a

app.py

@@ -16,18 +16,27 @@ if(input_image):
     image = Image.open(input_image)
     st.header("Original")
     st.image(input_image, use_column_width=True)
+
+    detection_model = st.selectbox("Select detection model", ("YOLO"))
+    embedding_model = st.selectbox("Select detection model", ("OML"))
+
     if st.button("Get prediction"):
         with st.spinner("Loading..."):
-            response = requests.post(os.path.join(api, "bounding-boxes/"), files = {"file": input_image.getvalue()})
+            response = requests.post(os.path.join(api, f"bounding-boxes-{detection_model}/"),
+                                    files = {"file": input_image.getvalue(), "model_id": detection_model})
             prediction = ast.literal_eval(response.text)
-            response = requests.post(os.path.join(api, "image-w-boxes/"), files = {"file": input_image.getvalue()})
+            response = requests.post(os.path.join(api, f"image-w-boxes-{detection_model}/"),
+                                    files = {"file": input_image.getvalue(), "model_id": detection_model})
+            
             image_with_boxes = response.content
             arr = []
             for b in prediction["bboxes"]:
                 stamp = image.crop((b["xmin"], b["ymin"], b["xmin"] + b["width"], b["ymin"] + b["height"]))
                 output = io.BytesIO()
                 stamp.save(output, format="BMP")
-                response = ast.literal_eval(requests.post(os.path.join(api, "embeddings-from-cropped/"), files = {"file": output.getvalue()}).text)
+                response = ast.literal_eval(requests.post(os.path.join(api, f"embeddings-from-cropped-{embedding_model}/"),
+                                                        files = {"file": output.getvalue(), "model_id": embedding_model}).text)
+                
                 arr.extend(response["embedding"])
 
         col1, col2, col3 = st.columns(3)