Update app.py

app.py

@@ -1,3 +1,22 @@
 import gradio as gr
+from transformers import AutoModelForQuestionAnswering, pipeline,AutoTokenizer
+import torch
+from optimum.onnxruntime import ORTModelForQuestionAnswering
+def question_answer(context, question):
+  AUTH_TOKEN = "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"
+  model_checkpoint = "letrunglinh/qa_pnc"
+  device = "cuda" if torch.cuda.is_available() else "cpu"
+  tokenizer = AutoTokenizer.from_pretrained(model_checkpoint, use_auth_token=AUTH_TOKEN)
+  model = ORTModelForQuestionAnswering.from_pretrained(model_checkpoint, from_transformers=True)
+  model = pipeline('question-answering', model=model,
+  tokenizer=tokenizer, use_auth_token=AUTH_TOKEN)
+  to_predict = [
+    {
+        "question": question,
+        "context": context,
+    }
+    ]
+  answers = model(to_predict)
+  return answers['answer'], answers['score']
 
-gr.Interface.load("models/letrunglinh/qa_pnc").launch()
+gr.Interface(fn=question_answer, inputs=["text", "text"], outputs=["textbox","textbox"], share = True).launch()