Create app.py

app.py

@@ -0,0 +1,18 @@
+import streamlit as st
+from transformers import pipeline
+from transformers import ViTFeatureExtractor, ViTForImageClassification
+from PIL import Image
+import requests
+
+url = 'http://images.cocodataset.org/val2017/000000039769.jpg'
+image = Image.open(requests.get(url, stream=True).raw)
+
+feature_extractor = ViTFeatureExtractor.from_pretrained('google/vit-base-patch16-224')
+model = ViTForImageClassification.from_pretrained('google/vit-base-patch16-224')
+
+inputs = feature_extractor(images=image, return_tensors="pt")
+outputs = model(**inputs)
+logits = outputs.logits
+
+predicted_class_idx = logits.argmax(-1).item()
+print("Predicted class:", model.config.id2label[predicted_class_idx])