instinct-org/yt4_chunked_speech_restorised_nfa_aligned

Forced-Aligned STT Dataset Source dataset: instinct-org/yt4_chunked_speech_restorised Aligned dataset: instinct-org/yt4_chunked_speech_restorised_nfa_aligned Rows: 965192 successfully aligned rows This dataset adds NeMo Forced Aligner metadata for STT training and timestamp quality control. Source rows that did not produce usable CTM alignments are excluded from the published data. Added columns: nfa_token_alignments: NeMo token/subword CTM spans nfa_word_alignments: word-level… See the full description on the dataset page: https://huggingface.co/datasets/instinct-org/yt4_chunked_speech_restorised_nfa_aligned.