united-nations/transcription-results
UN Transcription Benchmark Results Evaluation results for speech-to-text systems on UN Security Council and General Assembly meeting recordings, assessed against official UN verbatim records. See united-nations/transcription-corpus for the underlying audio and ground truth data. Metrics WER: Word Error Rate (reference = verbatim record, no normalization) normalized_wer: WER after lowercasing, punctuation removal, and filler word removal CER: Character Error Rate… See the full description on the dataset page: https://huggingface.co/datasets/united-nations/transcription-results.
Update results: 361 rows, 2026-04-06
Update results: 290 rows, 2026-04-06
Update results: 359 rows, 2026-04-06
Update results: 342 rows, 2026-04-06
Update results: 221 rows, 2026-04-05
Update results: 192 rows, 2026-04-05
Update results: 186 rows, 2026-03-03
Update results: 75 rows, 2026-03-02
Update results: 18 rows, 2026-02-28
initial commit
