README.md

metadata

language:
  - uz
  - ru
task_categories:
  - automatic-speech-recognition
tags:
  - forced-alignment
  - nemo
  - stt-training

Forced-Aligned STT Dataset

Source dataset: instinct-org/zy_chunked

Aligned dataset: instinct-org/zy_chunked_nfa_aligned

Rows: 534816 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 CTM spans
• nfa_segment_alignments: segment-level CTM spans
• nfa_character_alignments: character spans; true only when the model token CTM is character level, otherwise derived from word spans
• nfa_character_alignment_source
• nfa_true_character_alignment
• nfa_alignment_model
• nfa_alignment_status
• nfa_alignment_error

NeMo Forced Aligner emits token, word, and segment CTM levels for this model family. With BPE/subword ASR models, token level is not the same as character level.

Model: nvidia/stt_uz_fastconformer_hybrid_large_pc