Issues

• [Reproducibility] Training script / config for MiniMax-H3-TrainingAdapter (DeCFG adapter) itself

  modelscope/DiffSynth-Studio#1713 · Filter by author kmatzen opened on Sep 28, 2026
  1

• 麻烦问下，qwen image 2.1 如何进行图像编辑的训练？没看到相关数据集

  modelscope/DiffSynth-Studio#1709 · Filter by author whyPeanutbutter opened on Sep 24, 2026
  1

• WanToDance: hardcoded device='cuda' in music encoder construction crashes model loading on Ascend NPU

  modelscope/DiffSynth-Studio#1707 · Filter by author li-lizhe opened on Sep 23, 2026
  1

• qwen image 2.1 lora/full 训出来的模型 出图全是噪声

  modelscope/DiffSynth-Studio#1706 · Filter by author sjf18 opened on Sep 22, 2026
  6

• 怎么能使用minimax h3 turbo lora、我测试了最新的 1.2版本、输出视频都是噪声

  modelscope/DiffSynth-Studio#1691 · Filter by author 01-111162-173 opened on Sep 14, 2026

• Silently slower attention kernel when training MiniMax-H3 LoRA on B200

  modelscope/DiffSynth-Studio#1687 · Filter by author TarzanZhao opened on Sep 11, 2026

• FLUX LoRA Fine-Tuning: Border Outpainting and Pixel-Level Misalignment Issues in Image Editing

  modelscope/DiffSynth-Studio#1685 · Filter by author lllllialois opened on Sep 11, 2026
  1

• oom on minimax h3

  modelscope/DiffSynth-Studio#1682 · Filter by author perryyang-hue opened on Sep 10, 2026
  1

• Partial download surfaces as "Cannot detect the model type" rather than a download error

  modelscope/DiffSynth-Studio#1668 · Filter by author rakhimovv opened on Sep 2, 2026

• LingBot-Video 30B LoRA training OOM on H100

  modelscope/DiffSynth-Studio#1664 · Filter by author Zanghu-ze opened on Sep 2, 2026

• train_scale_cfg

  modelscope/DiffSynth-Studio#1663 · Filter by author smith-fugui opened on Sep 2, 2026
  3

• minimax-H3 FL2VA train error

  modelscope/DiffSynth-Studio#1660 · Filter by author stihuangyuan opened on Sep 1, 2026
  4

• Minimax h3 多卡如何推理？

  modelscope/DiffSynth-Studio#1645 · Filter by author 01-111162-173 opened on Aug 26, 2026
  1

• 训练minimax- h3的nf4模型的lora，使用adapter训练时融合lora会报错

  modelscope/DiffSynth-Studio#1642 · Filter by author professorhuanghaojing opened on Aug 26, 2026
  1

• minimax h3我看新增了 adater 训练，ref2av 目前是不支持这种训练方式吗

  modelscope/DiffSynth-Studio#1639 · Filter by author 01-111162-173 opened on Aug 26, 2026
  1

• DMD2训练损失震荡

  modelscope/DiffSynth-Studio#1617 · Filter by author 123lcy123 opened on Aug 19, 2026
  3

• [Proposal] 探讨对 Diffusion Language Model 的支持

  modelscope/DiffSynth-Studio#1609 · Filter by author tutao0123 opened on Aug 17, 2026
  1

• ltx2.5 会支持吗

  modelscope/DiffSynth-Studio#1595 · Filter by author 01-111162-173 opened on Aug 14, 2026
  2

• 会考虑支持d-opsd吗

  modelscope/DiffSynth-Studio#1588 · Filter by author NancyFyong opened on Aug 13, 2026
  5

• 训练minimax-h3的lora没办法在comfyui中使用

  modelscope/DiffSynth-Studio#1587 · Filter by author professorhuanghaojing opened on Aug 12, 2026
  3

• the training details of minimax h3 model

  modelscope/DiffSynth-Studio#1585 · Filter by author xinjiu612 opened on Aug 12, 2026
  4

• Is it possible to fine-tune Minimax H3 with silent videos?

  modelscope/DiffSynth-Studio#1576 · Filter by author plutoyuxie opened on Aug 7, 2026
  1

• Disk offload mode leaks host RAM during denoising (RSS grows every step, >15x slowdown from swap thrashing)

  modelscope/DiffSynth-Studio#1564 · Filter by author zhouxixiong4 opened on Aug 4, 2026
  1

• [Windows] Disk offload crashes with access violation after DiskMap.flush_files() (apparent use-after-free)

  modelscope/DiffSynth-Studio#1563 · Filter by author zhouxixiong4 opened on Aug 4, 2026

• MiniMax-H3这个模型经过了CFG-Distillation，它还能正常SFT吗？会不会SFT更多step以后丧失CFG蒸馏的能力？

  modelscope/DiffSynth-Studio#1557 · Filter by author 651961 opened on Aug 4, 2026
  30