gpu-document-processing

Extract text, tables, and embeddings from large PDFs via GPU pipelines.

1|Updated Apr 8, 2026
One-click install
npx skills add https://github.com/leesk212/dannys-coding-ai-agent-final --skill gpu-document-processing-leesk212
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: gpu-document-processing
Source: https://github.com/leesk212/dannys-coding-ai-agent-final/tree/main/ETC/deepagents_sourcecode/examples/nvidia_deep_agent/skills/gpu-document-processing
Command: npx skills add https://github.com/leesk212/dannys-coding-ai-agent-final --skill gpu-document-processing-leesk212

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Large documents and document collections often require time-consuming, manual processing to extract text, tables, and structured data. This skill uses GPU-accelerated tools to speed up parsing, extraction, and embedding generation while preserving layout and context.

Core Features & Use Cases

  • PDF Text Extraction: Preserve layout (headers, paragraphs, lists, tables) and capture page references.
  • Tabular Data Extraction: Convert PDF tables into structured formats (CSV/DataFrames) with type detection and multi-row headers.
  • Document Chunking & Embedding: Split large documents into meaningful chunks and generate embeddings for search and analysis.
  • Bulk Processing: Process large collections in parallel, extract metadata, and synthesize per-document summaries.
  • Workflow Orchestration: CPU-based reasoning orchestrates GPU sandbox tasks and aggregates results to the orchestrator.

Quick Start

Provide a document reference (or upload) and request GPU-accelerated extraction, embedding, and structured reporting.

Frequently Asked Questions about gpu-document-processing

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I extract text from large PDF files while preserving layout and page references?

GPU-accelerated document processing extracts text from large PDFs while preserving headers, paragraphs, lists, tables, and page references. A CPU-based reasoning layer orchestrates GPU sandbox tasks to parse layouts and write structured findings.

What is the best way to convert PDF tables into structured CSV or DataFrame formats?

GPU-accelerated tabular data extraction converts PDF tables into structured formats like CSV or DataFrames. It performs type detection and handles multi-row headers, outputting the structured table data to a shared directory.

How do I chunk large documents and generate embeddings for bulk text analysis?

Document chunking and embedding generation split large PDFs into meaningful chunks and produce embeddings for search and analysis. GPU sandboxes handle the bulk processing in parallel, aggregating metadata and per-document summaries.

Can I process collections of 10 or more PDF files in parallel with GPU acceleration?

Bulk processing handles collections of 10 or more files in parallel using GPU acceleration. The CPU orchestrator manages the GPU sandbox tasks, extracts metadata, and synthesizes per-document summaries for the entire collection.

Do I need a GPU to run document extraction and embedding generation tasks?

GPU acceleration is required for the sandbox processing tasks that handle text extraction, table extraction, chunking, and embedding generation. A CPU-based reasoning layer orchestrates these GPU tasks and aggregates the final structured results.

Where are the structured findings and extracted data written after GPU document processing?

Extracted text, structured tables, chunks, and embeddings are written to the /shared/ directory. The CPU-based orchestrator aggregates the results from the GPU sandbox and saves the final structured findings there.