Apex-Datasets commited on
Commit
da034a0
·
verified ·
1 Parent(s): ad7906a

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +0 -81
README.md CHANGED
@@ -1,81 +0,0 @@
1
- ---
2
- license: cc-by-nc-sa-4.0
3
- language:
4
- - de
5
- tags:
6
- - art-of-data
7
- - german-wikipedia
8
- - nlp-masterpiece
9
- pretty_name: "German Platinum Collection"
10
- ---
11
-
12
- <div align="center">
13
-
14
- <img src="https://via.placeholder.com/1500x400/1a1a1a/FFFFFF?text=GERMAN+PLATINUM+COLLECTION+%7C+APEX+DATASETS" alt="Banner" width="100%" style="border-radius: 10px;" />
15
-
16
- <br/><br/>
17
-
18
- [![License](https://img.shields.io/badge/LICENSE-CC_BY_NC_SA_4.0-000000?style=for-the-badge)]()
19
- [![Language](https://img.shields.io/badge/LANGUAGE-GERMAN_(DE)-000000?style=for-the-badge&color=FFCC00)]()
20
- [![Quality](https://img.shields.io/badge/QUALITY-PLATINUM_GRADE-000000?style=for-the-badge&color=E8E8E8)]()
21
-
22
- <br/>
23
-
24
- <h1>🇩🇪 THE GERMAN PROJECT</h1>
25
- <p style="font-size: 1.2em; font-style: italic; color: #666;">
26
- "Data is not just numbers. It is the architecture of intelligence."
27
- </p>
28
-
29
- </div>
30
-
31
- ---
32
-
33
- ## 🏛️ THE CONCEPT (Manifesto)
34
-
35
- This is not a mere collection of text. It is a **curated linguistic archive**.
36
- We have deconstructed the chaotic German Wikipedia, stripped away the digital noise (HTML, metadata), and reconstructed it into a pure, crystalline form ready for high-end Large Language Model training.
37
-
38
- > **Essentially, this dataset isn't about presenting raw data, but creating a space that allows AI to learn from the purest form of human language.**
39
-
40
- ---
41
-
42
- ## 🏗️ THE ARTIFACTS (File Structure)
43
-
44
- We present the data in **Apache Parquet**, the industry standard for performance and elegance.
45
-
46
- | **Component** | **Volume** | **Function** | **Design Intent** |
47
- | :--- | :---: | :--- | :--- |
48
- | **`german_sample_train.parquet`** | `400 rows` | **Training** | The foundation. Diverse sentence structures for deep learning. |
49
- | **`german_sample_test.parquet`** | `100 rows` | **Testing** | The critique. Unseen data to challenge and evaluate the model's perception. |
50
- | **`dataset_statistics.json`** | `Meta` | **Analysis** | The blueprint. Complete statistical breakdown of the dataset. |
51
-
52
- ---
53
-
54
- ## ⚡ THE FULL EXPERIENCE (15GB)
55
-
56
- This repository hosts the **Public Sample (1%)**.
57
- To access the complete **15GB Master Archive**, containing over **3,000,000 sentences**, you must acquire the access key.
58
-
59
- <div align="center">
60
- <br/>
61
-
62
- [![Unlock Full Version](https://img.shields.io/badge/🔓_ACCESS_FULL_ARCHIVE-LEMON_SQUEEZY-000000?style=for-the-badge&logo=shazam&color=ff0050)](https://lemonsqueezy.com/...)
63
-
64
- <p style="font-size: 0.9em; color: #888;">Only for those who demand the absolute best.</p>
65
-
66
- </div>
67
-
68
- ---
69
-
70
- ## 🛠️ USAGE (Python)
71
-
72
- Minimalist integration for your workflow:
73
-
74
- ```python
75
- from datasets import load_dataset
76
-
77
- # Initiate the experience
78
- ds = load_dataset("Apex-Datasets/German-Wikipedia-Cleaned-Sample")
79
-
80
- # Inspect the architecture
81
- print(ds['train'][0])