Repository navigation

feat: Qwen Image Edit 2511 — full pipeline with LoRA, GGUF, quantization, and UI - #131

Closed

lstein wants to merge 60 commits into

mainfrom

feat/qwen-image-edit-2511

Closed

feat: Qwen Image Edit 2511 — full pipeline with LoRA, GGUF, quantization, and UI#131
lstein wants to merge 60 commits into
mainfrom
feat/qwen-image-edit-2511

Commits on Mar 27, 2026
- Add Qwen Image Edit 2511 model support
  
  lstein
  and
  claude
  committed
  bc82599
- fix: diffusers compat, VL encoder image resize, badge wrapping
  
  lstein
  and
  claude
  committed
  a373e7e
- fix: free BnB-quantized encoder after use, suppress int8 cast warning
  
  lstein
  and
  claude
  committed
  1f248b7
- chore: ruff & uv lock
  lstein
  committed
  5cafafe
- Merge branch 'main' into feat/qwen-image-edit-2511
  lstein
  authored
  073b08b
- fix: bypass HfApi in model install test
  lstein
  committed
  c9f781e
- feat(backend): add support for a Comfy LoRA variant
  lstein
  committed
  ceaf97d
Commits on Mar 28, 2026
- refactor: rename QwenImageEdit -> QwenImage base type
  
  lstein
  and
  claude
  committed
  01d6f09
- feat: add QwenImage variant type to distinguish edit vs txt2img models
  
  lstein
  and
  claude
  committed
  c23f9ca
- fix: make QwenImage variant optional to fix model detection tags
  
  lstein
  and
  claude
  committed
  e13faf5
- fix: rename edit starter model variables to qwen_image_edit_* prefix
  
  lstein
  and
  claude
  committed
  87bda37
- fix: remove unused exports (zQwenImageVariantType, isQwenImageEditMainModelConfig)
  
  lstein
  and
  claude
  committed
  62ccb75
- fix: restore correct GGUF filenames in Qwen Image Edit starter model URLs
  
  lstein
  and
  claude
  committed
  cbfd804
- feat: tag edit starter models with variant, pass variant through install
  
  lstein
  and
  claude
  committed
  939f6f9
- fix: only set zero_cond_t=True for edit-variant GGUF models
  
  lstein
  and
  claude
  committed
  4982343
- fix: prevent Flux LoRAs from being detected as Qwen Image LoRAs
  
  lstein
  and
  claude
  committed
  3a39a79
- test: add Qwen Image LoRA model identification tests
  
  lstein
  and
  claude
  committed
  60cf18b
- chore: ruff
  lstein
  committed
  153f260
- chore: ruff
  lstein
  committed
  2d89174
- fix: don't force reference image to output aspect ratio in VAE encoding
  
  lstein
  and
  claude
  committed
  2eb2c55
Commits on Mar 29, 2026
- fix: clip denoise schedule by denoising_start/end, block GGUF enqueue without component source
  
  lstein
  and
  claude
  committed
  af1dc29
- test: add Qwen Image denoise, model loader, and text encoder tests
  
  lstein
  and
  claude
  committed
  e3449c1
Commits on Mar 30, 2026
- refactor: rename Qwen Image Edit node titles/descriptions to Qwen Image
  
  lstein
  and
  claude
  committed
  636b696
- chore: ruff
  lstein
  committed
  61d3a79
- chore: ruff
  lstein
  committed
  3d92014
- fix: skip negative text encoder node when CFG <= 1
  
  lstein
  and
  claude
  committed
  5bc4154
- Merge branch 'main' into feat/qwen-image-edit-2511
  JPPhoto
  authored
  ef4e41c
- feat: support Kohya-format Qwen Image LoRAs (LoKR)
  
  lstein
  and
  claude
  committed
  09119dc
Commits on Apr 1, 2026
- chore: ruff
  lstein
  committed
  f26a35e
- fix: generate noise in float32 instead of float16
  
  lstein
  and
  claude
  committed
  e2e16f7
Commits on Apr 4, 2026
- Agent-Logs-Url: https://github.com/lstein/InvokeAI/sessions/24879c8d-70b1-43ca-a314-4367e472cd4f
  
  Copilot
  and
  lstein
  authored
  9e828b9
- test: add Qwen Image LoRA identification and conversion tests
  
  lstein
  and
  claude
  committed
  b275ae0
- chore(backend): ruff
  lstein
  committed
  1ec8bad
- chore(frontend): update schema
  lstein
  committed
  08d40f9
Commits on Apr 5, 2026
Commits on Apr 6, 2026
- Merge branch 'main' into feat/qwen-image-edit-2511
  JPPhoto
  authored
  bbc1d7f
- fix(frontend): align CFG switch in front & backends; fix Qwen Image metadata recall
  lstein
  committed
  0c0af80
- chore: merge with main and fix typegen
  lstein
  committed
  7ae0eb9
- fix(backend): use correct latent factors for Qwen Image preview images
  lstein
  committed
  d92dffb
- feat: use chained collectors for deterministic reference image ordering
  
  lstein
  and
  claude
  committed
  cf05fb1
- test: add backend unit tests for Qwen Image preview image generation
  
  lstein
  and
  claude
  committed
  e5d3d65
Commits on Apr 7, 2026
Commits on Apr 10, 2026
- Merge branch 'main' into feat/qwen-image-edit-2511
  lstein
  committed
  34e64f0
- chore: regenerate schema.ts with merged Qwen Image + Anima types
  
  lstein
  and
  claude
  committed
  ce41b08
- chore(frontend): lint
  lstein
  committed
  9bf5b16
- Merge branch 'main' into feat/qwen-image-edit-2511
  lstein
  committed
  d1143e9
- feat: move Qwen Image shift to Generation accordion with slider UI
  
  lstein
  and
  claude
  committed
  7bf1369
- feat: auto-default Qwen Image component source to first available diffusers model
  
  lstein
  and
  claude
  committed
  0620c23
- fix(frontend): autoselect qwen3 image variant diffusers source when switching qwen3 denoisers
  lstein
  committed
  4d4536c
Commits on Apr 12, 2026
- fix(backend): use tensor structure to discriminate edit from generate denoise model variants
  lstein
  committed
  043bc0a
- test(backend): update Qwen GGUF variant assertion for non-edit models
  
  lstein
  and
  claude
  committed
  6e67d9a

Commits

Commits on Mar 27, 2026

Commits on Mar 28, 2026

Commits on Mar 29, 2026

Commits on Mar 30, 2026

Commits on Apr 1, 2026

Commits on Apr 4, 2026

Commits on Apr 5, 2026

Commits on Apr 6, 2026

Commits on Apr 7, 2026

Commits on Apr 10, 2026

Commits on Apr 12, 2026