add qwen-image-edit-2509 and update requirements

Signed-off-by: Vladimir Mandic <mandic00@live.com>

.gitignore

@@ -46,14 +46,6 @@ tunableop_results*.csv
 !package.json
 !requirements.txt
 
-# pyinstaller
-*.spec
-build/
-dist/
-
-# dynamically generated
-/repositories/deepseek-vl2/
-
 # all dynamic stuff
 /extensions/**/*
 /outputs/**/*

CHANGELOG.md

@@ -1,10 +1,12 @@
 # Change Log for SD.Next
 
-## Update for 2025-09-22
+## Update for 2025-09-23
 
 - **Models**
   - [WAN 2.2 14B VACE](https://huggingface.co/alibaba-pai/Wan2.2-VACE-Fun-A14B)  
     available for *text-to-image* and *text-to-video* and *image-to-video* workflows  
+  - [Qwen Image Edit 2509](https://huggingface.co/Qwen/Qwen-Image-Edit-2509)  
+    improved version of Qwen Image Edit with improved image consistency  
   - [Tencent FLUX.1 Dev SRPO](https://huggingface.co/tencent/SRPO)  
     SRPO is trained by Tencent with specific technique: Directly Aligning the Full Diffusion Trajectory with Fine-Grained Human Preference  
   - [Nunchaku SDXL](https://huggingface.co/nunchaku-tech/nunchaku-sdxl) and [Nunchaku SDXL Turbo](https://huggingface.co/nunchaku-tech/nunchaku-sdxl-turbo)  

html/reference.json

@@ -170,6 +170,13 @@
     "skip": true,
     "extras": ""
   },
+  "Qwen-Image-Edit-2509": {
+    "path": "Qwen/Qwen-Image-Edit-2509",
+    "preview": "Qwen--Qwen-Image-Edit.jpg",
+    "desc": "Qwen-Image-Edit, the image editing version of Qwen-Image. Built upon our 20B Qwen-Image model, Qwen-Image-Edit successfully extends Qwen-Image’s unique text rendering capabilities to image editing tasks, enabling precise text editing.",
+    "skip": true,
+    "extras": ""
+  },
   "Qwen-Image-Lightning": {
     "path": "vladmandic/Qwen-Lightning",
     "preview": "vladmandic--Qwen-Lightning.jpg",

installer.py

@@ -608,7 +608,7 @@ def check_diffusers():
     if args.skip_git:
         install('diffusers')
         return
-    sha = 'efb7a299af46d739dec6a57a5d2814165fba24b5' # diffusers commit hash
+    sha = '1448b035859dd57bbb565239dcdd79a025a85422' # diffusers commit hash
     pkg = pkg_resources.working_set.by_key.get('diffusers', None)
     minor = int(pkg.version.split('.')[1] if pkg is not None else -1)
     cur = opts.get('diffusers_version', '') if minor > -1 else ''
@@ -635,8 +635,8 @@ def check_transformers():
         target_transformers = '4.52.4'
         target_tokenizers = '0.21.4'
     else:
-        target_transformers = '4.56.1'
-        target_tokenizers = '0.22.0'
+        target_transformers = '4.56.2'
+        target_tokenizers = '0.22.1'
     if (pkg_transformers is None) or ((pkg_transformers.version != target_transformers) or (pkg_tokenizers is None) or ((pkg_tokenizers.version != target_tokenizers) and (not args.experimental))):
         if pkg_transformers is None:
             log.info(f'Transformers install: version={target_transformers}')
@@ -1273,8 +1273,8 @@ def install_insightface():
         uninstall('albumentations')
         install('albumentationsx')
     else:
-       uninstall('albumentationsx')
-       install('albumentations==1.4.3', ignore=True, quiet=True)
+        uninstall('albumentationsx')
+        install('albumentations==1.4.3', ignore=True, quiet=True)
     install_pydantic()
 
 

modules/cachedit.py

@@ -5,7 +5,7 @@ from modules import shared
 def apply_cache_dit(pipe):
     if not shared.opts.cache_dit_enabled:
         return
-    install('cache_dit')
+    install('git+https://github.com/vipshop/cache-dit', 'cache_dit')
     try:
         import cache_dit
     except Exception as e:
@@ -20,7 +20,9 @@ def apply_cache_dit(pipe):
     if getattr(pipe, 'has_cache_dit', False):
         unapply_cache_dir(pipe)
 
-    cache_config = cache_dit.BasicCacheConfig()
+    config_args = {}
+
+    cache_config = cache_dit.BasicCacheConfig(**config_args)
     if shared.opts.cache_dit_calibrator == "TaylorSeer":
         calibrator_config = cache_dit.TaylorSeerCalibratorConfig(taylorseer_order=1)
     elif shared.opts.cache_dit_calibrator == "FoCa":
@@ -52,6 +54,8 @@ def unapply_cache_dir(pipe):
         return
     try:
         import cache_dit
+        stats = cache_dit.summary(pipe)
+        shared.log.critical(f'Unapply Cache-DiT: {stats}')
         cache_dit.disable_cache(pipe)
         pipe.has_cache_dit = False
     except Exception:

modules/processing_args.py

@@ -103,9 +103,9 @@ def task_specific_kwargs(p, model):
         }
 
     # model specific args
-    if model_cls == 'QwenImageEditPipeline' and len(getattr(p, 'init_images', [])) == 0:
+    if 'QwenImageEdit' in model_cls and len(getattr(p, 'init_images', [])) == 0:
         task_args['image'] = [Image.new('RGB', (p.width, p.height), (0, 0, 0))] # monkey-patch so qwen-image-edit pipeline does not error-out on t2i
-    if model_cls == 'LatentConsistencyModelPipeline' and hasattr(p, 'init_images') and len(p.init_images) > 0:
+    if 'LatentConsistencyModelPipeline' in model_cls and hasattr(p, 'init_images') and len(p.init_images) > 0:
         p.ops.append('lcm')
         init_latents = [processing_vae.vae_encode(image, model=shared.sd_model, vae_type=p.vae_type).squeeze(dim=0) for image in p.init_images]
         init_latent = torch.stack(init_latents, dim=0).to(shared.device)
@@ -116,7 +116,7 @@ def task_specific_kwargs(p, model):
             'width': p.width if hasattr(p, 'width') else None,
             'height': p.height if hasattr(p, 'height') else None,
         }
-    if model_cls == 'BlipDiffusionPipeline':
+    if 'BlipDiffusionPipeline' in model_cls:
         if len(getattr(p, 'init_images', [])) == 0:
             shared.log.error('BLiP diffusion requires init image')
             return task_args
@@ -126,9 +126,9 @@ def task_specific_kwargs(p, model):
             'target_subject_category': getattr(p, 'prompt', '').split()[-1],
             'output_type': 'pil',
         }
-    if (model.__class__.__name__ == 'WanImageToVideoPipeline') and (getattr(p, 'init_images', None) is not None) and (len(p.init_images) > 0):
+    if ('WanImageToVideoPipeline' in model_cls) and (getattr(p, 'init_images', None) is not None) and (len(p.init_images) > 0):
         task_args['image'] = p.init_images[0]
-    if (model.__class__.__name__ == 'WanVACEPipeline') and (getattr(p, 'init_images', None) is not None) and (len(p.init_images) > 0):
+    if ('WanVACEPipeline' in model_cls) and (getattr(p, 'init_images', None) is not None) and (len(p.init_images) > 0):
         task_args['reference_images'] = p.init_images
 
     if debug_enabled:

pipelines/model_qwen.py

@@ -12,7 +12,12 @@ def load_qwen(checkpoint_info, diffusers_load_config={}):
     load_args, _quant_args = model_quant.get_dit_args(diffusers_load_config, module='Model')
     shared.log.debug(f'Load model: type=Qwen model="{checkpoint_info.name}" repo="{repo_id}" offload={shared.opts.diffusers_offload_mode} dtype={devices.dtype} args={load_args}')
 
-    if 'Edit' in repo_id:
+    if '2509' in repo_id :
+        cls_name = diffusers.QwenImageEditPlusPipeline
+        diffusers.pipelines.auto_pipeline.AUTO_TEXT2IMAGE_PIPELINES_MAPPING["qwen-image"] = diffusers.QwenImageEditPlusPipeline
+        diffusers.pipelines.auto_pipeline.AUTO_IMAGE2IMAGE_PIPELINES_MAPPING["qwen-image"] = diffusers.QwenImageEditPlusPipeline
+        diffusers.pipelines.auto_pipeline.AUTO_INPAINT_PIPELINES_MAPPING["qwen-image"] = diffusers.QwenImageEditPlusPipeline
+    elif 'Edit' in repo_id:
         cls_name = diffusers.QwenImageEditPipeline
         diffusers.pipelines.auto_pipeline.AUTO_TEXT2IMAGE_PIPELINES_MAPPING["qwen-image"] = diffusers.QwenImageEditPipeline
         diffusers.pipelines.auto_pipeline.AUTO_IMAGE2IMAGE_PIPELINES_MAPPING["qwen-image"] = diffusers.QwenImageEditPipeline
@@ -36,7 +41,13 @@ def load_qwen(checkpoint_info, diffusers_load_config={}):
         transformer_subfolder = "transformer"
 
     if transformer is None:
-        transformer = generic.load_transformer(repo_transformer, subfolder=transformer_subfolder, cls_name=diffusers.QwenImageTransformer2DModel, load_config=diffusers_load_config, modules_dtype_dict={"minimum_6bit": ["pos_embed", "time_text_embed", "img_in", "txt_in", "norm_out", "transformer_blocks.0.img_mod.1.weight"]})
+        transformer = generic.load_transformer(
+            repo_transformer,
+            subfolder=transformer_subfolder,
+            cls_name=diffusers.QwenImageTransformer2DModel,
+            load_config=diffusers_load_config,
+            modules_dtype_dict={"minimum_6bit": ["pos_embed", "time_text_embed", "img_in", "txt_in", "norm_out", "transformer_blocks.0.img_mod.1.weight"]},
+        )
 
     repo_te = 'Qwen/Qwen-Image'
     text_encoder = generic.load_text_encoder(repo_te, cls_name=transformers.Qwen2_5_VLForConditionalGeneration, load_config=diffusers_load_config)

requirements.txt

@@ -44,7 +44,7 @@ tqdm==4.67.1
 accelerate==1.10.1
 opencv-contrib-python-headless==4.11.0.86
 einops==0.8.1
-huggingface_hub==0.34.4
+huggingface_hub==0.35.0
 numexpr==2.11.0
 numpy==2.1.2
 pandas==2.3.1