ACM: Audio-Centric Multimodal Benchmark for Retrieval

ACM packages the Audio-Centric Multimodal benchmark introduced by the paper 'Efficient and High-Fidelity Omni Modality Retrieval'. The dataset contains four HuggingFace subsets, each using a single test split to preserve natural schemas for queries and candidates. It was authored by chuonghm and last updated on July 4, 2026.

Use Cases

• Benchmarking audio-text retrieval models based on the composed_audio_retrieval_queries and candidates subsets.
• Training multimodal embedding models for audio-centric tasks using the structured test splits.
• Evaluating the performance of omni-modality retrieval systems as described in the associated arXiv paper.

Strengths

• The dataset is structured with a single test split per subset to preserve natural schemas for queries and candidates.
• It is directly associated with a published research paper available at arXiv:2603.02098.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count and file formats are unknown, which may limit suitability assessment.

Provenance

Source

HuggingFace repository chuonghm/ACM.

Freshness

Last updated 2026-07-04 21:43:45.