1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18

SUPPORTED_TASKS = {
    # 其他省略......

    "question-answering": {
        "impl": QuestionAnsweringPipeline,
        "tf": (TFAutoModelForQuestionAnswering,) if is_tf_available() else (),
        "pt": (AutoModelForQuestionAnswering,) if is_torch_available() else (),
        "default": {
            "model": {
                "pt": ("distilbert/distilbert-base-cased-distilled-squad", "564e9b5"),
                "tf": ("distilbert/distilbert-base-cased-distilled-squad", "564e9b5"),
            },
        },
        "type": "text",
    },

    # 其他省略......
}

1

pip install transformers torch

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29

import os

# 设置代理
os.environ["HF_ENDPOINT"] = "https://hf-mirror.com"

# 设置本地缓存目录
cache_dir = os.path.join('D:', os.path.sep, 'ModelSpace', 'Cache')
os.environ['HF_HOME'] = cache_dir

from transformers import pipeline

# 创建Pipeline任务
nlp = pipeline("question-answering", model="distilbert/distilbert-base-cased-distilled-squad")

# 执行问答任务
if __name__ == "__main__":
    # 上下文
    context = r"""
    Extractive Question Answering is the task of extracting an answer from a text given a question.
    An example of a question answering dataset is the SQuAD dataset, which is entirely based on that task.
    If you would like to fine-tune a model on a SQuAD task,
    you may leverage the examples/pytorch/question-answering/run_squad.py script.
    """

    # 执行任务
    result = nlp(context=context, question="What is a good example of a question answering dataset?")

    print(result)
    # 输出：{'score': 0.5152314901351929, 'start': 155, 'end': 168, 'answer': 'SQuAD dataset'}

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14

from transformers import AutoTokenizer, AutoModelForQuestionAnswering, pipeline
import os

# 下载模型目录
model_dir = '/models/question_answering'

# 加载分词器和模型
tokenizer = AutoTokenizer.from_pretrained(model_dir, local_files_only=True)
model = AutoModelForQuestionAnswering.from_pretrained(model_dir, torch_dtype="auto", device_map="auto", local_files_only=True)

# 创建Pipeline任务
nlp = pipeline("question-answering", tokenizer=tokenizer, model=model)

# 后续用法与自动下载相同

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39

import sys

# 直接复用Pipeline实例
sys.path.append("./")
pipeline = __import__("04-question-answering")

import gradio as gr


# 定义问答函数
def answer_question(question, context):
    result = pipeline.nlp(question=question, context=context)
    return f"答案: {result['answer']}, 置信度: {result['score']:.4f}"


# 创建Gradio界面
with gr.Blocks() as demo:
    gr.Markdown("# 问答系统")
    gr.Markdown(
        "这是一个基于Transformers框架的问答工具。您可以输入一个问题和一段文本，点击“提交”按钮后，系统将尝试从中找到答案。")

    with gr.Row():
        input_context = gr.Textbox(placeholder="请输入相关文本...", label="上下文")

    with gr.Row():
        input_question = gr.Textbox(placeholder="请输入您的问题...", label="问题")

    with gr.Row():
        submit_button = gr.Button("提交")

    with gr.Row():
        output_answer = gr.Label(label="答案")

    # 设置按钮点击事件，触发问答函数
    submit_button.click(answer_question, inputs=[input_question, input_context], outputs=output_answer)

# 启动Gradio应用
if __name__ == "__main__":
    demo.launch()

1
2
3

* Running on local URL:  http://127.0.0.1:7860

To create a public link, set `share=True` in `launch()`.