metadata
pretty_name: MegaMath synthetic math
license: other
license_name: megamath
language:
- en
task_categories:
- text-generation
tags:
- finance
- pretraining
- corpus
- thefinai
extra_gated_heading: Request access to MegaMath synthetic math
extra_gated_description: >-
MegaMath synthetic math is released by The Fin AI for research. Access is
granted automatically after you complete this short form.
extra_gated_button_content: Agree and access
extra_gated_prompt: >-
By accessing this dataset you agree to its license and to cite the The Fin AI
paper in any resulting publication. Use of the content is also subject to the
terms of the original source.
extra_gated_fields:
Full name: text
Affiliation: text
Country: country
Intended use:
type: select
options:
- Research
- Education
- Commercial evaluation
- Other
I agree to the terms above and will cite the paper: checkbox
MegaMath synthetic math
Pretraining / reference corpus released by The Fin AI. Source: MegaMath (synthetic subset) — https://huggingface.co/datasets/LLM360/MegaMath.
Source
Synthetic math text from MegaMath.
Structure
- Rows: 3,271,606
- Columns:
Source,Date,Text,Token_count
Quick Start
from datasets import load_dataset
ds = load_dataset("TheFinAI/all_math", split="train", streaming=True)
print(next(iter(ds)))
License
The content keeps the license and terms of its original source (see above). The Fin AI does not relicense it; please check the source's terms before reuse.