img2dataset

The Reader component parses input URL lists from multiple formats (CSV, JSON, JSONL, Parquet) and extracts associated metadata like captions, alt text, and image attributes. It uses temporary feather files for memory-efficient handling of large datasets, sharding the input into work units that can be distributed across workers. This design allows processing of datasets ranging from thousands to billions of images without loading entire datasets into memory.

The Downloader component creates a thread pool to fetch multiple images concurrently from URLs, integrating HTTP request handling, optional hash verification, robots.txt directive checking, image decoding, and error handling throughout the pipeline. Each worker maintains its own thread pool, allowing fine-grained control over concurrency levels and connection pooling. The architecture supports custom HTTP headers, timeout configuration, and graceful handling of network failures with retry logic.

The Resizer component applies configurable image transformations including multiple resize modes (e.g., center crop, pad, stretch), format conversion, and quality normalization. It supports various resize strategies to handle aspect ratio preservation, enabling datasets with consistent dimensions for model training. The component integrates with the download pipeline to process images immediately after decoding, reducing memory footprint by avoiding storage of full-resolution intermediates.

The SampleWriter component outputs processed images and metadata in multiple formats optimized for different ML frameworks (WebDataset, Parquet, LMDB, TFRecord). It handles sharded output to avoid bottlenecks, writing data in parallel across workers. The component manages file organization, metadata serialization, and format-specific optimizations (e.g., tar-based streaming for WebDataset, columnar storage for Parquet). This architecture enables seamless integration with downstream ML pipelines.

The multiprocessing distributor allocates work units across multiple CPU cores on a single machine using Python's multiprocessing module. It spawns worker processes that each run independent Downloader instances, coordinating through a shared work queue and logger process. This strategy maximizes hardware utilization for datasets that fit within single-machine resources, avoiding the overhead of distributed computing frameworks.

The PySpark distributor scales image downloading across a Spark cluster by partitioning work units into RDDs and distributing them to Spark executors. Each executor runs a Downloader instance, with Spark handling fault tolerance, load balancing, and resource management. This strategy enables processing of massive datasets (billions of images) across commodity clusters while providing automatic recovery from node failures.

The Ray distributor scales image downloading across Ray clusters (on-premises or cloud-based) by creating remote tasks that execute Downloader instances on Ray workers. Ray handles dynamic resource allocation, auto-scaling, and fault recovery. This strategy enables elastic scaling on cloud platforms (AWS, GCP, Azure) with minimal infrastructure management, supporting both on-demand and spot instances.

The Logger component monitors the entire download pipeline in real-time, collecting statistics on download success rates, processing speed, error types, and resource utilization. It runs as a separate process to avoid blocking worker threads, aggregating metrics from all workers and writing periodic reports. The logger provides visibility into pipeline health, enabling detection of bottlenecks, network issues, or configuration problems.

The pipeline supports resuming interrupted downloads by tracking completed work units and skipping already-processed images. It uses metadata (URLs, hashes) to detect duplicates across runs, avoiding redundant downloads. This capability enables long-running pipelines to recover from failures without reprocessing, and supports incremental dataset growth by appending new images to existing datasets.

The downloader supports custom HTTP headers (User-Agent, Authorization, etc.) for accessing protected or restricted image sources. It integrates robots.txt checking to respect website crawling directives, parsing robots.txt files and validating URLs against allow/disallow rules before downloading. This enables ethical dataset creation while supporting authentication-protected image sources.