Update app.py

app.py

@@ -8,8 +8,6 @@ device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
 processor = AutoProcessor.from_pretrained("microsoft/udop-large", apply_ocr=True)
 model = UdopForConditionalGeneration.from_pretrained("microsoft/udop-large")
 
-question = "Question answering. How many unsafe practice of Lifting Operation"
-
 st.title("CIC Demo (by ITT)")
 st.write("Upload and Select a document (/an image) to test the model.")
 
@@ -21,17 +19,25 @@ selected_file = st.selectbox("Select a document (/an image):", uploaded_files, f
 if selected_file is not None and selected_file != "None":
     file_extension = selected_file.name.split(".")[-1]
     if file_extension in ["jpg", "jpeg", "png"]:
-        st.image(selected_file, caption="Selected Image")
+        image = st.image(selected_file, caption="Selected Image")
     else:
         st.write("Selected file: ", selected_file.name)
 
-# Model testing button
+# Model Testing 
+## Test button
 testButton = st.button("Test Model")
+
+## Question (/Prompt)
+question = "Question answering. How many unsafe practice of Lifting Operation?"
+
+## Perform Model Testing when Image is uploaded and selected as well as Test button is pressed
 if testButton and selected_file != "None":
     st.write("Testing the model with the selected image...")
-    encoding = processor(image, question, words, boxes=boxes, return_tensors="pt")
-    predicted_ids = model.generate(**encoding)
-    print(processor.batch_decode(predicted_ids, skip_special_tokens=True)[0])    
+    # encoding = processor(image, question, words, boxes=boxes, return_tensors="pt")
+    model_encoding = processor(iamges=image, text=question, return_tensors="pt")
+    model_output = model.generate(**model_encoding)
+    output_text = processor.batch_decode(model_output, skip_special_tokens=True)[0]
+    st.write(output_text)    
 elif testButton and selected_file == "None":
     st.write("Please upload and select a document (/an image).")