create function on image.py and call it on app.py

app.py

@@ -1,7 +1,11 @@
+from image import *
 import streamlit as st
 
-import os
 x = st.slider('Select a value')
 st.write(x, 'squared is', x * x)
 
-os.system("python image.py")
+
+url = "https://i.imgur.com/qs0CxjE_d.webp?maxwidth=760&fidelity=grand"
+text = "What is the two color's car ?"
+
+st.write(image(url, text))

image.py

@@ -1,24 +1,25 @@
 from transformers import ViltProcessor, ViltForQuestionAnswering
-import streamlit as st
 import requests
 from PIL import Image
+def image(url, text):
+    image = Image.open(requests.get(url, stream=True).raw)
+    processor = ViltProcessor.from_pretrained("dandelin/vilt-b32-finetuned-vqa")
+    model = ViltForQuestionAnswering.from_pretrained("dandelin/vilt-b32-finetuned-vqa")
 
-# prepare image + question
-url = "https://i.imgur.com/qs0CxjE_d.webp?maxwidth=760&fidelity=grand"
-image = Image.open(requests.get(url, stream=True).raw)
-text = "What is the two color's car ?"
+    # prepare inputs
+    encoding = processor(image, text, return_tensors="pt")
+
+    # forward pass
+    outputs = model(**encoding)
+    logits = outputs.logits
+    idx = logits.argmax(-1).item()
+    print("question asked:", text)
+    print("image link:", url)
+    print("Predicted answer:", model.config.id2label[idx])
+
+    return "Predicted answer:", model.config.id2label[idx]
 
-processor = ViltProcessor.from_pretrained("dandelin/vilt-b32-finetuned-vqa")
-model = ViltForQuestionAnswering.from_pretrained("dandelin/vilt-b32-finetuned-vqa")
 
-# prepare inputs
-encoding = processor(image, text, return_tensors="pt")
 
-# forward pass
-outputs = model(**encoding)
-logits = outputs.logits
-idx = logits.argmax(-1).item()
-st.write("Predicted answer:", model.config.id2label[idx])
-print("question asked:", text)
-print("image link:", url)
-print("Predicted answer:", model.config.id2label[idx])
+# prepare image + question
+