feat: add label to input component

app.py

@@ -19,6 +19,15 @@ model.load_state_dict(torch.load("./bin/model.ckpt", map_location=device)["state
 
 
 def augment(img: np.ndarray) -> torch.Tensor:
+    """
+    Helper Function to augment the image before we generate embeddings
+
+    Args:
+        img (np.ndarray): Input Image
+
+    Returns:
+        torch.Tensor
+    """
     img = Image.fromarray(img)
     if img.mode == "L":
         # Convert grayscale image to RGB by duplicating the single channel three times
@@ -31,14 +40,26 @@ def augment(img: np.ndarray) -> torch.Tensor:
     return transforms(img).unsqueeze(0)
 
 
-def search_index(input_image, k: int = 1):
+def search_index(input_image: np.ndarray, k: int = 1) -> list:
+    """
+    Retrieve the Top k images from the given input image
+
+    Args:
+        input_image (np.ndarray): Input Image
+        k (int): number of images to fetch
+
+    Returns:
+        list: List of top k images retrieved using the embeddings
+          generated from the input image
+    """
+    images = []
+
     with torch.no_grad():
         embedding = model(augment(input_image).to(device))
         index = read_index("./bin/dino.index")
         _, results = index.search(np.array(embedding[0].reshape(1, -1)), k)
         indices = results[0]
-        images = []
-        for i, index in enumerate(indices[:k]):
+        for _, index in enumerate(indices[:k]):
             retrieved_img = dataset["train"][int(index)]["image"]
             images.append(retrieved_img)
     return images
@@ -47,7 +68,7 @@ def search_index(input_image, k: int = 1):
 app = gr.Interface(
     search_index,
     inputs=[
-        gr.Image(),
+        gr.Image(label="Input Image"),
         gr.Slider(minimum=1, maximum=10, value=3, step=1, label="Top K"),
     ],
     outputs=[