- תיאור:
Eraser Multi RC הוא מערך נתונים לשאילתות על מעברים מרובי שורות, יחד עם תשובות ורציונל. לכל דוגמה במערך הנתונים הזה יש את 5 החלקים הבאים
- קטע Mutli-line 2. שאילתה אודות המעבר 3. תשובה לשאילתה
- סיווג האם התשובה נכונה או לא נכונה 5. הסבר המצדיק את הסיווג
קוד מקור:
tfds.text.EraserMultiRc
גרסאות:
-
0.1.1
(ברירת המחדל): הערות השחרור לא.
-
גודל ההורדה:
1.59 MiB
גודל בסיס הנתונים:
Unknown size
Auto-במטמון ( תיעוד ): לא ידוע
פיצולים:
לְפַצֵל | דוגמאות |
---|---|
'test' | 4,848 |
'train' | 24,029 |
'validation' | 3,214 |
- מאפיינים:
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),
})
מפתחות השגחה (ראה
as_supervised
doc ):None
איור ( tfds.show_examples ): לא נתמך.
דוגמאות ( tfds.as_dataframe ):
- ציטוט:
@unpublished{eraser2019,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}
}