Create app.py

app.py

@@ -0,0 +1,43 @@
+import streamlit as st
+from transformers import AutoModelForCausalLM, AutoTokenizer
+from PIL import Image
+import requests
+from io import BytesIO
+
+model_name = "nlpconnect/vit-gpt2-image-captioning"
+model = AutoModelForCausalLM.from_pretrained(model_name)
+tokenizer = AutoTokenizer.from_pretrained(model_name)
+
+def generate_caption(image):
+    image = image.convert("RGB")
+    
+    image = image.resize((224, 224))
+    
+    inputs = tokenizer("Image caption: ", return_tensors="pt", max_length=30, truncation=True)
+
+    with st.spinner("Generating caption..."):
+        caption_ids = model.generate(input_ids=inputs["input_ids"], attention_mask=inputs["attention_mask"])
+
+    generated_caption = tokenizer.decode(caption_ids[0], skip_special_tokens=True)
+    
+    return generated_caption
+
+def main():
+    st.title("Image Captioning App")
+    
+    with st.form("my_form"):
+        uploaded_file = st.file_uploader("Choose an image file", type=["jpg", "jpeg", "png"])
+
+        if uploaded_file is not None:
+            image = Image.open(uploaded_file)
+            st.image(image, caption="Uploaded Image", use_column_width=True)
+
+        clicked = st.form_submit_button("Generate Caption")
+        
+        if clicked and uploaded_file is not None:
+            caption = generate_caption(image)
+            st.success("Generated Caption:")
+            st.write(caption)
+
+if __name__ == "__main__":
+    main()