dataset-processing-multiprocessing

Community

Process large datasets in parallel with ease.

Authoranhvth
Version1.0.0
Installs0

System Documentation

What problem does it solve?

Large dataset preprocessing often becomes a bottleneck due to memory constraints and sequential processing. This Skill enables safe, parallelized handling of tokenization, packing, and merging for big HuggingFace datasets.

Core Features & Use Cases

  • Distributed sharding across CPU cores to maximize throughput while keeping workers isolated.
  • End-to-end data prep pipeline: load, shard, tokenize, pack, and merge into a final dataset.
  • Use Case: preprocess and tokenize multi-GB datasets for model pretraining with robust error handling and incremental saves.

Quick Start

Run the example_tokenize_pack.py script with your source dataset path and a tokenizer to start end-to-end processing.

Dependency Matrix

Required Modules

None required

Components

Standard package

💻 Claude Code Installation

Recommended: Let Claude install automatically. Simply copy and paste the text below to Claude Code.

Please help me install this Skill:
Name: dataset-processing-multiprocessing
Download link: https://github.com/anhvth/speedy_utils/archive/main.zip#dataset-processing-multiprocessing

Please download this .zip file, extract it, and install it in the .claude/skills/ directory.
View Source Repository

Agent Skills Search Helper

Install a tiny helper to your Agent, search and equip skill from 223,000+ vetted skills library on demand.