{"id":9368,"date":"2024-07-10T02:23:45","date_gmt":"2024-07-09T18:23:45","guid":{"rendered":"\/?p=9368"},"modified":"2024-07-10T02:23:45","modified_gmt":"2024-07-09T18:23:45","slug":"huggingface%e7%9a%84%e7%ae%a1%e9%81%93%e5%b7%a5%e5%85%b7pipeline","status":"publish","type":"post","link":"\/?p=9368","title":{"rendered":"HuggingFace\u7684\u7ba1\u9053\u5de5\u5177pipeline"},"content":{"rendered":"

\u4ec0\u4e48\u662f Hugging Face \u7684 pipeline<\/code>\uff1f<\/h3>\n

Hugging Face \u63d0\u4f9b\u7684 transformers<\/code> \u5e93\u662f\u4e00\u4e2a\u975e\u5e38\u5f3a\u5927\u7684\u5de5\u5177\u96c6\uff0c\u7528\u4e8e\u81ea\u7136\u8bed\u8a00\u5904\u7406\uff08NLP\uff09\u4efb\u52a1\u3002pipeline<\/code> \u662f\u8fd9\u4e2a\u5e93\u4e2d\u7684\u4e00\u4e2a\u6838\u5fc3\u5de5\u5177\uff0c\u65e8\u5728\u7b80\u5316\u5404\u79cdNLP\u4efb\u52a1\u7684\u5b9e\u73b0\u3002\u4ee5\u4e0b\u662f pipeline<\/code> \u7684\u4ecb\u7ecd\u53ca\u5176\u4f7f\u7528\u65b9\u6cd5\uff1a<\/p>\n

pipeline<\/code> \u662f\u4e00\u4e2a\u9ad8\u7ea7\u63a5\u53e3\uff0c\u53ef\u4ee5\u8ba9\u4f60\u65b9\u4fbf\u5730\u4f7f\u7528\u9884\u8bad\u7ec3\u7684\u6a21\u578b\u6765\u5904\u7406NLP\u4efb\u52a1\u3002\u5b83\u5c01\u88c5\u4e86\u6a21\u578b\u52a0\u8f7d\u3001\u9884\u5904\u7406\u3001\u63a8\u7406\u548c\u540e\u5904\u7406\u7b49\u6b65\u9aa4\uff0c\u4f7f\u5f97\u4f60\u53ef\u4ee5\u5feb\u901f\u4e0a\u624b\u8fdb\u884c\u6587\u672c\u5206\u7c7b\u3001\u60c5\u611f\u5206\u6790\u3001\u7ffb\u8bd1\u3001\u6587\u672c\u751f\u6210\u7b49\u4efb\u52a1\u3002<\/p>\n

\u652f\u6301\u7684\u4efb\u52a1<\/h3>\n

pipeline<\/code> \u652f\u6301\u7684\u4efb\u52a1\u5305\u62ec\u4f46\u4e0d\u9650\u4e8e\uff1a<\/p>\n

    \n
  • \u6587\u672c\u5206\u7c7b<\/strong> (Text Classification)<\/li>\n
  • \u547d\u540d\u5b9e\u4f53\u8bc6\u522b<\/strong> (Named Entity Recognition, NER)<\/li>\n
  • \u95ee\u7b54\u7cfb\u7edf<\/strong> (Question Answering)<\/li>\n
  • \u6587\u672c\u751f\u6210<\/strong> (Text Generation)<\/li>\n
  • \u7ffb\u8bd1<\/strong> (Translation)<\/li>\n
  • \u603b\u7ed3<\/strong> (Summarization)<\/li>\n
  • \u586b\u7a7a<\/strong> (Fill-Mask)<\/li>\n
  • \u7279\u5f81\u62bd\u53d6<\/strong> (Feature Extraction)<\/li>\n<\/ul>\n

    \u5982\u4f55\u4f7f\u7528 pipeline<\/code><\/h3>\n

    \u4ee5\u4e0b\u662f\u4e00\u4e9b\u5e38\u89c1\u4efb\u52a1\u7684\u793a\u4f8b\u4ee3\u7801\uff1a<\/p>\n

    1. \u6587\u672c\u5206\u7c7b\uff08\u60c5\u611f\u5206\u6790\uff09<\/h4>\n
    from transformers import pipeline\n\n# \u521b\u5efa\u60c5\u611f\u5206\u6790\u7ba1\u9053\nclassifier = pipeline('sentiment-analysis')\n\n# \u4f7f\u7528\u7ba1\u9053\u8fdb\u884c\u60c5\u611f\u5206\u6790\nresult = classifier("I love using Hugging Face's transformers library!")\nprint(result)<\/code><\/pre>\n
    2. \u547d\u540d\u5b9e\u4f53\u8bc6\u522b\uff08NER\uff09<\/h4>\n
    from transformers import pipeline\n\n# \u521b\u5efaNER\u7ba1\u9053\nner = pipeline('ner', grouped_entities=True)\n\n# \u4f7f\u7528\u7ba1\u9053\u8fdb\u884c\u547d\u540d\u5b9e\u4f53\u8bc6\u522b\nresult = ner("Hugging Face Inc. is a company based in New York City.")\nprint(result)<\/code><\/pre>\n
    3. \u95ee\u7b54\u7cfb\u7edf<\/h4>\n
    from transformers import pipeline\n\n# \u521b\u5efa\u95ee\u7b54\u7ba1\u9053\nqa = pipeline('question-answering')\n\n# \u4f7f\u7528\u7ba1\u9053\u8fdb\u884c\u95ee\u7b54\nresult = qa(question="Where is Hugging Face based?", context="Hugging Face Inc. is a company based in New York City.")\nprint(result)<\/code><\/pre>\n
    4. \u6587\u672c\u751f\u6210<\/h4>\n
    from transformers import pipeline\n\n# \u521b\u5efa\u6587\u672c\u751f\u6210\u7ba1\u9053\ngenerator = pipeline('text-generation', model='gpt-2')\n\n# \u4f7f\u7528\u7ba1\u9053\u751f\u6210\u6587\u672c\nresult = generator("Once upon a time, in a land far, far away", max_length=50, num_return_sequences=1)\nprint(result)<\/code><\/pre>\n
    5. \u7ffb\u8bd1<\/h4>\n
    from transformers import pipeline\n\n# \u521b\u5efa\u7ffb\u8bd1\u7ba1\u9053\uff08\u82f1\u8bed\u5230\u6cd5\u8bed\uff09\ntranslator = pipeline('translation_en_to_fr')\n\n# \u4f7f\u7528\u7ba1\u9053\u8fdb\u884c\u7ffb\u8bd1\nresult = translator("Hugging Face is creating a tool that democratizes AI.", max_length=40)\nprint(result)<\/code><\/pre>\n
    \u5207\u6362\u6a21\u578b\u7684\u6b65\u9aa4<\/h3>\n
    Hugging Face \u7684 pipeline<\/code> \u5de5\u5177\u652f\u6301\u901a\u8fc7\u6307\u5b9a\u6a21\u578b\u540d\u79f0\u6765\u5207\u6362\u4e0d\u540c\u7684\u9884\u8bad\u7ec3\u6a21\u578b\u3002\u4f60\u53ef\u4ee5\u9009\u62e9\u4e0d\u540c\u7684\u6a21\u578b\u6765\u6267\u884c\u76f8\u540c\u7684\u4efb\u52a1\uff0c\u4ece\u800c\u6bd4\u8f83\u5b83\u4eec\u7684\u6027\u80fd\u6216\u9002\u5e94\u4e0d\u540c\u7684\u9700\u6c42\u3002\u4ee5\u4e0b\u662f\u5982\u4f55\u5728 pipeline<\/code> \u4e2d\u5207\u6362\u4e0d\u540c\u6a21\u578b\u7684\u793a\u4f8b\uff1a<\/p>\n
      \n
    1. \u9009\u62e9\u6a21\u578b\u540d\u79f0<\/strong>: \u4ece Hugging Face \u6a21\u578b\u5e93\u4e2d\u9009\u62e9\u4f60\u60f3\u8981\u4f7f\u7528\u7684\u6a21\u578b\u3002\u6a21\u578b\u5e93\u4e2d\u6709\u8bb8\u591a\u9884\u8bad\u7ec3\u6a21\u578b\uff0c\u6bcf\u4e2a\u6a21\u578b\u90fd\u6709\u4e00\u4e2a\u552f\u4e00\u7684\u540d\u79f0\u3002<\/li>\n
    2. \u5728 pipeline<\/code> \u4e2d\u6307\u5b9a\u6a21\u578b\u540d\u79f0<\/strong>: \u5728\u521b\u5efa pipeline<\/code> \u5bf9\u8c61\u65f6\uff0c\u901a\u8fc7 model<\/code> \u53c2\u6570\u6307\u5b9a\u4f60\u9009\u62e9\u7684\u6a21\u578b\u540d\u79f0\u3002<\/li>\n<\/ol>\n
      \u793a\u4f8b<\/h3>\n
      \u4ee5\u4e0b\u662f\u4e00\u4e9b\u5e38\u89c1\u4efb\u52a1\u5982\u4f55\u5207\u6362\u4e0d\u540c\u7684\u6a21\u578b\uff1a<\/p>\n
      1. \u6587\u672c\u5206\u7c7b\uff08\u60c5\u611f\u5206\u6790\uff09<\/h4>\n
      from transformers import pipeline\n\n# \u4f7f\u7528\u9ed8\u8ba4\u6a21\u578b\ndefault_classifier = pipeline('sentiment-analysis')\ndefault_result = default_classifier("I love using Hugging Face's transformers library!")\nprint("Default model:", default_result)\n\n# \u4f7f\u7528\u7279\u5b9a\u6a21\u578b\uff08\u4f8b\u5982distilbert-base-uncased-finetuned-sst-2-english\uff09\ncustom_classifier = pipeline('sentiment-analysis', model='distilbert-base-uncased-finetuned-sst-2-english')\ncustom_result = custom_classifier("I love using Hugging Face's transformers library!")\nprint("Custom model:", custom_result)<\/code><\/pre>\n
      2. \u547d\u540d\u5b9e\u4f53\u8bc6\u522b\uff08NER\uff09<\/h4>\n
      from transformers import pipeline\n\n# \u4f7f\u7528\u9ed8\u8ba4\u6a21\u578b\ndefault_ner = pipeline('ner', grouped_entities=True)\ndefault_result = default_ner("Hugging Face Inc. is a company based in New York City.")\nprint("Default model:", default_result)\n\n# \u4f7f\u7528\u7279\u5b9a\u6a21\u578b\uff08\u4f8b\u5982dbmdz\/bert-large-cased-finetuned-conll03-english\uff09\ncustom_ner = pipeline('ner', model='dbmdz\/bert-large-cased-finetuned-conll03-english', grouped_entities=True)\ncustom_result = custom_ner("Hugging Face Inc. is a company based in New York City.")\nprint("Custom model:", custom_result)<\/code><\/pre>\n
      3. \u95ee\u7b54\u7cfb\u7edf<\/h4>\n
      from transformers import pipeline\n\n# \u4f7f\u7528\u9ed8\u8ba4\u6a21\u578b\ndefault_qa = pipeline('question-answering')\ndefault_result = default_qa(question="Where is Hugging Face based?", context="Hugging Face Inc. is a company based in New York City.")\nprint("Default model:", default_result)\n\n# \u4f7f\u7528\u7279\u5b9a\u6a21\u578b\uff08\u4f8b\u5982distilbert-base-cased-distilled-squad\uff09\ncustom_qa = pipeline('question-answering', model='distilbert-base-cased-distilled-squad')\ncustom_result = custom_qa(question="Where is Hugging Face based?", context="Hugging Face Inc. is a company based in New York City.")\nprint("Custom model:", custom_result)<\/code><\/pre>\n
      4. \u6587\u672c\u751f\u6210<\/h4>\n
      from transformers import pipeline\n\n# \u4f7f\u7528\u9ed8\u8ba4\u6a21\u578b\ndefault_generator = pipeline('text-generation', model='gpt-2')\ndefault_result = default_generator("Once upon a time, in a land far, far away", max_length=50, num_return_sequences=1)\nprint("Default model:", default_result)\n\n# \u4f7f\u7528\u7279\u5b9a\u6a21\u578b\uff08\u4f8b\u5982gpt2-medium\uff09\ncustom_generator = pipeline('text-generation', model='gpt2-medium')\ncustom_result = custom_generator("Once upon a time, in a land far, far away", max_length=50, num_return_sequences=1)\nprint("Custom model:", custom_result)<\/code><\/pre>\n
      5. \u7ffb\u8bd1<\/h4>\n
      from transformers import pipeline\n\n# \u4f7f\u7528\u9ed8\u8ba4\u6a21\u578b\ndefault_translator = pipeline('translation_en_to_fr')\ndefault_result = default_translator("Hugging Face is creating a tool that democratizes AI.", max_length=40)\nprint("Default model:", default_result)\n\n# \u4f7f\u7528\u7279\u5b9a\u6a21\u578b\uff08\u4f8b\u5982Helsinki-NLP\/opus-mt-en-fr\uff09\ncustom_translator = pipeline('translation_en_to_fr', model='Helsinki-NLP\/opus-mt-en-fr')\ncustom_result = custom_translator("Hugging Face is creating a tool that democratizes AI.", max_length=40)\nprint("Custom model:", custom_result)<\/code><\/pre>\n
      \u603b\u7ed3<\/h3>\n
      Hugging Face \u7684 pipeline<\/code> \u5de5\u5177\u6781\u5927\u5730\u7b80\u5316\u4e86NLP\u4efb\u52a1\u7684\u5b9e\u73b0\uff0c\u4f7f\u5f97\u5373\u4fbf\u662f\u975e\u4e13\u4e1a\u4eba\u5458\u4e5f\u80fd\u8f7b\u677e\u5730\u4f7f\u7528\u9884\u8bad\u7ec3\u6a21\u578b\u5904\u7406\u590d\u6742\u7684\u81ea\u7136\u8bed\u8a00\u5904\u7406\u4efb\u52a1\u3002\u901a\u8fc7\u7b80\u5355\u7684\u51e0\u884c\u4ee3\u7801\uff0c\u4f60\u53ef\u4ee5\u5feb\u901f\u6784\u5efa\u548c\u90e8\u7f72\u5404\u79cdNLP\u5e94\u7528\u3002<\/p>\n
      \u901a\u8fc7\u5728 pipeline<\/code> \u4e2d\u6307\u5b9a\u4e0d\u540c\u7684\u6a21\u578b\u540d\u79f0\uff0c\u4f60\u53ef\u4ee5\u8f7b\u677e\u5207\u6362\u4e0d\u540c\u7684\u9884\u8bad\u7ec3\u6a21\u578b\uff0c\u6bd4\u8f83\u5b83\u4eec\u7684\u6027\u80fd\uff0c\u5e76\u9009\u62e9\u6700\u9002\u5408\u4f60\u9700\u6c42\u7684\u6a21\u578b\u3002\u8fd9\u4f7f\u5f97 pipeline<\/code> \u5de5\u5177\u66f4\u52a0\u7075\u6d3b\u548c\u5f3a\u5927\u3002<\/p>\n","protected":false},"excerpt":{"rendered":"
      \u4ec0\u4e48\u662f Hugging Face \u7684 pipeline\uff1f Hugging Face \u63d0\u4f9b\u7684 transformers \u5e93\u662f\u4e00\u4e2a\u975e\u5e38\u5f3a\u5927\u7684   \u2026 …<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"ngg_post_thumbnail":0},"categories":[225],"tags":[],"_links":{"self":[{"href":"\/index.php?rest_route=\/wp\/v2\/posts\/9368"}],"collection":[{"href":"\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=9368"}],"version-history":[{"count":1,"href":"\/index.php?rest_route=\/wp\/v2\/posts\/9368\/revisions"}],"predecessor-version":[{"id":9369,"href":"\/index.php?rest_route=\/wp\/v2\/posts\/9368\/revisions\/9369"}],"wp:attachment":[{"href":"\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=9368"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=9368"},{"taxonomy":"post_tag","embeddable":true,"href":"\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=9368"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}