by author zhangrongxiang opened on Jul 25, 2025

• No hidden text for image generation?

  VectorSpaceLab/OmniGen2#118 · Filter by author quantumiracle opened on Jul 22, 2025

• When will the training data construction pipeline be open-sourced?

  VectorSpaceLab/OmniGen2#117 · Filter by author nini0919 opened on Jul 22, 2025

• How was the mllm fine tuned

  VectorSpaceLab/OmniGen2#115 · Filter by author rishabh063 opened on Jul 20, 2025

• 训练生成后的图有些模糊

  VectorSpaceLab/OmniGen2#114 · Filter by author cjymz886 opened on Jul 19, 2025
  4

• /pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:144: operator(): …… Assertion idx_dim >= 0 && idx_dim < index_size && "index out of bounds" failed.

  VectorSpaceLab/OmniGen2#112 · Filter by author cb1cyf opened on Jul 16, 2025

• How to load LoRA weights in Omnigen2ChatPipeline?

  VectorSpaceLab/OmniGen2#111 · Filter by author Vinayak2104 opened on Jul 15, 2025
  1

• Room Staging & Redesigning

  VectorSpaceLab/OmniGen2#110 · Filter by author MohamedAziz-Khezami opened on Jul 15, 2025

• Changes in training code to send both reference images and instruction to MLLM

  VectorSpaceLab/OmniGen2#109 · Filter by author Vinayak2104 opened on Jul 15, 2025