BEANS-Next: Bioacoustic Event Annotation and Sound Benchmark

EarthSpeciesProject created this audio dataset for bioacoustic analysis, last updated on May 7, 2026. The dataset includes audio files and a metadata file with columns for file paths and benchmark task identifiers. It is structured with a tier system, likely containing different levels of annotation complexity.

Use Cases

• Benchmarking audio event detection models based on the described tiered annotation system.
• Training machine learning models for animal sound classification using the provided audio clips.
• Developing query-by-example audio retrieval systems based on the context and query audio path structure.

Strengths

• Metadata includes a structured tier system (1-4) for organizing annotation complexity.
• Audio files are organized for direct playback in the Hugging Face Dataset Viewer via the file_name column convention.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Freshness should be verified as the last update timestamp is in the future (2026-05-07).

Provenance

Source

EarthSpeciesProject

Freshness

Last updated 2026-05-07 05:59:30.