Create app.py

app.py

@@ -0,0 +1,29 @@
+import streamlit as st
+from transformers import pipeline
+from PIL import Image
+
+#pipeline 1
+generator = pipeline("fill-mask", model="distilbert/distilbert-base-uncased")
+
+#pipeline 2
+classifier = pipeline("text-classification", model="swangfr/distilbert-multi-label-amazon", return_all_scores=True)
+
+
+# Streamlit application title
+st.title("Amazon Product Image classifier")
+st.write("Classification for 24 categories")
+
+#upload image file
+file_name = st.file_uploader("Upload a product image file")
+
+if st.button("Classify") & file_name is not None:
+    col1, col2 = st.columns(2)
+
+    image = Image.open(file_name)
+    col1.image(image, use_column_width=True)
+    generation = generator(image)
+    prediction = classifier(str(generation[0]['generated_text']))
+
+    col2.header("Probabilities")
+    for p in predictions:
+        col2.subheader(f"{ p['label'] }: { round(p['score'] * 100, 1)}%")