Update app.py

app.py

@@ -1,16 +1,16 @@
-import streamlit as st
-from transformers import DetrImageProcessor, DetrForObjectDetection
+import io
+import json
 import torch
 from PIL import Image
-import requests
+from transformers import DetrImageProcessor, DetrForObjectDetection
 
-st.set_page_config(page_title="SnapSpot", page_icon="📸", layout="wide", initial_sidebar_state="collapsed")
+# Initialize the DETR model and processor
+processor = DetrImageProcessor.from_pretrained("facebook/detr-resnet-50", revision="no_timm")
+model = DetrForObjectDetection.from_pretrained("facebook/detr-resnet-50", revision="no_timm")
 
-# Function to perform object detection
-def detect_objects(image):
-    # Load DETR model and processor
-    processor = DetrImageProcessor.from_pretrained("facebook/detr-resnet-50", revision="no_timm")
-    model = DetrForObjectDetection.from_pretrained("facebook/detr-resnet-50", revision="no_timm")
+def predict(inputs):
+    # Load the image from the provided inputs
+    image = Image.open(io.BytesIO(inputs["image"]))
 
     # Preprocess the image
     inputs = processor(images=image, return_tensors="pt")
@@ -22,48 +22,16 @@ def detect_objects(image):
     target_sizes = torch.tensor([image.size[::-1]])
     results = processor.post_process_object_detection(outputs, target_sizes=target_sizes, threshold=0.9)[0]
 
-    return results
-
-# Main Streamlit app
-def main():
-    st.title("SnapSpot")
-    st.markdown(
-        """
-        <style>
-            .reportview-container {
-                background: #0e1117;
-                color: #f0f6fc;
-            }
-            .st-bq {
-                background-color: #0e1117;
-            }
-            .st-bm {
-                padding-top: 2rem;
-            }
-        </style>
-        """,
-        unsafe_allow_html=True,
-    )
-
-    # Upload image
-    uploaded_image = st.file_uploader("Upload an image", type=["jpg", "jpeg", "png"])
-
-    if uploaded_image is not None:
-        # Display uploaded image
-        image = Image.open(uploaded_image)
-        st.image(image, caption="Uploaded Image", use_column_width=True)
+    # Prepare the results in a dictionary format
+    detections = [{"label": model.config.id2label[label.item()], "confidence": score.item(), "box": box.tolist()} 
+                  for score, label, box in zip(results["scores"], results["labels"], results["boxes"])]
 
-        # Perform object detection
-        results = detect_objects(image)
+    return detections
 
-        # Display detection results
-        st.subheader("Detection Results:")
-        for score, label, box in zip(results["scores"], results["labels"], results["boxes"]):
-            box = [round(i, 2) for i in box.tolist()]
-            st.write(
-                f"Detected {model.config.id2label[label.item()]} with confidence "
-                f"{round(score.item(), 3)} at location {box}"
-            )
+# Define the API endpoint for Hugging Face Spaces
+def huggingface_spaces_endpoint(inputs):
+    # Call the predict function with the provided inputs
+    detections = predict(inputs)
 
-if __name__ == "__main__":
-    main()
+    # Return the detections as a JSON object
+    return json.dumps(detections)