गोंद

  • विवरण :

GLUE, सामान्य भाषा समझ मूल्यांकन बेंचमार्क ( https://gluebenchmark.com/ ) प्राकृतिक भाषा समझ प्रणालियों के प्रशिक्षण, मूल्यांकन और विश्लेषण के लिए संसाधनों का एक संग्रह है।

गोंद/कोला (डिफ़ॉल्ट कॉन्फ़िगरेशन)

  • कॉन्फ़िग विवरण : भाषाई स्वीकार्यता के संग्रह में भाषाई सिद्धांत पर पुस्तकों और जर्नल लेखों से लिए गए अंग्रेजी स्वीकार्यता निर्णय शामिल हैं। प्रत्येक उदाहरण शब्दों का एक क्रम है जो इस बात से जुड़ा है कि क्या यह एक व्याकरणिक अंग्रेजी वाक्य है।

  • होमपेज : https://nyu-mll.github.io/CoLA/

  • डाउनलोड आकार : 368.14 KiB

  • डेटासेट का आकार : 965.49 KiB

  • विभाजन :

विभाजित करना उदाहरण
'test' 1,063
'train' 8,551
'validation' 1,043
  • फ़ीचर संरचना :
FeaturesDict({
    'idx': int32,
    'label': ClassLabel(shape=(), dtype=int64, num_classes=2),
    'sentence': Text(shape=(), dtype=string),
})
  • फ़ीचर दस्तावेज़ीकरण :
विशेषता कक्षा आकार डीटाइप विवरण
विशेषताएं डिक्ट
idx टेन्सर int32
लेबल क्लासलेबल int64
वाक्य मूलपाठ डोरी
  • उद्धरण :
@article{warstadt2018neural,
  title={Neural Network Acceptability Judgments},
  author={Warstadt, Alex and Singh, Amanpreet and Bowman, Samuel R},
  journal={arXiv preprint arXiv:1805.12471},
  year={2018}
}
@inproceedings{wang2019glue,
  title={ {GLUE}: A Multi-Task Benchmark and Analysis Platform for Natural Language Understanding},
  author={Wang, Alex and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R.},
  note={In the Proceedings of ICLR.},
  year={2019}
}

Note that each GLUE dataset has its own citation. Please see the source to see
the correct citation for each contained dataset.

गोंद/sst2

  • कॉन्फिग विवरण : स्टैनफोर्ड सेंटीमेंट ट्रीबैंक में फिल्म समीक्षा के वाक्य और उनकी भावना के मानवीय एनोटेशन शामिल हैं। कार्य किसी दिए गए वाक्य की भावना की भविष्यवाणी करना है। हम दो-तरफा (सकारात्मक/नकारात्मक) वर्ग विभाजन का उपयोग करते हैं, और केवल वाक्य-स्तर के लेबल का उपयोग करते हैं।

  • होमपेज : https://nlp.stanford.edu/sentiment/index.html

  • डाउनलोड आकार : 7.09 MiB

  • डेटासेट का आकार : 7.22 MiB

  • विभाजन :

विभाजित करना उदाहरण
'test' 1,821
'train' 67,349
'validation' 872
  • फ़ीचर संरचना :
FeaturesDict({
    'idx': int32,
    'label': ClassLabel(shape=(), dtype=int64, num_classes=2),
    'sentence': Text(shape=(), dtype=string),
})
  • फ़ीचर दस्तावेज़ीकरण :
विशेषता कक्षा आकार डीटाइप विवरण
विशेषताएं डिक्ट
idx टेन्सर int32
लेबल क्लासलेबल int64
वाक्य मूलपाठ डोरी
  • उद्धरण :
@inproceedings{socher2013recursive,
  title={Recursive deep models for semantic compositionality over a sentiment treebank},
  author={Socher, Richard and Perelygin, Alex and Wu, Jean and Chuang, Jason and Manning, Christopher D and Ng, Andrew and Potts, Christopher},
  booktitle={Proceedings of the 2013 conference on empirical methods in natural language processing},
  pages={1631--1642},
  year={2013}
}
@inproceedings{wang2019glue,
  title={ {GLUE}: A Multi-Task Benchmark and Analysis Platform for Natural Language Understanding},
  author={Wang, Alex and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R.},
  note={In the Proceedings of ICLR.},
  year={2019}
}

Note that each GLUE dataset has its own citation. Please see the source to see
the correct citation for each contained dataset.

गोंद / एमआरपीसी

  • कॉन्फिग विवरण : माइक्रोसॉफ्ट रिसर्च पैराफ्रेज कॉर्पस (डोलन एंड ब्रॉकेट, 2005) ऑनलाइन समाचार स्रोतों से स्वचालित रूप से निकाले गए वाक्य जोड़े का एक कोष है, मानव एनोटेशन के साथ कि क्या जोड़ी में वाक्य शब्दार्थ के समकक्ष हैं।

  • होमपेज : https://www.microsoft.com/en-us/download/details.aspx?id=52398

  • डाउनलोड आकार : 1.43 MiB

  • डेटासेट का आकार : 1.74 MiB

  • विभाजन :

विभाजित करना उदाहरण
'test' 1,725
'train' 3,668
'validation' 408
  • फ़ीचर संरचना :
FeaturesDict({
    'idx': int32,
    'label': ClassLabel(shape=(), dtype=int64, num_classes=2),
    'sentence1': Text(shape=(), dtype=string),
    'sentence2': Text(shape=(), dtype=string),
})
  • फ़ीचर दस्तावेज़ीकरण :
विशेषता कक्षा आकार डीटाइप विवरण
विशेषताएं डिक्ट
idx टेन्सर int32
लेबल क्लासलेबल int64
वाक्य1 मूलपाठ डोरी
वाक्य2 मूलपाठ डोरी
  • उद्धरण :
@inproceedings{dolan2005automatically,
  title={Automatically constructing a corpus of sentential paraphrases},
  author={Dolan, William B and Brockett, Chris},
  booktitle={Proceedings of the Third International Workshop on Paraphrasing (IWP2005)},
  year={2005}
}
@inproceedings{wang2019glue,
  title={ {GLUE}: A Multi-Task Benchmark and Analysis Platform for Natural Language Understanding},
  author={Wang, Alex and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R.},
  note={In the Proceedings of ICLR.},
  year={2019}
}

Note that each GLUE dataset has its own citation. Please see the source to see
the correct citation for each contained dataset.

गोंद / क्यूक्यूपी

  • कॉन्फिग विवरण : Quora Question Pairs2 डेटासेट समुदाय प्रश्न-उत्तर देने वाली वेबसाइट Quora से प्रश्न जोड़ियों का एक संग्रह है। कार्य यह निर्धारित करना है कि प्रश्नों की एक जोड़ी शब्दार्थ के समकक्ष है या नहीं।

  • होमपेज : https://data.quora.com/First-Quora-Dataset-Release-Question-Pairs

  • डाउनलोड आकार : 39.76 MiB

  • डेटासेट का आकार : 150.37 MiB

  • विभाजन :

विभाजित करना उदाहरण
'test' 390,965
'train' 363,846
'validation' 40,430
  • फ़ीचर संरचना :
FeaturesDict({
    'idx': int32,
    'label': ClassLabel(shape=(), dtype=int64, num_classes=2),
    'question1': Text(shape=(), dtype=string),
    'question2': Text(shape=(), dtype=string),
})
  • फ़ीचर दस्तावेज़ीकरण :
विशेषता कक्षा आकार डीटाइप विवरण
विशेषताएं डिक्ट
idx टेन्सर int32
लेबल क्लासलेबल int64
प्रश्न 1 मूलपाठ डोरी
प्रश्न 2 मूलपाठ डोरी
  • उद्धरण :
@online{WinNT,
  author = {Iyer, Shankar and Dandekar, Nikhil and Csernai, Kornel},
  title = {First Quora Dataset Release: Question Pairs},
  year = 2017,
  url = {https://data.quora.com/First-Quora-Dataset-Release-Question-Pairs},
  urldate = {2019-04-03}
}
@inproceedings{wang2019glue,
  title={ {GLUE}: A Multi-Task Benchmark and Analysis Platform for Natural Language Understanding},
  author={Wang, Alex and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R.},
  note={In the Proceedings of ICLR.},
  year={2019}
}

Note that each GLUE dataset has its own citation. Please see the source to see
the correct citation for each contained dataset.

गोंद / एसटीएसबी

  • कॉन्फिग विवरण : सिमेंटिक टेक्स्टुअल सिमिलैरिटी बेंचमार्क (सेर एट अल।, 2017) समाचार सुर्खियों, वीडियो और छवि कैप्शन, और प्राकृतिक भाषा अनुमान डेटा से तैयार किए गए वाक्य जोड़े का एक संग्रह है। प्रत्येक जोड़ी को 0 से 5 तक समानता स्कोर के साथ मानव-एनोटेट किया गया है।

  • होमपेज : http://ixa2.si.ehu.es/stswiki/index.php/STSbenchmark

  • डाउनलोड आकार : 784.05 KiB

  • डेटासेट का आकार : 1.58 MiB

  • विभाजन :

विभाजित करना उदाहरण
'test' 1,379
'train' 5,749
'validation' 1,500
  • फ़ीचर संरचना :
FeaturesDict({
    'idx': int32,
    'label': float32,
    'sentence1': Text(shape=(), dtype=string),
    'sentence2': Text(shape=(), dtype=string),
})
  • फ़ीचर दस्तावेज़ीकरण :
विशेषता कक्षा आकार डीटाइप विवरण
विशेषताएं डिक्ट
idx टेन्सर int32
लेबल टेन्सर फ्लोट32
वाक्य1 मूलपाठ डोरी
वाक्य2 मूलपाठ डोरी
  • उद्धरण :
@article{cer2017semeval,
  title={Semeval-2017 task 1: Semantic textual similarity-multilingual and cross-lingual focused evaluation},
  author={Cer, Daniel and Diab, Mona and Agirre, Eneko and Lopez-Gazpio, Inigo and Specia, Lucia},
  journal={arXiv preprint arXiv:1708.00055},
  year={2017}
}
@inproceedings{wang2019glue,
  title={ {GLUE}: A Multi-Task Benchmark and Analysis Platform for Natural Language Understanding},
  author={Wang, Alex and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R.},
  note={In the Proceedings of ICLR.},
  year={2019}
}

Note that each GLUE dataset has its own citation. Please see the source to see
the correct citation for each contained dataset.

गोंद/mnli

  • कॉन्फिग विवरण : मल्टी-जेनर नेचुरल लैंग्वेज इन्वेंशन कॉर्पस टेक्स्टुअल एनलमेंट एनोटेशन के साथ वाक्य जोड़े का एक क्राउडसोर्स्ड संग्रह है। एक आधार वाक्य और एक परिकल्पना वाक्य को देखते हुए, यह भविष्यवाणी करना है कि क्या आधार परिकल्पना (एंटेलमेंट) पर जोर देता है, परिकल्पना (विरोधाभास) का खंडन करता है, या न ही (तटस्थ)। आधार वाक्यों को दस अलग-अलग स्रोतों से इकट्ठा किया जाता है, जिसमें अनुलेखित भाषण, कल्पना और सरकारी रिपोर्ट शामिल हैं। हम मानक परीक्षण सेट का उपयोग करते हैं, जिसके लिए हमने लेखकों से निजी लेबल प्राप्त किए, और मिलान किए गए (इन-डोमेन) और बेमेल (क्रॉस-डोमेन) दोनों अनुभागों पर मूल्यांकन किया। हम सहायक प्रशिक्षण डेटा के 550k उदाहरणों के रूप में SNLI कॉर्पस का उपयोग और अनुशंसा भी करते हैं।

  • होमपेज : http://www.nyu.edu/projects/bowman/multinli/

  • डाउनलोड आकार : 298.29 MiB

  • डेटासेट का आकार : 100.56 MiB

  • विभाजन :

विभाजित करना उदाहरण
'test_matched' 9,796
'test_mismatched' 9,847
'train' 392,702
'validation_matched' 9,815
'validation_mismatched' 9,832
  • फ़ीचर संरचना :
FeaturesDict({
    'hypothesis': Text(shape=(), dtype=string),
    'idx': int32,
    'label': ClassLabel(shape=(), dtype=int64, num_classes=3),
    'premise': Text(shape=(), dtype=string),
})
  • फ़ीचर दस्तावेज़ीकरण :
विशेषता कक्षा आकार डीटाइप विवरण
विशेषताएं डिक्ट
परिकल्पना मूलपाठ डोरी
idx टेन्सर int32
लेबल क्लासलेबल int64
आधार मूलपाठ डोरी
  • उद्धरण :
@InProceedings{N18-1101,
  author = "Williams, Adina
            and Nangia, Nikita
            and Bowman, Samuel",
  title = "A Broad-Coverage Challenge Corpus for
           Sentence Understanding through Inference",
  booktitle = "Proceedings of the 2018 Conference of
               the North American Chapter of the
               Association for Computational Linguistics:
               Human Language Technologies, Volume 1 (Long
               Papers)",
  year = "2018",
  publisher = "Association for Computational Linguistics",
  pages = "1112--1122",
  location = "New Orleans, Louisiana",
  url = "http://aclweb.org/anthology/N18-1101"
}
@article{bowman2015large,
  title={A large annotated corpus for learning natural language inference},
  author={Bowman, Samuel R and Angeli, Gabor and Potts, Christopher and Manning, Christopher D},
  journal={arXiv preprint arXiv:1508.05326},
  year={2015}
}
@inproceedings{wang2019glue,
  title={ {GLUE}: A Multi-Task Benchmark and Analysis Platform for Natural Language Understanding},
  author={Wang, Alex and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R.},
  note={In the Proceedings of ICLR.},
  year={2019}
}

Note that each GLUE dataset has its own citation. Please see the source to see
the correct citation for each contained dataset.

गोंद/mnli_mismatched

  • Config विवरण : बेमेल मान्यता और परीक्षण MNLI से विभाजित। अतिरिक्त जानकारी के लिए "mnli" BuilderConfig देखें।

  • होमपेज : http://www.nyu.edu/projects/bowman/multinli/

  • डाउनलोड आकार : 298.29 MiB

  • डेटासेट का आकार : 4.79 MiB

  • विभाजन :

विभाजित करना उदाहरण
'test' 9,847
'validation' 9,832
  • फ़ीचर संरचना :
FeaturesDict({
    'hypothesis': Text(shape=(), dtype=string),
    'idx': int32,
    'label': ClassLabel(shape=(), dtype=int64, num_classes=3),
    'premise': Text(shape=(), dtype=string),
})
  • फ़ीचर दस्तावेज़ीकरण :
विशेषता कक्षा आकार डीटाइप विवरण
विशेषताएं डिक्ट
परिकल्पना मूलपाठ डोरी
idx टेन्सर int32
लेबल क्लासलेबल int64
आधार मूलपाठ डोरी
  • उद्धरण :
@InProceedings{N18-1101,
  author = "Williams, Adina
            and Nangia, Nikita
            and Bowman, Samuel",
  title = "A Broad-Coverage Challenge Corpus for
           Sentence Understanding through Inference",
  booktitle = "Proceedings of the 2018 Conference of
               the North American Chapter of the
               Association for Computational Linguistics:
               Human Language Technologies, Volume 1 (Long
               Papers)",
  year = "2018",
  publisher = "Association for Computational Linguistics",
  pages = "1112--1122",
  location = "New Orleans, Louisiana",
  url = "http://aclweb.org/anthology/N18-1101"
}
@article{bowman2015large,
  title={A large annotated corpus for learning natural language inference},
  author={Bowman, Samuel R and Angeli, Gabor and Potts, Christopher and Manning, Christopher D},
  journal={arXiv preprint arXiv:1508.05326},
  year={2015}
}
@inproceedings{wang2019glue,
  title={ {GLUE}: A Multi-Task Benchmark and Analysis Platform for Natural Language Understanding},
  author={Wang, Alex and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R.},
  note={In the Proceedings of ICLR.},
  year={2019}
}

Note that each GLUE dataset has its own citation. Please see the source to see
the correct citation for each contained dataset.

गोंद/mnli_matched

  • Config विवरण : मिलान किया गया सत्यापन और परीक्षण MNLI से विभाजित होता है। अतिरिक्त जानकारी के लिए "mnli" BuilderConfig देखें।

  • होमपेज : http://www.nyu.edu/projects/bowman/multinli/

  • डाउनलोड आकार : 298.29 MiB

  • डेटासेट का आकार : 4.58 MiB

  • विभाजन :

विभाजित करना उदाहरण
'test' 9,796
'validation' 9,815
  • फ़ीचर संरचना :
FeaturesDict({
    'hypothesis': Text(shape=(), dtype=string),
    'idx': int32,
    'label': ClassLabel(shape=(), dtype=int64, num_classes=3),
    'premise': Text(shape=(), dtype=string),
})
  • फ़ीचर दस्तावेज़ीकरण :
विशेषता कक्षा आकार डीटाइप विवरण
विशेषताएं डिक्ट
परिकल्पना मूलपाठ डोरी
idx टेन्सर int32
लेबल क्लासलेबल int64
आधार मूलपाठ डोरी
  • उद्धरण :
@InProceedings{N18-1101,
  author = "Williams, Adina
            and Nangia, Nikita
            and Bowman, Samuel",
  title = "A Broad-Coverage Challenge Corpus for
           Sentence Understanding through Inference",
  booktitle = "Proceedings of the 2018 Conference of
               the North American Chapter of the
               Association for Computational Linguistics:
               Human Language Technologies, Volume 1 (Long
               Papers)",
  year = "2018",
  publisher = "Association for Computational Linguistics",
  pages = "1112--1122",
  location = "New Orleans, Louisiana",
  url = "http://aclweb.org/anthology/N18-1101"
}
@article{bowman2015large,
  title={A large annotated corpus for learning natural language inference},
  author={Bowman, Samuel R and Angeli, Gabor and Potts, Christopher and Manning, Christopher D},
  journal={arXiv preprint arXiv:1508.05326},
  year={2015}
}
@inproceedings{wang2019glue,
  title={ {GLUE}: A Multi-Task Benchmark and Analysis Platform for Natural Language Understanding},
  author={Wang, Alex and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R.},
  note={In the Proceedings of ICLR.},
  year={2019}
}

Note that each GLUE dataset has its own citation. Please see the source to see
the correct citation for each contained dataset.

गोंद/qnli

  • कॉन्फिग विवरण : स्टैनफोर्ड क्वेश्चन आंसरिंग डेटासेट एक प्रश्न-उत्तर डेटासेट है जिसमें प्रश्न-पैराग्राफ जोड़े होते हैं, जहां पैराग्राफ में एक वाक्य (विकिपीडिया से लिया गया) में संबंधित प्रश्न का उत्तर होता है (एक एनोटेटर द्वारा लिखित)। हम प्रत्येक प्रश्न और प्रत्येक वाक्य के बीच संबंधित संदर्भ में एक जोड़ी बनाकर और प्रश्न और संदर्भ वाक्य के बीच कम शाब्दिक ओवरलैप वाले जोड़े को फ़िल्टर करके कार्य को वाक्य जोड़ी वर्गीकरण में परिवर्तित करते हैं। कार्य यह निर्धारित करना है कि संदर्भ वाक्य में प्रश्न का उत्तर है या नहीं। मूल कार्य का यह संशोधित संस्करण इस आवश्यकता को हटा देता है कि मॉडल सटीक उत्तर का चयन करता है, लेकिन सरल धारणाओं को भी हटा देता है कि उत्तर हमेशा इनपुट में मौजूद होता है और लेक्सिकल ओवरलैप एक विश्वसनीय क्यू है।

  • होमपेज : https://rajpurkar.github.io/SQuAD-explorer/

  • डाउनलोड आकार : 10.14 MiB

  • डेटासेट का आकार : 32.99 MiB

  • विभाजन :

विभाजित करना उदाहरण
'test' 5,463
'train' 104,743
'validation' 5,463
  • फ़ीचर संरचना :
FeaturesDict({
    'idx': int32,
    'label': ClassLabel(shape=(), dtype=int64, num_classes=2),
    'question': Text(shape=(), dtype=string),
    'sentence': Text(shape=(), dtype=string),
})
  • फ़ीचर दस्तावेज़ीकरण :
विशेषता कक्षा आकार डीटाइप विवरण
विशेषताएं डिक्ट
idx टेन्सर int32
लेबल क्लासलेबल int64
प्रश्न मूलपाठ डोरी
वाक्य मूलपाठ डोरी
  • उद्धरण :
@article{rajpurkar2016squad,
  title={Squad: 100,000+ questions for machine comprehension of text},
  author={Rajpurkar, Pranav and Zhang, Jian and Lopyrev, Konstantin and Liang, Percy},
  journal={arXiv preprint arXiv:1606.05250},
  year={2016}
}
@inproceedings{wang2019glue,
  title={ {GLUE}: A Multi-Task Benchmark and Analysis Platform for Natural Language Understanding},
  author={Wang, Alex and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R.},
  note={In the Proceedings of ICLR.},
  year={2019}
}

Note that each GLUE dataset has its own citation. Please see the source to see
the correct citation for each contained dataset.

गोंद / आरटीई

  • कॉन्फ़िग विवरण : रिकॉग्नाइज़िंग टेक्सचुअल एनटेलमेंट (आरटीई) डेटासेट वार्षिक टेक्स्टुअल एंटेलमेंट चुनौतियों की एक श्रृंखला से आते हैं। हम RTE1 (Dagan et al., 2006), RTE2 (Bar Haim et al., 2006), RTE3 (Giampiccolo et al., 2007), और RTE5 (Bentivogli et al., 2009) से डेटा को जोड़ते हैं। 4 उदाहरण हैं। समाचार और विकिपीडिया पाठ के आधार पर निर्मित। हम सभी डेटासेट को दो-श्रेणी के विभाजन में परिवर्तित करते हैं, जहाँ तीन-श्रेणी के डेटासेट के लिए हम तटस्थता को समाप्त कर देते हैं और निरंतरता के लिए अंतर्विरोध नहीं करते हैं।

  • होमपेज : https://aclweb.org/aclwiki/Recognizing_Textual_Entailment

  • डाउनलोड आकार : 680.81 KiB

  • डेटासेट का आकार : 2.15 MiB

  • विभाजन :

विभाजित करना उदाहरण
'test' 3,000
'train' 2,490
'validation' 277
  • फ़ीचर संरचना :
FeaturesDict({
    'idx': int32,
    'label': ClassLabel(shape=(), dtype=int64, num_classes=2),
    'sentence1': Text(shape=(), dtype=string),
    'sentence2': Text(shape=(), dtype=string),
})
  • फ़ीचर दस्तावेज़ीकरण :
विशेषता कक्षा आकार डीटाइप विवरण
विशेषताएं डिक्ट
idx टेन्सर int32
लेबल क्लासलेबल int64
वाक्य1 मूलपाठ डोरी
वाक्य2 मूलपाठ डोरी
  • उद्धरण :
@inproceedings{dagan2005pascal,
  title={The PASCAL recognising textual entailment challenge},
  author={Dagan, Ido and Glickman, Oren and Magnini, Bernardo},
  booktitle={Machine Learning Challenges Workshop},
  pages={177--190},
  year={2005},
  organization={Springer}
}
@inproceedings{bar2006second,
  title={The second pascal recognising textual entailment challenge},
  author={Bar-Haim, Roy and Dagan, Ido and Dolan, Bill and Ferro, Lisa and Giampiccolo, Danilo and Magnini, Bernardo and Szpektor, Idan},
  booktitle={Proceedings of the second PASCAL challenges workshop on recognising textual entailment},
  volume={6},
  number={1},
  pages={6--4},
  year={2006},
  organization={Venice}
}
@inproceedings{giampiccolo2007third,
  title={The third pascal recognizing textual entailment challenge},
  author={Giampiccolo, Danilo and Magnini, Bernardo and Dagan, Ido and Dolan, Bill},
  booktitle={Proceedings of the ACL-PASCAL workshop on textual entailment and paraphrasing},
  pages={1--9},
  year={2007},
  organization={Association for Computational Linguistics}
}
@inproceedings{bentivogli2009fifth,
  title={The Fifth PASCAL Recognizing Textual Entailment Challenge.},
  author={Bentivogli, Luisa and Clark, Peter and Dagan, Ido and Giampiccolo, Danilo},
  booktitle={TAC},
  year={2009}
}
@inproceedings{wang2019glue,
  title={ {GLUE}: A Multi-Task Benchmark and Analysis Platform for Natural Language Understanding},
  author={Wang, Alex and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R.},
  note={In the Proceedings of ICLR.},
  year={2019}
}

Note that each GLUE dataset has its own citation. Please see the source to see
the correct citation for each contained dataset.

गोंद/wnli

  • विन्यास विवरण : विनोग्रैड स्कीमा चैलेंज (लेवेस्क और अन्य, 2011) एक पठन बोध कार्य है जिसमें एक प्रणाली को एक सर्वनाम के साथ एक वाक्य पढ़ना चाहिए और विकल्पों की सूची से उस सर्वनाम के दिग्दर्शन का चयन करना चाहिए। उदाहरण मैन्युअल रूप से सरल सांख्यिकीय विधियों को विफल करने के लिए बनाए गए हैं: प्रत्येक वाक्य में एक शब्द या वाक्यांश द्वारा प्रदान की गई प्रासंगिक जानकारी पर आकस्मिक है। समस्या को वाक्य युग्म वर्गीकरण में बदलने के लिए, हम अस्पष्ट सर्वनाम को प्रत्येक संभावित संदर्भ के साथ बदलकर वाक्य युग्म बनाते हैं। कार्य भविष्यवाणी करना है कि सर्वनाम के साथ प्रतिस्थापित वाक्य मूल वाक्य से जुड़ा हुआ है या नहीं। हम एक छोटे से मूल्यांकन सेट का उपयोग करते हैं जिसमें फिक्शन किताबों से प्राप्त नए उदाहरण शामिल हैं जिन्हें मूल संग्रह के लेखकों द्वारा निजी तौर पर साझा किया गया था। जबकि शामिल प्रशिक्षण सेट दो वर्गों के बीच संतुलित है, परीक्षण सेट उनके बीच असंतुलित है (65% प्रवेश नहीं)। इसके अलावा, एक डेटा विचित्रता के कारण, विकास सेट प्रतिकूल है: कभी-कभी परिकल्पनाओं को प्रशिक्षण और विकास उदाहरणों के बीच साझा किया जाता है, इसलिए यदि कोई मॉडल प्रशिक्षण उदाहरणों को याद करता है, तो वे संबंधित विकास सेट उदाहरण पर गलत लेबल की भविष्यवाणी करेंगे। क्यूएनएलआई के साथ, प्रत्येक उदाहरण का अलग-अलग मूल्यांकन किया जाता है, इसलिए इस कार्य पर मॉडल के स्कोर और अपरिवर्तनीय मूल कार्य पर इसके स्कोर के बीच व्यवस्थित पत्राचार नहीं होता है। हम परिवर्तित डेटासेट WNLI (विनोग्रैड एनएलआई) कहते हैं।

  • होमपेज : https://cs.nyu.edu/facademy/davise/papers/WinogradSchemas/WS.html

  • डाउनलोड आकार : 28.32 KiB

  • डेटासेट का आकार : 198.88 KiB

  • विभाजन :

विभाजित करना उदाहरण
'test' 146
'train' 635
'validation' 71
  • फ़ीचर संरचना :
FeaturesDict({
    'idx': int32,
    'label': ClassLabel(shape=(), dtype=int64, num_classes=2),
    'sentence1': Text(shape=(), dtype=string),
    'sentence2': Text(shape=(), dtype=string),
})
  • फ़ीचर दस्तावेज़ीकरण :
विशेषता कक्षा आकार डीटाइप विवरण
विशेषताएं डिक्ट
idx टेन्सर int32
लेबल क्लासलेबल int64
वाक्य1 मूलपाठ डोरी
वाक्य2 मूलपाठ डोरी
  • उद्धरण :
@inproceedings{levesque2012winograd,
  title={The winograd schema challenge},
  author={Levesque, Hector and Davis, Ernest and Morgenstern, Leora},
  booktitle={Thirteenth International Conference on the Principles of Knowledge Representation and Reasoning},
  year={2012}
}
@inproceedings{wang2019glue,
  title={ {GLUE}: A Multi-Task Benchmark and Analysis Platform for Natural Language Understanding},
  author={Wang, Alex and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R.},
  note={In the Proceedings of ICLR.},
  year={2019}
}

Note that each GLUE dataset has its own citation. Please see the source to see
the correct citation for each contained dataset.

गोंद / कुल्हाड़ी

  • कॉन्फिग विवरण : भाषाई घटनाओं की एक विस्तृत श्रृंखला पर सिस्टम प्रदर्शन के ठीक-ठीक विश्लेषण के लिए मैन्युअल रूप से क्यूरेटेड मूल्यांकन डेटासेट। यह डेटासेट प्राकृतिक भाषा अनुमान (एनएलआई) समस्याओं के माध्यम से वाक्य समझ का मूल्यांकन करता है। इस डेटासेट के लिए भविष्यवाणियां करने के लिए MulitNLI पर प्रशिक्षित मॉडल का उपयोग करें।

  • होमपेज : https://gluebenchmark.com/diagnostics

  • डाउनलोड आकार : 217.05 KiB

  • डेटासेट का आकार : 299.16 KiB

  • विभाजन :

विभाजित करना उदाहरण
'test' 1,104
  • फ़ीचर संरचना :
FeaturesDict({
    'hypothesis': Text(shape=(), dtype=string),
    'idx': int32,
    'label': ClassLabel(shape=(), dtype=int64, num_classes=3),
    'premise': Text(shape=(), dtype=string),
})
  • फ़ीचर दस्तावेज़ीकरण :
विशेषता कक्षा आकार डीटाइप विवरण
विशेषताएं डिक्ट
परिकल्पना मूलपाठ डोरी
idx टेन्सर int32
लेबल क्लासलेबल int64
आधार मूलपाठ डोरी
  • उद्धरण :
@inproceedings{wang2019glue,
  title={ {GLUE}: A Multi-Task Benchmark and Analysis Platform for Natural Language Understanding},
  author={Wang, Alex and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R.},
  note={In the Proceedings of ICLR.},
  year={2019}
}

Note that each GLUE dataset has its own citation. Please see the source to see
the correct citation for each contained dataset.