Datasetpublic

lighthouse-emnlp2024/AudioMomentRetrievalFromLongAudio_DCASE2026EvaluationData

What is this? This repository contains data for DCASE Challenge 2026 Task 6.  Audio and text features using CLAP and a sliding window, following the same feature extraction protocol as the CASTELLA dataset. submission template File structure clap     └──dcase2026_evaluation_audio_{vid}.npz clap_text    └──qiddcase2026_evaluation_q{qid}.npz Raw audio files If participants require the raw audio, please contact the… See the full description on the dataset page: https://huggingface.co/datasets/lighthouse-emnlp2024/AudioMomentRetrievalFromLongAudio_DCASE2026EvaluationData.

sourceHugging Facecc-by-4.0updated 4mo agoView on Hugging Face

0likes21downloads

Dataset Card

What is this?

This repository contains data for DCASE Challenge 2026 Task 6.

—Audio and text features using CLAP and a sliding window, following the same feature extraction protocol as the CASTELLA dataset.
—submission template

File structure

clap 
   └──dcase2026_evaluation_audio_{vid}.npz
clap_text
   └──qiddcase2026_evaluation_q{qid}.npz

Raw audio files

If participants require the raw audio, please contact the organizers

Submission

—Predict the relevant moments for each qid in the submission template. The audio and text features in this repository correspond to the qid and vid in the template.

—Fill in the predicted moments in the pred_relevant_windows field. Multiple moments can be submitted, which will be used to compute the mAP. Moments should be listed in descending order of confidence, and only the first moment in the list will be used for the final ranking.