प्राकृतिक_प्रश्न

  • विवरण :

एनक्यू कॉर्पस में वास्तविक उपयोगकर्ताओं के प्रश्न शामिल हैं, और इसके लिए पूरे विकिपीडिया लेख को पढ़ने और समझने के लिए क्यूए सिस्टम की आवश्यकता होती है, जिसमें प्रश्न का उत्तर हो भी सकता है और नहीं भी। वास्तविक उपयोगकर्ता प्रश्नों का समावेश, और आवश्यकता है कि उत्तर खोजने के लिए समाधानों को एक पूरे पृष्ठ को पढ़ना चाहिए, एनक्यू को पिछले क्यूए डेटासेट की तुलना में अधिक यथार्थवादी और चुनौतीपूर्ण कार्य बनाता है।

विभाजित करना उदाहरण
'train' 307,373
'validation' 7,830
  • पर्यवेक्षित कुंजियाँ ( as_supervised doc देखें): None

  • चित्र ( tfds.show_examples ): समर्थित नहीं है।

  • उद्धरण :

@article{47761,
title = {Natural Questions: a Benchmark for Question Answering Research},
author = {Tom Kwiatkowski and Jennimaria Palomaki and Olivia Redfield and Michael Collins and Ankur Parikh and Chris Alberti and Danielle Epstein and Illia Polosukhin and Matthew Kelcey and Jacob Devlin and Kenton Lee and Kristina N. Toutanova and Llion Jones and Ming-Wei Chang and Andrew Dai and Jakob Uszkoreit and Quoc Le and Slav Petrov},
year = {2019},
journal = {Transactions of the Association of Computational Linguistics}
}

प्राकृतिक_प्रश्न/डिफ़ॉल्ट (डिफ़ॉल्ट कॉन्फ़िगरेशन)

  • विन्यास विवरण : डिफ़ॉल्ट प्राकृतिक_प्रश्न विन्यास

  • डेटासेट का आकार : 90.26 GiB

  • फ़ीचर संरचना :

FeaturesDict({
    'annotations': Sequence({
        'id': string,
        'long_answer': FeaturesDict({
            'end_byte': int64,
            'end_token': int64,
            'start_byte': int64,
            'start_token': int64,
        }),
        'short_answers': Sequence({
            'end_byte': int64,
            'end_token': int64,
            'start_byte': int64,
            'start_token': int64,
            'text': Text(shape=(), dtype=string),
        }),
        'yes_no_answer': ClassLabel(shape=(), dtype=int64, num_classes=2),
    }),
    'document': FeaturesDict({
        'html': Text(shape=(), dtype=string),
        'title': Text(shape=(), dtype=string),
        'tokens': Sequence({
            'is_html': bool,
            'token': Text(shape=(), dtype=string),
        }),
        'url': Text(shape=(), dtype=string),
    }),
    'id': string,
    'question': FeaturesDict({
        'text': Text(shape=(), dtype=string),
        'tokens': Sequence(string),
    }),
})
  • फ़ीचर दस्तावेज़ीकरण :
विशेषता कक्षा आकार डीटाइप विवरण
विशेषताएं डिक्ट
एनोटेशन अनुक्रम
एनोटेशन / आईडी टेन्सर डोरी
एनोटेशन/long_answer विशेषताएं डिक्ट
एनोटेशन/long_answer/end_byte टेन्सर int64
एनोटेशन/long_answer/end_token टेन्सर int64
एनोटेशन/long_answer/start_byte टेन्सर int64
एनोटेशन/long_answer/start_token टेन्सर int64
एनोटेशन/short_answers अनुक्रम
एनोटेशन/short_answers/end_byte टेन्सर int64
एनोटेशन/short_answers/end_token टेन्सर int64
एनोटेशन/short_answers/start_byte टेन्सर int64
एनोटेशन/short_answers/start_token टेन्सर int64
एनोटेशन/short_answers/text मूलपाठ डोरी
एनोटेशन/yes_no_answer क्लासलेबल int64
दस्तावेज़ विशेषताएं डिक्ट
दस्तावेज़/एचटीएमएल मूलपाठ डोरी
दस्तावेज़ का शीर्षक मूलपाठ डोरी
दस्तावेज़/टोकन अनुक्रम
दस्तावेज़/टोकन/is_html टेन्सर बूल
दस्तावेज़/टोकन/token मूलपाठ डोरी
दस्तावेज़/यूआरएल मूलपाठ डोरी
पहचान टेन्सर डोरी
सवाल विशेषताएं डिक्ट
प्रश्न / पाठ मूलपाठ डोरी
प्रश्न/टोकन अनुक्रम (टेंसर) (कोई नहीं,) डोरी

प्राकृतिक_प्रश्न/longt5

  • कॉन्फिग विवरण : longT5 बेंचमार्क के रूप में प्राकृतिक_प्रश्न पूर्वसंसाधित

  • डेटासेट का आकार : 8.91 GiB

  • फ़ीचर संरचना :

FeaturesDict({
    'all_answers': Sequence(Text(shape=(), dtype=string)),
    'answer': Text(shape=(), dtype=string),
    'context': Text(shape=(), dtype=string),
    'id': Text(shape=(), dtype=string),
    'question': Text(shape=(), dtype=string),
    'title': Text(shape=(), dtype=string),
})
  • फ़ीचर दस्तावेज़ीकरण :
विशेषता कक्षा आकार डीटाइप विवरण
विशेषताएं डिक्ट
सभी_उत्तर अनुक्रम (पाठ) (कोई नहीं,) डोरी
उत्तर मूलपाठ डोरी
प्रसंग मूलपाठ डोरी
पहचान मूलपाठ डोरी
सवाल मूलपाठ डोरी
शीर्षक मूलपाठ डोरी