用30行Python代码实现一个AI文本摘要工具

# 用30行Python代码实现一个AI文本摘要工具 最近需要批量处理新闻摘要，调API太贵了，干脆用开源模型搞一个本地版。 ## 核心代码 ```python # 只需要30行就能跑起来 from transformers import pipeline # 加载预训练的摘要模型 summarizer = pipeline( "summarization", model="facebook/bart-large-cnn", device=0 # 用GPU加速，没有GPU就改成-1 ) def summarize(text, max_length=130, min_length=30): """生成文本摘要""" result = summarizer( text, max_length=max_length, min_length=min_length, do_sample=False # 贪心解码，结果更稳定 ) return result[0]['summary_text'] # 测试一下 article = """ Artificial intelligence has made significant progress in recent years, particularly in the field of natural language processing. Large language models can now generate human-like text, translate languages, and even write code. However, concerns about AI safety and alignment continue to grow among researchers and policymakers. """ print(summarize(article)) # 输出：AI has made significant progress in NLP. LLMs can generate text, # translate and write code. AI safety concerns continue to grow. ``` ## 处理中文 ```python # 中文摘要用这个模型 summarizer_zh = pipeline( "summarization", model="fnlp/bart-base-chinese" ) ``` BART模型效果不错，处理一篇1000字的文章大概0.5秒（RTX 3060）。如果量不大，比调API划算多了。