Update app.py

app.py

@@ -17,7 +17,6 @@ from transformers import (
     Sam3VideoModel, Sam3VideoProcessor
 )
 
-# --- THEME CONFIGURATION ---
 colors.steel_blue = colors.Color(
     name="steel_blue",
     c50="#EBF3F8",
@@ -80,7 +79,6 @@ class CustomBlueTheme(Soft):
 
 app_theme = CustomBlueTheme()
 
-# --- MODEL MANAGEMENT & UTILS ---
 MODEL_CACHE = {} 
 device = "cuda" if torch.cuda.is_available() else "cpu"
 print(f"Using compute device: {device}")
@@ -103,7 +101,6 @@ def load_segmentation_model(model_key):
     
     try:
         if model_key == "img_seg_model":
-            # Using generic internal names
             seg_model = Sam3Model.from_pretrained("facebook/sam3").to(device)
             seg_processor = Sam3Processor.from_pretrained("facebook/sam3")
             MODEL_CACHE[model_key] = (seg_model, seg_processor)
@@ -185,7 +182,8 @@ def run_image_segmentation(source_img, text_query, conf_thresh=0.5):
             mask_threshold=0.5,
             target_sizes=model_inputs.get("original_sizes").tolist()
         )[0]
-        
+
+        # Use AnnotatedImage format
         annotation_list = []
         raw_masks = processed_results['masks'].cpu().numpy()
         raw_scores = processed_results['scores'].cpu().numpy()
@@ -250,7 +248,6 @@ def run_video_segmentation(source_vid, text_query, frame_limit, time_limit):
     except Exception as e:
         return None, f"Error during video processing: {str(e)}"
 
-# --- GUI ---
 custom_css="""
 #col-container { margin: 0 auto; max-width: 1100px; }
 #main-title h1 { font-size: 2.1em !important; }
@@ -274,6 +271,17 @@ with gr.Blocks(css=custom_css, theme=app_theme) as main_interface:
                     with gr.Column(scale=1.5):
                         image_result = gr.AnnotatedImage(label="Segmented Result", height=450)
                 
+                gr.Examples(
+                    examples=[
+                        ["examples/player.jpg", "player in white", 0.5],
+                    ],
+                    inputs=[image_input, txt_prompt_img, conf_slider],
+                    outputs=[image_result],
+                    fn=run_image_segmentation,
+                    cache_examples=False,
+                    label="Image Examples"
+                )
+
                 btn_process_img.click(
                     fn=run_image_segmentation, 
                     inputs=[image_input, txt_prompt_img, conf_slider], 
@@ -296,6 +304,17 @@ with gr.Blocks(css=custom_css, theme=app_theme) as main_interface:
                         video_result = gr.Video(label="Processed Video")
                         process_status = gr.Textbox(label="System Status", interactive=False)
                 
+                gr.Examples(
+                    examples=[
+                        ["examples/sample_video.mp4", "ball", 60, 60],
+                    ],
+                    inputs=[video_input, txt_prompt_vid, frame_limiter, time_limiter],
+                    outputs=[video_result, process_status],
+                    fn=run_video_segmentation,
+                    cache_examples=False,
+                    label="Video Examples"
+                )
+
                 btn_process_vid.click(
                     run_video_segmentation, 
                     inputs=[video_input, txt_prompt_vid, frame_limiter, time_limiter],