Update README.md
Browse files
README.md
CHANGED
|
@@ -18,6 +18,8 @@ size_categories:
|
|
| 18 |
Large-scale LaTeX corpus from arXiv (math, CS, physics, statistics)
|
| 19 |
</p>
|
| 20 |
|
|
|
|
|
|
|
| 21 |
## 📚 Overview
|
| 22 |
|
| 23 |
**arxiv-tex-corpus-full (80GB)** is a large-scale dataset of LaTeX source content extracted from papers hosted on arXiv.
|
|
@@ -42,8 +44,6 @@ The dataset is designed for research in:
|
|
| 42 |
|
| 43 |
This repository provides the **full 80GB version** of the corpus.
|
| 44 |
|
| 45 |
-
---
|
| 46 |
-
|
| 47 |
## 📦 Files
|
| 48 |
|
| 49 |
```
|
|
@@ -65,8 +65,6 @@ Validation split in JSON Lines format.
|
|
| 65 |
Internal processing checkpoint file used during dataset construction.
|
| 66 |
This file is not required for training.
|
| 67 |
|
| 68 |
-
---
|
| 69 |
-
|
| 70 |
## 🧾 Data Format
|
| 71 |
|
| 72 |
Each line in `train.jsonl` and `val.jsonl` is a JSON object.
|
|
@@ -95,8 +93,6 @@ Example schema:
|
|
| 95 |
* `stat.TH`
|
| 96 |
* `latex` — extracted LaTeX source content
|
| 97 |
|
| 98 |
-
---
|
| 99 |
-
|
| 100 |
## 📊 Dataset Statistics
|
| 101 |
|
| 102 |
* Total size: ~80GB
|
|
@@ -105,8 +101,6 @@ Example schema:
|
|
| 105 |
* Snapshot date: *(add date here)*
|
| 106 |
* Deduplication: *(state yes/no)*
|
| 107 |
|
| 108 |
-
---
|
| 109 |
-
|
| 110 |
## 🧹 Processing Pipeline
|
| 111 |
|
| 112 |
The dataset was constructed by:
|
|
@@ -123,8 +117,6 @@ Unless otherwise specified:
|
|
| 123 |
* No normalization of LaTeX commands was performed.
|
| 124 |
* No compilation validation was enforced.
|
| 125 |
|
| 126 |
-
---
|
| 127 |
-
|
| 128 |
## ⚖️ Licensing
|
| 129 |
|
| 130 |
This dataset contains LaTeX source files from papers hosted on arXiv.
|
|
@@ -138,8 +130,6 @@ For license details, see:
|
|
| 138 |
|
| 139 |
If you filtered by specific licenses (e.g., CC-BY only), please state that here explicitly.
|
| 140 |
|
| 141 |
-
---
|
| 142 |
-
|
| 143 |
## 🎯 Intended Use
|
| 144 |
|
| 145 |
This dataset is intended for:
|
|
@@ -149,8 +139,6 @@ This dataset is intended for:
|
|
| 149 |
* Modeling structured scientific documents
|
| 150 |
* Studying LaTeX generation and transformation
|
| 151 |
|
| 152 |
-
---
|
| 153 |
-
|
| 154 |
## 🚫 Limitations
|
| 155 |
|
| 156 |
* LaTeX sources may contain compilation errors.
|
|
@@ -159,8 +147,6 @@ This dataset is intended for:
|
|
| 159 |
* Licensing varies across papers.
|
| 160 |
* Category labeling follows arXiv metadata and may not reflect full topical scope.
|
| 161 |
|
| 162 |
-
---
|
| 163 |
-
|
| 164 |
## 📌 Citation
|
| 165 |
|
| 166 |
If you use this dataset, please cite:
|
|
@@ -178,8 +164,6 @@ Example:
|
|
| 178 |
}
|
| 179 |
```
|
| 180 |
|
| 181 |
-
---
|
| 182 |
-
|
| 183 |
## 🤝 Acknowledgements
|
| 184 |
|
| 185 |
This dataset is derived from papers made publicly available by authors via arXiv. We thank the research community for openly sharing their work.
|
|
|
|
| 18 |
Large-scale LaTeX corpus from arXiv (math, CS, physics, statistics)
|
| 19 |
</p>
|
| 20 |
|
| 21 |
+
📄 Paper: https://arxiv.org/abs/2602.17288
|
| 22 |
+
|
| 23 |
## 📚 Overview
|
| 24 |
|
| 25 |
**arxiv-tex-corpus-full (80GB)** is a large-scale dataset of LaTeX source content extracted from papers hosted on arXiv.
|
|
|
|
| 44 |
|
| 45 |
This repository provides the **full 80GB version** of the corpus.
|
| 46 |
|
|
|
|
|
|
|
| 47 |
## 📦 Files
|
| 48 |
|
| 49 |
```
|
|
|
|
| 65 |
Internal processing checkpoint file used during dataset construction.
|
| 66 |
This file is not required for training.
|
| 67 |
|
|
|
|
|
|
|
| 68 |
## 🧾 Data Format
|
| 69 |
|
| 70 |
Each line in `train.jsonl` and `val.jsonl` is a JSON object.
|
|
|
|
| 93 |
* `stat.TH`
|
| 94 |
* `latex` — extracted LaTeX source content
|
| 95 |
|
|
|
|
|
|
|
| 96 |
## 📊 Dataset Statistics
|
| 97 |
|
| 98 |
* Total size: ~80GB
|
|
|
|
| 101 |
* Snapshot date: *(add date here)*
|
| 102 |
* Deduplication: *(state yes/no)*
|
| 103 |
|
|
|
|
|
|
|
| 104 |
## 🧹 Processing Pipeline
|
| 105 |
|
| 106 |
The dataset was constructed by:
|
|
|
|
| 117 |
* No normalization of LaTeX commands was performed.
|
| 118 |
* No compilation validation was enforced.
|
| 119 |
|
|
|
|
|
|
|
| 120 |
## ⚖️ Licensing
|
| 121 |
|
| 122 |
This dataset contains LaTeX source files from papers hosted on arXiv.
|
|
|
|
| 130 |
|
| 131 |
If you filtered by specific licenses (e.g., CC-BY only), please state that here explicitly.
|
| 132 |
|
|
|
|
|
|
|
| 133 |
## 🎯 Intended Use
|
| 134 |
|
| 135 |
This dataset is intended for:
|
|
|
|
| 139 |
* Modeling structured scientific documents
|
| 140 |
* Studying LaTeX generation and transformation
|
| 141 |
|
|
|
|
|
|
|
| 142 |
## 🚫 Limitations
|
| 143 |
|
| 144 |
* LaTeX sources may contain compilation errors.
|
|
|
|
| 147 |
* Licensing varies across papers.
|
| 148 |
* Category labeling follows arXiv metadata and may not reflect full topical scope.
|
| 149 |
|
|
|
|
|
|
|
| 150 |
## 📌 Citation
|
| 151 |
|
| 152 |
If you use this dataset, please cite:
|
|
|
|
| 164 |
}
|
| 165 |
```
|
| 166 |
|
|
|
|
|
|
|
| 167 |
## 🤝 Acknowledgements
|
| 168 |
|
| 169 |
This dataset is derived from papers made publicly available by authors via arXiv. We thank the research community for openly sharing their work.
|