metadata
license: apache-2.0
tags:
- balanced
- leave-one-out
- lenient
- multimodal3
- news-media
- none
- progressive
- tfrecord
- weak-supervision
dataset.py
Dataset Summary
A news media dataset with multimodal3 modality, stored in tfrecord format.
Preprocessing & Augmentation
- Preprocessing: progressive
- Augmentation: none
Splits & Sampling
- Split strategy: leave one out
- Sampling: balanced
Quality & Labeling
- Quality filtering: lenient
- Labeling: weak supervision
Files
dataset.py— main artifact of this repository
License
See the license field above.