- Description:
The movie rationale dataset contains human annotated rationales for movie reviews.
Source code:
tfds.text.MovieRationales
Versions:
0.1.0
(default): No release notes.
Download size:
3.72 MiB
Dataset size:
8.37 MiB
Auto-cached (documentation): Yes
Splits:
Split | Examples |
---|---|
'test' |
199 |
'train' |
1,600 |
'validation' |
200 |
- Feature structure:
FeaturesDict({
'evidences': Sequence(Text(shape=(), dtype=string)),
'label': ClassLabel(shape=(), dtype=int64, num_classes=2),
'review': Text(shape=(), dtype=string),
})
- Feature documentation:
Feature | Class | Shape | Dtype | Description |
---|---|---|---|---|
FeaturesDict | ||||
evidences | Sequence(Text) | (None,) | string | |
label | ClassLabel | int64 | ||
review | Text | string |
Supervised keys (See
as_supervised
doc):None
Figure (tfds.show_examples): Not supported.
Examples (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@unpublished{eraser2019,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{zaidan-eisner-piatko-2008:nips,
author = {Omar F. Zaidan and Jason Eisner and Christine Piatko},
title = {Machine Learning with Annotator Rationales to Reduce Annotation Cost},
booktitle = {Proceedings of the NIPS*2008 Workshop on Cost Sensitive Learning},
month = {December},
year = {2008}
}