Category: File Formats
Definition
TAR.GZ files combine TAR archiving with GZIP compression, the standard format for distributing large model files and datasets in the ML community.
How It Works
TAR (Tape Archive) bundles multiple files while preserving directory structure and permissions. GZIP compression then reduces the archive size by 50-90%.
The format supports streaming extraction, allowing processing to begin before the entire file is downloaded.
Why It Matters
TAR.GZ is ubiquitous in ML research for distributing multi-gigabyte models and datasets. Its efficiency and universal support make it ideal for academic sharing.
Most pre-trained models and benchmark datasets are distributed as TAR.GZ files on platforms like AWS S3.
← Back to File Formats | All Terms