README.md

metadata

configs:
  - config_name: ru
    data_files:
      - split: train
        path: corpus/ru.train.clean.jsonl
      - split: eval
        path: corpus/ru.eval.clean.jsonl
  - config_name: ru-5000h
    data_files:
      - split: train
        path: corpus/ru.train.coverage-5000h.jsonl
      - split: eval
        path: corpus/ru.eval.clean.jsonl
  - config_name: audio-aadonis_multilingual_audio_alignments__ru
    data_files:
      - split: train
        path: audio/aadonis_multilingual_audio_alignments__ru/*/part-*.parquet
  - config_name: audio-common_voice_21_ru
    data_files:
      - split: train
        path: audio/common_voice_21_ru/*/part-*.parquet
  - config_name: audio-d3vastated_common_voice_ru_en__ru
    data_files:
      - split: train
        path: audio/d3vastated_common_voice_ru_en__ru/*/part-*.parquet
  - config_name: audio-dialogs
    data_files:
      - split: train
        path: audio/dialogs/*/part-*.parquet
  - config_name: audio-dusha_crowd
    data_files:
      - split: train
        path: audio/dusha_crowd/*/part-*.parquet
  - config_name: audio-espnet_yodas_granary__ru
    data_files:
      - split: train
        path: audio/espnet_yodas_granary__ru/*/part-*.parquet
  - config_name: audio-fleurs_ru
    data_files:
      - split: train
        path: audio/fleurs_ru/*/part-*.parquet
  - config_name: audio-golos
    data_files:
      - split: train
        path: audio/golos/*/part-*.parquet
  - config_name: audio-rmndrnts_common_voice_ru_noised__default
    data_files:
      - split: train
        path: audio/rmndrnts_common_voice_ru_noised__default/*/part-*.parquet
  - config_name: audio-ru_audiobooks_devices
    data_files:
      - split: train
        path: audio/ru_audiobooks_devices/*/part-*.parquet
  - config_name: audio-rudevices_audio_records
    data_files:
      - split: train
        path: audio/rudevices_audio_records/*/part-*.parquet
  - config_name: audio-ruls
    data_files:
      - split: train
        path: audio/ruls/*/part-*.parquet
  - config_name: audio-ruslan
    data_files:
      - split: train
        path: audio/ruslan/*/part-*.parquet

ru cleaned speech corpus

This refresh provides a converged full train manifest, a frozen eval manifest, and a token-coverage subset of approximately 5000 hours. The validated FLAC payloads are reused without decoding or transcoding. Completely referenced Parquet shards are hard-linked locally; partial shards are filtered into exact row-only Parquet files.

Corpus overview

Variant	Rows	Hours
train	4,650,979	4465.559
coverage_5000h	4,304,751	3879.927
eval	1,563	2.013

All audio locators are portable and have the form audio/<subset>/<group>/part-N.parquet#row=N.

The release audit records exact input/output hashes, rejected audio mappings, language-tag validation, and zero train/eval audio-identity overlap.

from datasets import load_dataset
ds = load_dataset("YuKuanFu/ru", "ru-5000h", split="train")

Coverage-size note

The completed strict eligible pool contains 3,880.000 source hours and could not reach 5,000 hours. After packaging exclusions, the uploaded ru-5000h train contains 3,879.927 hours. The default configuration uses the larger conservative full-clean train manifest.

Packaging exclusions

Rows that could not be represented by the validated FLAC pool are omitted and recorded here rather than silently replaced:

Variant	Rows	Hours	Reasons
train	70	0.073172	all_zero_audio: 70
coverage_5000h	70	0.073172	all_zero_audio: 70