Hugging Face Datasets

🤗 Datasets is a library that makes it easy to access and share AI datasets for audio, vision, and NLP tasks. With a single line of code you can load datasets from the Hugging Face Hub or local files, and its processing and streaming features let you quickly prepare them for deep learning training. Built on the Apache Arrow format, it handles even very large datasets without memory limits. It is deeply integrated with the Hugging Face Hub, so you can easily share datasets with the machine learning community.

• Load Hub or local datasets with a single load_dataset() line
• Supports many formats, including CSV, JSON, JSONL, Parquet, Arrow, XML, text, and Webdataset
• Supports multimodal data: text, audio, images, video, PDF, and NIfTI (3D medical imaging)
• Streaming mode that reads data directly without downloading it
• Zero-copy reads based on Apache Arrow for handling large data
• Fast, reproducible preprocessing with dataset.map()
• Integrates with NumPy, Pandas, PyTorch, TensorFlow, JAX, and Polars
• Share datasets on the Hugging Face Hub

Open source (free)

GitHub에 LICENSE 파일과 함께 공개된 Python 라이브러리로, 받은 본문에 이 라이브러리 자체의 가격·영업 문의 안내 없음(요금 페이지 본문은 비어 있음)

공식 페이지 확인 2026-10-14(산출 규칙 2026-10-15)

• One of 8 AI tools in the AI Data & Human Feedback field.
• Connected to 0 other tools — #7 in its field.