Added samples

.gitattributes

@@ -14,6 +14,11 @@
 *.npy filter=lfs diff=lfs merge=lfs -text
 *.npz filter=lfs diff=lfs merge=lfs -text
 *.onnx filter=lfs diff=lfs merge=lfs -text
+*.mp4 filter=lfs diff=lfs merge=lfs -text
+*.jpg filter=lfs diff=lfs merge=lfs -text
+*.gif filter=lfs diff=lfs merge=lfs -text
+*.png filter=lfs diff=lfs merge=lfs -text
+*.webp filter=lfs diff=lfs merge=lfs -text
 *.ot filter=lfs diff=lfs merge=lfs -text
 *.parquet filter=lfs diff=lfs merge=lfs -text
 *.pb filter=lfs diff=lfs merge=lfs -text

app.py

@@ -15,6 +15,9 @@ MODEL_PATH = hf_hub_download(
 backend_id = cv.dnn.DNN_BACKEND_OPENCV
 target_id  = cv.dnn.DNN_TARGET_CPU
 
+car_on_road_video = "examples/car.mp4"
+car_in_desert_video = "examples/desert_car.mp4"
+
 # Global state
 state = {
     "points": [],
@@ -23,6 +26,12 @@ state = {
     "first_frame": None
 }
 
+#Example bounding boxes
+bbox_dict = {
+    "car.mp4": "(152, 356, 332, 104)",
+    "desert_car.mp4": "(758, 452, 119, 65)",
+}
+
 def load_first_frame(video_path):
     """Load video, grab first frame, reset state."""
     state["video_path"] = video_path
@@ -32,11 +41,8 @@ def load_first_frame(video_path):
     if not has_frame:
         return None
     state["first_frame"] = frame.copy()
-    state["points"].clear()
-    state["bbox"] = None
     return cv.cvtColor(frame, cv.COLOR_BGR2RGB)
 
-
 def select_point(img, evt: gr.SelectData):
     """Accumulate up to 4 clicks, draw polygon + bounding box."""
     if state["first_frame"] is None:
@@ -64,7 +70,6 @@ def select_point(img, evt: gr.SelectData):
 
     return cv.cvtColor(vis, cv.COLOR_BGR2RGB)
 
-
 def clear_points():
     """Reset selected points only."""
     state["points"].clear()
@@ -73,7 +78,6 @@ def clear_points():
         return None
     return cv.cvtColor(state["first_frame"], cv.COLOR_BGR2RGB)
 
-
 def clear_all():
     """Reset everything."""
     state["points"].clear()
@@ -82,7 +86,6 @@ def clear_all():
     state["first_frame"] = None
     return None, None, None
 
-
 def track_video():
     """Init VitTrack and process entire video, return output path."""
     if state["video_path"] is None or state["bbox"] is None:
@@ -145,8 +148,21 @@ def track_video():
     writer.release()
     return out_path
 
+def example_pipeline(video_path):
+    clear_all()
+
+    filename = video_path.split('/')[-1]
+    state["video_path"] = video_path
+    state["bbox"] = eval(bbox_dict[filename])
+
+    return track_video()
+
+with gr.Blocks(css='''.example * {
+    font-style: italic;
+    font-size: 18px !important;
+    color: #0ea5e9 !important;
+    }''') as demo:
 
-with gr.Blocks() as demo:
     gr.Markdown("## VitTrack: Interactive Video Object Tracking")
     gr.Markdown(
         """
@@ -166,7 +182,7 @@ with gr.Blocks() as demo:
     )
 
     with gr.Row():
-        video_in     = gr.File(label="Upload Video", file_types=[".mp4", ".avi"])
+        video_in     = gr.Video(label="Upload Video")
         first_frame  = gr.Image(label="First Frame", interactive=True)
         output_video = gr.Video(label="Tracking Result")
 
@@ -175,6 +191,24 @@ with gr.Blocks() as demo:
         clear_pts_btn = gr.Button("Clear Points")
         clear_all_btn = gr.Button("Clear All")
 
+    gr.Markdown("Click any row to load an example.", elem_classes=["example"])
+    
+    examples = [
+        [car_on_road_video],
+        [car_in_desert_video],
+    ]
+
+    gr.Examples(
+        examples=examples,
+        inputs=[video_in],
+        outputs=[output_video],
+        fn=example_pipeline,
+        cache_examples=False,
+        run_on_click=True 
+    )
+
+    gr.Markdown("Example videos credit: https://pixabay.com/")
+
     video_in.change(fn=load_first_frame, inputs=video_in, outputs=first_frame)
     first_frame.select(fn=select_point, inputs=first_frame, outputs=first_frame)
     clear_pts_btn.click(fn=clear_points, outputs=first_frame)

examples/car.mp4

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:40b4545bba02d4f35238c8ea4382bc96e7e17d192bd115c7c0f6fe781d5717a3
+size 2088343

examples/desert_car.mp4

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5e8c2492b28d60539b5d61dec2a36a94f61a2f04707eab6f2ab6830d672c0f50
+size 3971502