Datasets:
Tasks:
Text Generation
Modalities:
Text
Sub-tasks:
language-modeling
Languages:
English
Size:
100K - 1M
License:
Commit ·
52d97b5
1
Parent(s): ed95795
Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,17 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
Note: this repo is a WIP and does not yet implement all features described below. It is certainly not ready to be used to train a model.
|
| 2 |
+
# Dataset card for the proof-pile
|
| 3 |
+
The `proof-pile` is a 45GB pre-training dataset of mathematical text. The dataset is composed of diverse sources of both informal and formal mathematics, namely
|
| 4 |
+
- ArXiv.math (40GB)
|
| 5 |
+
- Open-source math textbooks (50MB)
|
| 6 |
+
- Formal mathematics libraries (500MB)
|
| 7 |
+
- Lean mathlib and other Lean repositories
|
| 8 |
+
- Isabelle AFP
|
| 9 |
+
- Coq mathematical components and other Coq repositories
|
| 10 |
+
- HOL Light
|
| 11 |
+
- set.mm
|
| 12 |
+
- Mizar Mathematical Library
|
| 13 |
+
- Math Overflow and Math Stack Exchange (500MB)
|
| 14 |
+
- Wiki-style sources (50MB)
|
| 15 |
+
- ProofWiki
|
| 16 |
+
- Wikipedia math articles
|
| 17 |
+
- MATH dataset (6MB)
|