diffsynth-engine 0.6.1.dev17__tar.gz → 0.6.1.dev19__tar.gz

{diffsynth_engine-0.6.1.dev17 → diffsynth_engine-0.6.1.dev19}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: diffsynth_engine
-Version: 0.6.1.dev17
+Version: 0.6.1.dev19
 Author: MuseAI x ModelScope
 Classifier: Programming Language :: Python :: 3
 Classifier: Operating System :: OS Independent

{diffsynth_engine-0.6.1.dev17 → diffsynth_engine-0.6.1.dev19}/diffsynth_engine/pipelines/flux_image.py

@@ -983,8 +983,9 @@ class FluxImagePipeline(BasePipeline):
         elif self.ip_adapter is not None:
             image_emb = self.ip_adapter.encode_image(ref_image)
         elif self.redux is not None:
-            image_prompt_embeds = self.redux(ref_image)
-            positive_prompt_emb = torch.cat([positive_prompt_emb, image_prompt_embeds], dim=1)
+            ref_prompt_embeds = self.redux(ref_image)
+            flattened_ref_emb = ref_prompt_embeds.view(1, -1, ref_prompt_embeds.size(-1))
+            positive_prompt_emb = torch.cat([positive_prompt_emb, flattened_ref_emb], dim=1)
 
         # Extra input
         image_ids, text_ids, guidance = self.prepare_extra_input(

{diffsynth_engine-0.6.1.dev17 → diffsynth_engine-0.6.1.dev19}/diffsynth_engine/pipelines/qwen_image.py

@@ -147,9 +147,10 @@ class QwenImagePipeline(BasePipeline):
         self.prompt_template_encode = "<|im_start|>system\nDescribe the image by detailing the color, shape, size, texture, quantity, text, spatial relationships of the objects and background:<|im_end|>\n<|im_start|>user\n{}<|im_end|>\n<|im_start|>assistant\n"
         self.prompt_template_encode_start_idx = 34
         # qwen image edit
-        self.edit_prompt_template_encode = "<|im_start|>system\nDescribe the key features of the input image (color, shape, size, texture, objects, background), then explain how the user's text instruction should alter or modify the image. Generate a new image that meets the user's requirements while maintaining consistency with the original input where appropriate.<|im_end|>\n<|im_start|>user\n<|vision_start|><|image_pad|><|vision_end|>{}<|im_end|>\n<|im_start|>assistant\n"
+        self.edit_system_prompt = "Describe the key features of the input image (color, shape, size, texture, objects, background), then explain how the user's text instruction should alter or modify the image. Generate a new image that meets the user's requirements while maintaining consistency with the original input where appropriate."
+        self.edit_prompt_template_encode = "<|im_start|>system\n" + self.edit_system_prompt + "<|im_end|>\n<|im_start|>user\n<|vision_start|><|image_pad|><|vision_end|>{}<|im_end|>\n<|im_start|>assistant\n"
         # qwen image edit plus
-        self.edit_plus_prompt_template_encode = "<|im_start|>system\nDescribe the key features of the input image (color, shape, size, texture, objects, background), then explain how the user's text instruction should alter or modify the image. Generate a new image that meets the user's requirements while maintaining consistency with the original input where appropriate.<|im_end|>\n<|im_start|>user\n{}<|im_end|>\n<|im_start|>assistant\n"
+        self.edit_plus_prompt_template_encode = "<|im_start|>system\n" + self.edit_system_prompt + "<|im_end|>\n<|im_start|>user\n{}<|im_end|>\n<|im_start|>assistant\n"
 
         self.edit_prompt_template_encode_start_idx = 64
 

{diffsynth_engine-0.6.1.dev17 → diffsynth_engine-0.6.1.dev19}/diffsynth_engine.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: diffsynth_engine
-Version: 0.6.1.dev17
+Version: 0.6.1.dev19
 Author: MuseAI x ModelScope
 Classifier: Programming Language :: Python :: 3
 Classifier: Operating System :: OS Independent