metadata
dataset_info:
features:
- name: document
dtype: string
- name: summary
dtype: string
- name: source
dtype: string
- name: __index_level_0__
dtype: int64
splits:
- name: train
num_bytes: 1143746144
num_examples: 83254
- name: validation
num_bytes: 142815263
num_examples: 10405
- name: test
num_bytes: 143020108
num_examples: 10405
download_size: 637677002
dataset_size: 1429581515
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*
task_categories:
- summarization
language:
- en
tags:
- finance
size_categories:
- 100K<n<1M
Dataset Card for Dataset Name
This dataset is designed for text summarization tasks, specifically focusing on financial and liquidity data. It combines structured text from different segments of financial reports, allowing for both automatic and human evaluation in text summarization tasks.
Dataset Details
This dataset was built using the dataset presented in the research paper "Long Text and Multi-Table Summarization: Dataset and Method". The dataset consists of financial documents with detailed reports and their corresponding summaries, which aim to condense lengthy documents into shorter, coherent summaries.
Paper Reference: Long Text and Multi-Table Summarization: Dataset and Method
Dataset Description
Dataset Structure
The dataset is divided into:
- Train: The primary dataset for model training.
- Validation: Used for validation during training.
- Test: Used for final evaluation of the summarization models.
Each entry consists of:
- text: The full input document, which is around 2500 words in length.
- summary: A condensed version of the document, around 350 words long.