| ---
|
| license: bsd-3-clause
|
| tags:
|
| - aggressive
|
| - hard-negative
|
| - light
|
| - moderate
|
| - nlp-summarization
|
| - semi-auto
|
| - temporal
|
| - text-tabular
|
| - tfrecord
|
| ---
|
|
|
| # build_dataset.py
|
|
|
| ## Dataset Summary
|
|
|
| A **nlp summarization** dataset with **text tabular** modality, stored in **tfrecord** format.
|
|
|
| ## Preprocessing & Augmentation
|
|
|
| - **Preprocessing**: aggressive
|
| - **Augmentation**: light
|
|
|
| ## Splits & Sampling
|
|
|
| - **Split strategy**: temporal
|
| - **Sampling**: hard negative
|
|
|
| ## Quality & Labeling
|
|
|
| - **Quality filtering**: moderate
|
| - **Labeling**: semi auto
|
|
|
| ## Files
|
|
|
| - `build_dataset.py` — main artifact of this repository
|
|
|
| ## License
|
|
|
| See the license field above.
|
|
|
