Intel-Web
Collection
This is a complete collection of full scrapped Intel's Website • 3 items • Updated • 2
url string | title string | html string | markdown string | links list | assets string |
|---|---|---|---|---|---|
https://ark.intel.com/content/www/us/en/ark.html | Intel® Product Specifications | "<html class=\"rwd geo-override js vis no-rtl headerfooter-menu3 no-applicationcache audio audio-og(...TRUNCATED) | "Intel® Product Specifications\n\n[Skip To Main Content](#primary-content) \n\n[ \n\n[ \n\n[ \n\n[ \n\n[ \n\n[ \n\n[ \n\n[ \n\n[ \n\n[![In(...TRUNCATED) | ["https://ark.intel.com/content/www/us/en/homepage.html","https://ark.intel.com/content/www/us/en/pr(...TRUNCATED) | "[{\"url\": \"https://cdn.cookielaw.org/logos/215b83c1-e050-4276-906c-e65481a36156/4e0ec566-94a6-4ca(...TRUNCATED) |
End of preview. Expand in Data Studio
Intel Ark Specifications Dataset
This dataset contains the complete HTML structure, markdown representations, and specifications for processors and hardware from the Intel Ark database (ark.intel.com). It is designed for large-scale data extraction, web-agent grounding, and LLM pre-training.
📊 Dataset Statistics
We utilized tiktoken (cl100k_base encoding) to calculate the precise token volume of this dataset for LLM training applications.
| Metric | Count / Size |
|---|---|
| Total Pages Scraped | 3,791 |
| Total HTML Tokens | 890,315,124 tokens |
| Total Markdown Tokens | 52,182,332 tokens |
| Parquet Partitions | 38 (100 pages per file) |
| Compression | zstd (Level 7) |
📁 Repository Structure
pages_part_*.parquet: Partitioned dataset files containing the raw HTML and Markdown. Partitioning guarantees low RAM overhead when streaming this large corpus.images.tar.zst: An incredibly dense, deduplicated archive of all visual assets (SVG logos, WebP product badges, flags) mapped perfectly to their source pages.
⚙️ Architecture & Data Sourcing
The raw HTML is stored exactly as it was delivered via Server-Side Rendering (SSR) from Intel's Adobe Experience Manager (AEM) infrastructure. Because the processor specifications are strictly baked into the server-rendered DOM, absolutely no CSS or client-side JavaScript execution is required to extract tabular data.
Note: Frontend Stylesheets and Scripts have been specifically decoupled into the sphita/intel-stylesheet-javascript dataset for targeted web-agent UI training.
📄 License
This dataset is published under the GPL-3.0 License.
- Downloads last month
- -