Issues

• Support for Qwen 2.1

  tdrussell/diffusion-pipe#547 · Filter by author Dream-Making-Git opened on Sep 21, 2026
  1

• Consider using pre-built Flash Attention kernels via kernels

  tdrussell/diffusion-pipe#540 · Filter by author sayakpaul opened on Aug 11, 2026
  1

• Using eval_datasets doubles memory usage leading to OOMs

  tdrussell/diffusion-pipe#538 · Filter by author thrnz opened on Aug 9, 2026
  1

• (Minimax h3 training) Comfy video get_frame_count() does not match actual frame count, causing errors

  tdrussell/diffusion-pipe#537 · Filter by author gitmylo opened on Aug 8, 2026
  7

• ostris/minimax_h3_training_adapter

  tdrussell/diffusion-pipe#534 · Filter by author Jamzsam opened on Aug 6, 2026
  15

• Support for MiniMax H3 Model

  tdrussell/diffusion-pipe#533 · Filter by author TianDongL opened on Aug 3, 2026
  41

• LTX-2 nan loss

  tdrussell/diffusion-pipe#531 · Filter by author paju1986 opened on Jul 30, 2026

• resume from checkpoint fails [Krea 2]

  tdrussell/diffusion-pipe#528 · Filter by author paju1986 opened on Jun 29, 2026
  3

• Qwen image training error, AttributeError: 'AutoencoderKLQwenImage' object has no attribute 'add_embedding'

  tdrussell/diffusion-pipe#525 · Filter by author victrue opened on May 26, 2026

• Add merge_adapter-support to Qwen-image module (fork exists).

  tdrussell/diffusion-pipe#524 · Filter by author fbbcool opened on May 13, 2026

• regularization dataset - how to?

  tdrussell/diffusion-pipe#523 · Filter by author tedbiv opened on May 5, 2026

• Feature request: Multi-image edit / multi-control training support for FLUX.2 Klein and Qwen-Image-Edit

  tdrussell/diffusion-pipe#522 · Filter by author Xabsurd opened on Apr 30, 2026
  1

• Support Ernie model

  tdrussell/diffusion-pipe#519 · Filter by author LorisTecnology opened on Apr 15, 2026

• Please add the multi-image control feature(Qwen,Flux2...). Let's keep pushing forward.

  tdrussell/diffusion-pipe#516 · Filter by author samliu315 opened on Apr 6, 2026
  1

• Save lora in diffusers format?

  tdrussell/diffusion-pipe#514 · Filter by author Sector14 opened on Mar 12, 2026

• Video "is being skipped because it is too short" and a quick hack to fix it.

  tdrussell/diffusion-pipe#513 · Filter by author mkultra333 opened on Feb 24, 2026

• Wan 2.1 T2V 14B training issue

  tdrussell/diffusion-pipe#510 · Filter by author piszczu4 opened on Feb 13, 2026
  1

• cache_text_embedding option for Flux2 and Anima

  tdrussell/diffusion-pipe#509 · Filter by author thojmr opened on Feb 8, 2026

• Add support for ACE Step 1.5

  tdrussell/diffusion-pipe#508 · Filter by author LorisTecnology opened on Feb 8, 2026

• Multi CPU with Runpod

  tdrussell/diffusion-pipe#506 · Filter by author BenDes21 opened on Feb 5, 2026
  1

• zimage base support

  tdrussell/diffusion-pipe#503 · Filter by author TianDongL opened on Jan 27, 2026
  22

• RuntimeError: CUDA error: CUBLAS_STATUS_INVALID_VALUE when calling `cublasGemmEx(

  tdrussell/diffusion-pipe#500 · Filter by author LorisTecnology opened on Jan 22, 2026
  1

• [rank0]: ValueError: Hidden size 4096 must be divisible by num_heads 48 （training：flux-2-klein-base-9b）

  tdrussell/diffusion-pipe#499 · Filter by author BillusA1111 opened on Jan 22, 2026
  3

• Is there any plan to release the training code for flux2-klein’s multi-image editing?

  tdrussell/diffusion-pipe#498 · Filter by author Iriya99 opened on Jan 19, 2026
  4

• flux-2-klein-9B lora training error

  tdrussell/diffusion-pipe#497 · Filter by author wuyuetiger opened on Jan 19, 2026
  10