- Açıklama:
Eraser Multi RC, cevaplar ve bir mantıkla birlikte çok satırlı pasajlar üzerindeki sorgular için bir veri setidir. Bu veri kümesindeki her örnek aşağıdaki 5 bölüme sahiptir
- Çok Satırlı Bir Pasaj 2. Parça hakkında bir soru 3. Sorguya bir cevap
- Cevabın doğru mu yanlış mı olduğuna dair bir Sınıflandırma 5. Sınıflandırmayı haklı çıkaran bir açıklama
Anasayfa: https://cogcomp.seas.upenn.edu/multirc/
Kaynak kodu:
tfds.text.EraserMultiRc
sürümleri:
-
0.1.1
(varsayılan): Hayır sürüm notları.
-
İndirme boyutu:
1.59 MiB
Veri kümesi boyutu:
Unknown size
Otomatik önbelleğe ( dokümantasyon ): Bilinmeyen
Splits:
Bölmek | Örnekler |
---|---|
'test' | 4.848 |
'train' | 24.029 |
'validation' | 3.214 |
- Özellikler:
FeaturesDict({
'evidences': Sequence(Text(shape=(), dtype=tf.string)),
'label': ClassLabel(shape=(), dtype=tf.int64, num_classes=2),
'passage': Text(shape=(), dtype=tf.string),
'query_and_answer': Text(shape=(), dtype=tf.string),
})
Denetimli tuşları (Bkz
as_supervised
doc ):None
Şekil ( tfds.show_examples ): desteklenmez.
Örnekler ( tfds.as_dataframe ):
- Citation:
@unpublished{eraser2019,
title = {ERASER: A Benchmark to Evaluate Rationalized NLP Models},
author = {Jay DeYoung and Sarthak Jain and Nazneen Fatema Rajani and Eric Lehman and Caiming Xiong and Richard Socher and Byron C. Wallace}
}
@inproceedings{MultiRC2018,
author = {Daniel Khashabi and Snigdha Chaturvedi and Michael Roth and Shyam Upadhyay and Dan Roth},
title = {Looking Beyond the Surface:A Challenge Set for Reading Comprehension over Multiple Sentences},
booktitle = {NAACL},
year = {2018}
}