huggingface-datasets

You can query and preview Hugging Face datasets by sending standard HTTP requests to the Dataset Viewer API. This allows you to list splits, validate datasets, and preview rows directly without manual browsing.

Yes, you can perform server-side filtering on large-scale machine learning datasets via the Dataset Viewer API. This enables efficient row-level filtering, searching, and statistical analysis without downloading the entire dataset locally.

Yes, the huggingface-datasets skill supports uploading local parquet shards and agent session traces. It uses the Hub CLI and an optional Node.js environment to manage data versioning and collaborative analysis on the Hugging Face Hub.

The best way to manage dataset versioning for agent session traces is using the Hub CLI through this skill. It automates uploading parquet shards or session traces to the Hugging Face Hub for collaborative analysis.

You do not need a Node.js environment solely to explore Hugging Face datasets, as data discovery and querying use standard HTTP requests. Node.js is only optionally required for Hub CLI operations like uploading parquet files.

Yes, you can access private Hugging Face datasets through the Dataset Viewer API. The skill facilitates programmatic access to validate, list splits, and preview rows of both public and private datasets directly.