gr.Galleryからgr.Imageに変更します。

app.py

@@ -81,7 +81,7 @@ MAX_SEED = np.iinfo(np.int32).max
 # --- Main Inference Function (Combined LoRA) ---
 @spaces.GPU()
 def infer(
-    images,
+    image,
     seed=42,
     randomize_seed=False,
     true_guidance_scale=1.0,
@@ -96,7 +96,7 @@ def infer(
     Run single inference with combined LoRAs: Lightning + Stage1 + Stage2.
 
     Parameters:
-        images (list): Input images from the Gradio gallery (PIL or path-based).
+        image: Input image (PIL Image or path string).
         seed (int): Random seed for reproducibility.
         randomize_seed (bool): If True, overrides seed with a random value.
         true_guidance_scale (float): CFG scale used by Qwen-Image.
@@ -108,7 +108,7 @@ def infer(
         progress: Gradio progress callback.
 
     Returns:
-        tuple: (result_images, seed_used)
+        tuple: (result_image, seed_used)
     """
 
     # Hardcode the negative prompt
@@ -120,19 +120,13 @@ def infer(
     # Set up the generator for reproducibility
     generator = torch.Generator(device=device).manual_seed(seed)
 
-    # Load input images into PIL Images
-    pil_images = []
-    if images is not None:
-        for item in images:
-            try:
-                if isinstance(item[0], Image.Image):
-                    pil_images.append(item[0].convert("RGB"))
-                elif isinstance(item[0], str):
-                    pil_images.append(Image.open(item[0]).convert("RGB"))
-                elif hasattr(item, "name"):
-                    pil_images.append(Image.open(item.name).convert("RGB"))
-            except Exception:
-                continue
+    # Load input image into PIL Image
+    pil_image = None
+    if image is not None:
+        if isinstance(image, Image.Image):
+            pil_image = image.convert("RGB")
+        elif isinstance(image, str):
+            pil_image = Image.open(image).convert("RGB")
 
     if height==256 and width==256:
         height, width = None, None
@@ -147,7 +141,7 @@ def infer(
     pipe.set_adapters(["lightning", "stage1", "stage2"], adapter_weights=[1.0, stage1_weight, stage2_weight])
 
     result_images = pipe(
-        image=pil_images if len(pil_images) > 0 else None,
+        image=[pil_image] if pil_image is not None else None,
         prompt=STAGE1_PROMPT,
         height=height,
         width=width,
@@ -158,8 +152,8 @@ def infer(
         num_images_per_prompt=1,
     ).images
 
-    # Return result images and seed
-    return result_images, seed
+    # Return first result image and seed
+    return result_images[0] if result_images else None, seed
 
 # --- Examples and UI Layout ---
 examples = []
@@ -185,15 +179,14 @@ with gr.Blocks(css=css) as demo:
         with gr.Row():
             with gr.Column(scale=1):
                 gr.Markdown("### 📥 Input")
-                input_images = gr.Gallery(label="Input Images",
-                                          show_label=False,
-                                          type="pil",
-                                          interactive=True,
-                                          object_fit="contain")
+                input_image = gr.Image(label="Input Image",
+                                       show_label=False,
+                                       type="pil",
+                                       interactive=True)
 
             with gr.Column(scale=1):
                 gr.Markdown("### 📤 Result")
-                result = gr.Gallery(label="Result", show_label=False, type="pil", interactive=False, object_fit="contain")
+                result = gr.Image(label="Result", show_label=False, type="pil", interactive=False)
 
         run_button = gr.Button("🚀 Generate", variant="primary", size="lg")
 
@@ -263,7 +256,7 @@ with gr.Blocks(css=css) as demo:
     run_button.click(
         fn=infer,
         inputs=[
-            input_images,
+            input_image,
             seed,
             randomize_seed,
             true_guidance_scale,