tomcat/Ask_and_Answer_v2
0
1import gradio as gr2from transformers import pipeline3 4pp_en = pipeline("question-answering",model="deepset/roberta-base-squad2")5pp_ch = pipeline("question-answering",model="luhua/chinese_pretrain_mrc_roberta_wwm_ext_large")6 7def qa_fn(ask,ctxt,model):8 pp = (pp_en if model=="english" else pp_ch);9 ret = pp(context=ctxt, question=ask);10 ret['entity']='Answer';11 return {"text":ctxt,"entities":[ret]}, ret['answer'], ret['score'] 12 #注意HighlightedText的用法。有两种不同用法:https://gradio.app/named_entity_recognition/13 # 一种是list of dict ,一种是list of tuple. 详细用法参考https://gradio.app/named_entity_recognition/吧14 15samples= [16 ["乔治的哥哥叫什么名字?","我是小猪佩奇,我是乔治的哥哥,我家住在北京","english"],17 ["图书馆主页的网址是多少啊?","读者进入图书馆主页(http://lib.tjut.edu.cn)后,点击文献传递菜单,即可查看文献传递的具体流程步骤","chinese"],18 ];19introStr = "用于演示使用人工智能自动寻找问题答案,这将是一种更加高效便捷的新型信息检索方式。";20titleStr = "智能问答演示程序";21 22demo = gr.Interface(qa_fn, 23 inputs=[gr.Textbox(label="Question",placeholder='请输入问题'), 24 gr.Textbox(label="Context",lines=10,placeholder="请输入一段文本"),25 gr.Radio(["english","chinese"],label="Select a Model", value="english"),26 ],27 outputs=[gr.HighlightedText(label='答案位置'),gr.Textbox(label="答案"),gr.Number(label="Score")],28 examples=samples,29 description=introStr,30 title=titleStr);31 32demo.launch()33 