Access is gated because this repository contains speech-derived pair metadata. Request access only for approved training, evaluation, or research use.

By requesting access, you agree not to redistribute files, attempt speaker re-identification, or use the data for unauthorized voice impersonation.

tbp_chunked_speech_restorised_tts_train_clone_pairs_raw

This is a gated Russian raw TTS clone-pair plan for tbp_chunked_speech_restorised. It contains speaker-reference and target metadata rows used to build tokenized clone-pair training shards.

clone_pair_plan.jsonl: one reference/target pair per row
clone_pair_plan_summary.json: upload-time summary

Dataset Summary

Field	Value
Source dataset key	`tbp_chunked_speech_restorised`
Pair rows	42793

Intended Use

text-to-speech clone-pair training preparation
speaker-reference conditioning experiments
Internal dataset curation, quality checks, and evaluation
Research or commercial use only after access approval and license review

Data Notes

This is a raw metadata plan, not tokenized audio shards.
Pair rows contain reference and target identifiers, text fields, audio references, and quality metadata.
Tokenized versions should be used for direct training pipelines that expect shard manifests.

Access

This repository is gated with manual approval. Access requests are reviewed by the dataset maintainers. Users requesting access must provide their affiliation, intended use, and agreement to the dataset terms.