Maleo Short 1.5H: A Speaker Diarization Benchmark for Complex Media

Maleo Short 1.5H is a manually curated and rigorously annotated dataset designed to benchmark State-of-the-Art speaker diarization models. It focuses on complex, 'in-the-wild' media domains where models typically struggle, such as content with overlapping speech and sound effects. The dataset was created by maleo-ai and was last updated on Hugging Face in May 2026.

Use Cases

• Benchmarking diarization model performance based on the described complex 'in-the-wild' media domains.
• Training models to handle overlapping speech based on the annotation of this specific challenge.
• Improving diarization robustness against sound effects based on the dataset's focus on noisy audio environments.
• Evaluating model performance on rapid speaker shifts based on the described dataset characteristics.

Strengths

• The dataset is manually curated and rigorously annotated, suggesting high-quality labels.
• It is specifically designed to benchmark SOTA models against complex, 'in-the-wild' audio scenarios.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count and dataset size are unknown, which may limit suitability assessment.
• The description notes models 'heavily struggle' with the data, indicating it presents a very difficult challenge that may not be suitable for all applications.

Provenance

Source

maleo-ai

Collection Method

Manually curated and annotated.

Freshness

Last updated 2026-05-25 01:28:13.