Create app.py

app.py

@@ -0,0 +1,34 @@
+import re
+import gradio as gr
+import torch
+from transformers import UdopProcessor, UdopForConditionalGeneration
+
+repo_id = "microsoft/udop-large"
+
+processor = UdopProcessor.from_pretrained(repo_id)
+model = UdopForConditionalGeneration.from_pretrained(repo_id)
+
+def process_document(image, question):
+    
+    pixel_values = processor(image, return_tensors="pt").pixel_values
+    encoding = processor(images=image, text=question, return_tensors="pt")
+    outputs = model.generate(**encoding, max_new_tokens=20)
+    generated_text = processor.batch_decode(outputs, skip_special_tokens=True)[0]
+    
+    return generated_text
+
+description = "Unofficial Gradio Demo for UDOP for DocVQA (document visual question answering). To use it, simply upload your image and type a question and click 'submit', or click one of the examples to load them. Read more at the links below."
+article = "<p style='text-align: center'><a href='https://arxiv.org/pdf/2212.02623.pdf' target='_blank'>Unifying Vision, Text, and Layout for Universal Document Processing</a> | <a href='https://github.com/microsoft/UDOP' target='_blank'>Github Repo</a></p>"
+
+demo = gr.Interface(
+    fn=process_document,
+    inputs=["image", "text"],
+    outputs="text",
+    title="Demo: UDOP for DocVQA",
+    description=description,
+    article=article,
+    enable_queue=True,
+    # examples=[["example_1.png", "When is the coffee break?"], ["example_2.jpeg", "What's the population of Stoddard?"]],
+    cache_examples=False)
+
+demo.launch()