Update app.py

app.py

@@ -16,60 +16,81 @@ if 'models_loaded' not in st.session_state:
     st.session_state.models_loaded = True
 
 # Define image processing and classification functions
-def imageprocessing(image):
-    encoding = st.session_state.image_processor(image.convert("RGB"), return_tensors="pt")
-    st.write(encoding)
-    return encoding
-
-def topwear(encoding):
+def topwear(encoding,top_wear_model):
+    # Make prediction
     outputs = st.session_state.top_wear_model(**encoding)
-    predicted_class_idx = outputs.logits.argmax(-1).item()
+    logits = outputs.logits
+    predicted_class_idx = logits.argmax(-1).item()
+    # Print the result
     return st.session_state.top_wear_model.config.id2label[predicted_class_idx]
 
-def patterns(encoding):
+def patterns(encoding,pattern_model):
+    # Make prediction
     outputs = st.session_state.pattern_model(**encoding)
-    predicted_class_idx = outputs.logits.argmax(-1).item()
+    logits = outputs.logits
+    predicted_class_idx = logits.argmax(-1).item()
+    # Print the result
     return st.session_state.pattern_model.config.id2label[predicted_class_idx]
 
-def prints(encoding):
+def prints(encoding,print_model):
+    # Make prediction
     outputs = st.session_state.print_model(**encoding)
-    predicted_class_idx = outputs.logits.argmax(-1).item()
+    logits = outputs.logits
+    predicted_class_idx = logits.argmax(-1).item()
+    # Print the result
     return st.session_state.print_model.config.id2label[predicted_class_idx]
 
-def sleevelengths(encoding):
+def sleevelengths(encoding,sleeve_length_model):
+    # Make prediction
     outputs = st.session_state.sleeve_length_model(**encoding)
-    predicted_class_idx = outputs.logits.argmax(-1).item()
+    logits = outputs.logits
+    predicted_class_idx = logits.argmax(-1).item()
+    # Print the result
     return st.session_state.sleeve_length_model.config.id2label[predicted_class_idx]
 
+def imageprocessing(url):
+    response = requests.get(url)
+    if response.status_code == 200:
+        image = Image.open(BytesIO(response.content))
+        encoding = image_processor(image.convert("RGB"), return_tensors="pt")
+    return encoding
+    
+# Define the function that will be used in each thread
+def call_model(func, encoding, model, results, index):
+    results[index] = func(encoding, model)
 # Run all models in parallel
-def pipes(image):
-    encoding = imageprocessing(image)
+def pipes(imagepath):
+    # Process the image once and reuse the encoding
+    encoding = imageprocessing(imagepath)
 
+    # Prepare a list to store the results from each thread
     results = [None] * 4
 
-    def update_results(index, func):
-        results[index] = func(encoding)
-
+    # Create threads for each function call
     threads = [
-        threading.Thread(target=update_results, args=(0, topwear)),
-        threading.Thread(target=update_results, args=(1, patterns)),
-        threading.Thread(target=update_results, args=(2, prints)),
-        threading.Thread(target=update_results, args=(3, sleevelengths)),
+        threading.Thread(target=call_model, args=(topwear, encoding, top_wear_model, results, 0)),
+        threading.Thread(target=call_model, args=(patterns, encoding, pattern_model, results, 1)),
+        threading.Thread(target=call_model, args=(prints, encoding, print_model, results, 2)),
+        threading.Thread(target=call_model, args=(sleevelengths, encoding, sleeve_length_model, results, 3)),
     ]
 
+    # Start all threads
     for thread in threads:
         thread.start()
+
+    # Wait for all threads to finish
     for thread in threads:
         thread.join()
 
-    result_dict = {
-        "topwear": results[0],
+    # Combine the results into a dictionary
+    dicts = {
+        "top": results[0],
         "pattern": results[1],
         "print": results[2],
         "sleeve_length": results[3]
     }
 
-    return result_dict
+    return dicts
 
 # Streamlit app UI
 st.title("Clothing Classification Pipeline")
@@ -79,12 +100,13 @@ if url:
     response = requests.get(url)
     if response.status_code == 200:
         image = Image.open(BytesIO(response.content))
+        encoding = image_processor(image.convert("RGB"), return_tensors="pt")
         st.image(image.resize((200, 200)), caption="Uploaded Image", use_column_width=False)
         
         start_time = time.time()
 
         try:
-            result = pipes(image)
+            result = pipes(url)
             st.write("Classification Results (JSON):")
             st.json(result)  # Display results in JSON format
             st.write(f"Time taken: {time.time() - start_time:.2f} seconds")