{"id":714,"slug":"ath-maas--marco_longspeech","name":"Marco_Longspeech","author":"ATH-MaaS","description":"\n\n\t\n\t\t\n\t\n\t\n\t\tMarco-LongSpeech Dataset\n\t\n\n\n\n\n\n\n\nMarco-LongSpeech is a multi-task long speech understanding dataset containing 8 different speech understanding tasks designed to benchmark Large Language Models on lengthy audio inputs.\n\n\t\n\t\t\n\t\n\t\n\t\t📊 Dataset Statistics\n\t\n\n\n\t\n\t\t\n\t\n\t\n\t\tTask Statistics\n\t\n\n\n\t\n\t\t\nTask\nTrain\nVal\nTest\nTotal\nUnique Audios\n\n\n\t\t\nASR\n71,275\n15,273\n15,274\n101,822\n101,822\n\n\nTemporal_Relative_QA\n5,886\n1,261\n1,262\n8,409\n8,409\n\n\nsummary\n4,366\n935\n937\n6,238\n6,238… See the full description on the dataset page: https://huggingface.co/datasets/ATH-MaaS/Marco_Longspeech.","tags":"[\"Task_categories:automatic-Speech-Recognition\",\"Task_categories:audio-Classification\",\"Task_categories:text-Generation\",\"Language:en\",\"Language:zh\",\"Size_categories:10K<n<100K\"]","license":null,"framework":null,"parameters":null,"downloads":133785,"likes":18,"verified":0,"created_at":"2026-07-11 11:23:33","updated_at":"2026-07-11 18:23:35","source_url":"https://huggingface.co/datasets/ATH-MaaS/Marco_Longspeech","source_platform":"huggingface","hf_repo_id":"ATH-MaaS/Marco_Longspeech","ollama_name":"","category":"dataset","latest_version":"v1.0.0","version_count":1,"signature_count":1,"risk_level":null,"risk_score":null,"versions":[{"id":713,"model_id":714,"version":"v1.0.0","manifest_hash":"b3ca1afa14008d7db0b90acd87f586afc1e4944f0c54d43d02600d7b65799bd1","file_count":0,"total_size":0,"r2_manifest_key":"manifests/datasets/ath-maas--marco_longspeech/v1.0.0.json","created_at":"2026-07-11 11:23:33"}],"files":[],"signatures":[{"id":1277,"version_id":713,"signer_did":"did:quantamrkt:registry:shield-v1","algorithm":"ML-DSA-65","signature_hex":"c6608f638b5a3712a044f4b09a9c4afd2183aa0ecc6a4c2414b032dbdfbcb084","attestation_type":"registry","signed_at":"2026-07-11 11:23:33"}],"hndl":null}