Version 0.2.13

Highlights

• Support for Qwen-Image-Edit-2509: This release introduces support for the Qwen-Image-Edit-2509 model, enabling training and inference with multiple control images for more complex image editing tasks. (in #590)
• Reduced VRAM Usage for Block Swap: The shared VRAM usage for the block swap feature has been significantly reduced on Windows. (in #585)
• Dataset Handling Fix: A bug has been fixed where the first data item of each epoch was being handled incorrectly. This ensures that all data is processed properly throughout the epoch. (in #601)

What's Changed

• feat: add option to force 2.1 style time embedding in WanModel by @kohya-ss in #586
• doc: Wan update offloading instructions for DiT model on Windows by @kohya-ss in #597
• feat: Add flag to disable cuDNN PyTorch backend when caching by @xzuyn in #592
• Free VRAM for lazy loading in batch prompt mode by @JCBrouwer in #593
• doc: update dataset configuration for control images with mask by @kohya-ss in #600
• feat: Add support for Qwen-Image-Edit-2509 by @kohya-ss in #590
• fix: Qwen-Image-Edit (not 2509) incorrect prompt for VLM. by @kohya-ss in #606
• fix: Qwen-Image-Edit-2509 cannot handle arbitrary number of control images by @kohya-ss in #613
• Fix Prodigy optimizer logs by @wenyifancc in #623
• feat: Reducing shared VRAM usage for block swap by @kohya-ss in #585
• Organizing document structure by @kohya-ss in #630
• fix: first data of the epoch from dataset is inappropriate by @kohya-ss in #601
• chore: bump version to 0.2.13 by @kohya-ss in #631

New Contributors

• @JCBrouwer made their first contribution in #593
• @wenyifancc made their first contribution in #623

Full Changelog: v0.2.12...v0.2.13

Version 0.2.12

Highlights

📝 Code Quality: Ruff Formatting: PR #538

We introduced Ruff for unified linting and formatting across the codebase. This standardization improves contributor experience, ensures consistent style, and streamlines the review process. Thank you @arledesma for this great contribution!

⚡ CPU Offloading for Gradient Checkpointing: PR #537

Added support for activation CPU offloading during gradient checkpointing. This feature reduces VRAM usage by up to 20–30% in large-scale video training, enabling larger batch sizes. The trade-off is slightly slower training (a few percent to ~20%).

🚀 Faster Model Loading with MemoryEfficientSafeOpen: PR #556

Improved .safetensors loading with np.memmap and non-blocking GPU transfer, making model load times up to 1.5× faster. This significantly reduces waiting time for large model initialization.

🔬 FP8 Quantization with Block-wise Scaling: PR #575

Changed the --fp8_scaled option from per-tensor quantization to block-wise scaling, resulting in improved accuracy and stability. For Qwen-Image LoRA training, this reduces VRAM usage by about 5GB. Training and inference speed may be slightly slower.

What's Changed

• Configure ruff format for code quality standardization by @arledesma in #538
• doc: update contributing guidelines for clarity and consistency by @kohya-ss in #541
• feat: add CPU offloading support for gradient checkpointing by @kohya-ss in #537
• docs: update README to include ruff code analysis and activation CPU offloading by @kohya-ss in #545
• bugfix: fix mask initialization for wan flf2v inference by @LittleNyima in #548
• feat: faster MemoryEfficientSafeOpen by @kohya-ss in #556
• from_file with qwen_image_generate_image is broken by @nmfisher in #553
• fix: Qwen-Image training not working with fp8_base by @kohya-ss in #559
• chore: add .vscode/settings.json to .gitignore by @kohya-ss in #562
• feat: faster LoRA merging and stable fp8 max calculation by @kohya-ss in #563
• fix(wan): FlashAttention-3 call error by @jimlee2048 in #570
• Fix qwen image from file generation by @kohya-ss in #557
• fix: add dtype handling for fp8 model weights in weight_hook_func by @kohya-ss in #582
• fix: VAE tiling is always enabled for FramePack by @kohya-ss in #583
• feat: FP8 quantization with block-wise scaling by @kohya-ss in #575
• chore: bump version to 0.2.12 by @kohya-ss in #584

New Contributors

• @LittleNyima made their first contribution in #548
• @nmfisher made their first contribution in #553
• @jimlee2048 made their first contribution in #570

Full Changelog: v0.2.11...v0.2.12

Version 0.2.11

What's Changed

• fix: correct metadata key for session ID in NetworkTrainer by @kohya-ss in #516
• Code Quality - F821 by @arledesma in #483
• Code Quality - Configure ruff linting with Exclusions by @arledesma in #488
• doc: README Add ruff code analysis introduction by @kohya-ss in #522
• feat: add QwenImageTrainer for fine-tuning with Adafactor optimizer by @kohya-ss in #492
• doc: split HunyuanVideo documentation from README by @kohya-ss in #525
• feat: Add REX learning rate scheduler by @xzuyn in #513
• Feat small refactoring for rex scheduler by @kohya-ss in #535
• Fix bugs when qwen-image use lycoris by @sdbds in #530
• bump version to v0.2.11 by @kohya-ss in #540

New Contributors

• @arledesma made their first contribution in #483
• @xzuyn made their first contribution in #513

Full Changelog: v0.2.10...v0.2.11

Version 0.2.10

What's Changed

• Minor usability improvements by @kohya-ss in #486
• Add sponsor logo by @kohya-ss in #487
• feat: Wan use original dtype in AttentionBlock to reduce memory usage by @kohya-ss in #493
• Feat add codex prompt by @kohya-ss in #507
• feat: support another LoRA format in convert_lora.py by @kohya-ss in #508
• feat: remove bitsandbytes version and update sentencepiece by @kohya-ss in #509
• feat: Fix support for schedulefree optimizer by @am7coffee in #505
• fix: streamline optimizer training mode by @kohya-ss in #510
• doc: add Schedule Free Optimizer support and update documentation by @kohya-ss in #511

New Contributors

• @am7coffee made their first contribution in #505

Full Changelog: v0.2.9...v0.2.10

v0.2.9

What's Changed

• doc: update documentation to clarify LoRA weight compatibility with ComfyUI for Wan by @kohya-ss in #465
• fix: update documentation to include network_module for Qwen-Image training by @kohya-ss in #467
• fix: load appropriate lora (high/low) for lazy loading by @kohya-ss in #479
• Update pyproject.toml by @SamTyurenkov in #472
• feat: Qwen-Image-Edit inference and training by @kohya-ss in #473

New Contributors

• @SamTyurenkov made their first contribution in #472

Full Changelog: v0.2.8...v0.2.9

v0.2.8

What's Changed

• fix: handle wandb initialization when wandb is not installed by @kohya-ss in #425
• Feat wan generate lazy loading by @kohya-ss in #427
• Add qwen_shift and qinglong_qwen by @sdbds in #428
• doc: update documents for qwen-image sampling method by @kohya-ss in #429
• fix: convert embed dtype to bfloat16 for QwenVL-2.5 fp8 caching by @kohya-ss in #440
• fix: fp8_base not working for Qwen-Image by @kohya-ss in #441
• Supports conversion of Qwen-Image LoRA to diffuser format by @kohya-ss in #444
• Fix error occurs in sampling image during the Qwen-Image training by @zhao-kun in #451
• fix: improve error handling for missing LoRA modules by @kohya-ss in #452
• feat: add support for timestep bucketing in training by @kohya-ss in #418
• fix: hv_generate_video requires a lot of VRAM by @kohya-ss in #459
• feat: Captioning with Qwen2.5-VL by @kohya-ss in #460

New Contributors

• @zhao-kun made their first contribution in #451

Full Changelog: v0.2.7...v0.2.8

v0.2.7

What's Changed

• doc: update wan.md to include for --discrete_flow_shift in Wan2.2 models by @kohya-ss in #422
• feature: Qwen-Image support by @kohya-ss in #408

Full Changelog: v0.2.6...v0.2.7

v0.2.6

What's Changed

• feat: Logging sample images and videos with wandb by @xhiroga in #420
• doc: update README to include wandb logging feature and recent changes by @kohya-ss in #421

Full Changelog: v0.2.5...v0.2.6

v0.2.5

What's Changed

• Dynamic fp8 scaling and lora merging by @kohya-ss in #406
• fix: update attention mask handling for --split_attn by @kohya-ss in #413
• Fixed Wan2.1 model dtype casting with non-fp8 dtypes by @kohya-ss in #414
• Add new SNR function for better training by @sdbds in #407
• Add documentation for new sampling methods logsnr and qinglong by @kohya-ss in #415
• feature: Wan2.2 14B support by @kohya-ss in #399

Full Changelog: v0.2.4...v0.2.5

v0.2.4

What's Changed

• Add prompt files for AI coding agents by @kohya-ss in #398
• Fix bugs about flux kontext part by @sdbds in #402
• fix: fix function name for offloading in FLUX1 Kontext by @kohya-ss in #403
• fix: add fp_latent_window_size to dataset schemas by @kohya-ss in #404

Full Changelog: v0.2.3...v0.2.4