allow users to pick inference resolution

.gitignore

@@ -0,0 +1,3 @@
+.idea/
+venv/
+*.pth

app.py

@@ -4,6 +4,8 @@ import supervision as sv
 from rfdetr import RFDETRBase, RFDETRLarge
 from rfdetr.util.coco_classes import COCO_CLASSES
 
+from utils.video import create_directory
+
 MARKDOWN = """
 # RF-DETR 🔥
 
@@ -24,9 +26,9 @@ by [Roboflow](https://roboflow.com/) and released under the Apache 2.0 license.
 """
 
 IMAGE_EXAMPLES = [
-    ['https://media.roboflow.com/supervision/image-examples/people-walking.png', 0.3, "large"],
-    ['https://media.roboflow.com/supervision/image-examples/vehicles.png', 0.3, "large"],
-    ['https://media.roboflow.com/notebooks/examples/dog-2.jpeg', 0.5, "base"],
+    ['https://media.roboflow.com/supervision/image-examples/people-walking.png', 0.3, 728, "large"],
+    ['https://media.roboflow.com/supervision/image-examples/vehicles.png', 0.3, 728, "large"],
+    ['https://media.roboflow.com/notebooks/examples/dog-2.jpeg', 0.5, 560, "base"],
 ]
 
 COLOR = sv.ColorPalette.from_hex([
@@ -34,15 +36,16 @@ COLOR = sv.ColorPalette.from_hex([
     "#9999ff", "#3399ff", "#66ffff", "#33ff99", "#66ff66", "#99ff00"
 ])
 
-MODEL_BASE = RFDETRBase(resolution=728)
-MODEL_LARGE = RFDETRLarge(resolution=728)
+VIDEO_SCALE_FACTOR = 0.5
+VIDEO_TARGET_DIRECTORY = "tmp"
+create_directory(directory_path=VIDEO_TARGET_DIRECTORY)
 
 
 @spaces.GPU()
-def inference(image, confidence: float, checkpoint: str):
-    detections = MODEL_BASE.predict(image, threshold=confidence) \
-        if checkpoint == "base" \
-        else MODEL_LARGE.predict(image, threshold=confidence)
+def inference(image, confidence: float, resolution: int, checkpoint: str):
+    model_class = RFDETRBase if checkpoint == "base" else RFDETRLarge
+    model = model_class(resolution=resolution)
+    detections = model.predict(image, threshold=confidence)
 
     text_scale = sv.calculate_optimal_text_scale(resolution_wh=image.size)
     thickness = sv.calculate_optimal_line_thickness(resolution_wh=image.size)
@@ -83,6 +86,13 @@ with gr.Blocks() as demo:
                 step=0.05,
                 value=0.5,
             )
+            resolution_slider = gr.Slider(
+                label="Inference resolution",
+                minimum=560,
+                maximum=1120,
+                step=56,
+                value=728,
+            )
             with gr.Row():
                 checkpoint_dropdown = gr.Dropdown(
                     label="Checkpoint",
@@ -100,13 +110,13 @@ with gr.Blocks() as demo:
     gr.Examples(
         fn=inference,
         examples=IMAGE_EXAMPLES,
-        inputs=[input_image, confidence_slider, checkpoint_dropdown],
+        inputs=[input_image, confidence_slider, resolution_slider, checkpoint_dropdown],
         outputs=output_image
     )
 
     submit_button.click(
         inference,
-        inputs=[input_image, confidence_slider, checkpoint_dropdown],
+        inputs=[input_image, confidence_slider, resolution_slider, checkpoint_dropdown],
         outputs=output_image
     )
 

utils/video.py

@@ -0,0 +1,26 @@
+import datetime
+import os
+import shutil
+import uuid
+
+
+def create_directory(directory_path: str) -> None:
+    if not os.path.exists(directory_path):
+        os.makedirs(directory_path)
+
+
+def delete_directory(directory_path: str) -> None:
+    if not os.path.exists(directory_path):
+        raise FileNotFoundError(f"Directory '{directory_path}' does not exist.")
+
+    try:
+        shutil.rmtree(directory_path)
+    except PermissionError:
+        raise PermissionError(
+            f"Permission denied: Unable to delete '{directory_path}'.")
+
+
+def generate_unique_name():
+    current_datetime = datetime.datetime.now().strftime("%Y%m%d%H%M%S")
+    unique_id = uuid.uuid4()
+    return f"{current_datetime}_{unique_id}"