Update README.md
Browse files
README.md
CHANGED
|
@@ -0,0 +1,71 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: cc-by-nc-sa-4.0
|
| 3 |
+
task_categories:
|
| 4 |
+
- text-generation
|
| 5 |
+
- fill-mask
|
| 6 |
+
language:
|
| 7 |
+
- de
|
| 8 |
+
tags:
|
| 9 |
+
- wikipedia
|
| 10 |
+
- german
|
| 11 |
+
- cleaned
|
| 12 |
+
- nlp
|
| 13 |
+
size_categories:
|
| 14 |
+
- 10k<n<100k
|
| 15 |
+
---
|
| 16 |
+
|
| 17 |
+
<div align="center">
|
| 18 |
+
|
| 19 |
+
# 💎 APEX DATASETS
|
| 20 |
+
### "Believe in Apex. Become what's Next."
|
| 21 |
+
|
| 22 |
+
<br>
|
| 23 |
+
|
| 24 |
+
# 🇩🇪 German Wikipedia Cleaned: Platinum Sample
|
| 25 |
+
|
| 26 |
+
[](https://creativecommons.org/licenses/by-nc-sa/4.0/)
|
| 27 |
+
[-yellow.svg)]()
|
| 28 |
+
[]()
|
| 29 |
+
[]()
|
| 30 |
+
|
| 31 |
+
**Cleaned • Deduplicated • Optimized for LLMs**
|
| 32 |
+
|
| 33 |
+
</div>
|
| 34 |
+
|
| 35 |
+
---
|
| 36 |
+
|
| 37 |
+
## 💎 The "Art" of Clean Data
|
| 38 |
+
This is not just a dataset; it's a **masterpiece of data engineering**. We have processed raw German Wikipedia dumps, removing all HTML tags, boilerplate, and noise, leaving only the **purest linguistic structures** for your AI training needs.
|
| 39 |
+
|
| 40 |
+
> *"Quality over Quantity. But we give you both."* — **Apex Datasets**
|
| 41 |
+
|
| 42 |
+
---
|
| 43 |
+
|
| 44 |
+
## 🏗️ Dataset Structure (Sample)
|
| 45 |
+
|
| 46 |
+
We provide a **ready-to-use** split, formatted in Parquet for maximum efficiency.
|
| 47 |
+
|
| 48 |
+
| File Name | Rows | Usage Scenario | Description |
|
| 49 |
+
| :--- | :---: | :--- | :--- |
|
| 50 |
+
| **`sample_train.parquet`** | **400** | 🏋️ Fine-Tuning | Rich, diverse sentences for teaching the model context. |
|
| 51 |
+
| **`sample_test.parquet`** | **100** | 🧪 Evaluation | Unseen data for testing model performance (perplexity). |
|
| 52 |
+
| **`dataset_statistics.json`** | - | 📊 Analysis | Complete metadata and statistical breakdown. |
|
| 53 |
+
|
| 54 |
+
---
|
| 55 |
+
|
| 56 |
+
## 📊 Quick Stats (Full Version)
|
| 57 |
+
* **Total Size:** 15GB+
|
| 58 |
+
* **Sentences:** 3,000,000+
|
| 59 |
+
* **Format:** Apache Parquet (Snappy compression)
|
| 60 |
+
* **Encoding:** UTF-8
|
| 61 |
+
|
| 62 |
+
---
|
| 63 |
+
|
| 64 |
+
## 🚀 How to Use
|
| 65 |
+
```python
|
| 66 |
+
from datasets import load_dataset
|
| 67 |
+
|
| 68 |
+
# Download the sample directly
|
| 69 |
+
dataset = load_dataset("Apex-Datasets/German-Wikipedia-Cleaned-Sample")
|
| 70 |
+
|
| 71 |
+
print(dataset['train'][0])
|