Datasets:

marcosremar2
/

pt-br-tts-synth

Tasks:

Text-to-Speech

Modalities:

Formats:

Languages:

Size:

Libraries:

License:

Dataset card Data Studio Files Files and versions

xet

Community

pt-br-tts-synth / README.md

marcosremar2

Upload folder using huggingface_hub

a573a7b verified 3 months ago

preview code

Raw

History Blame Contribute Delete

525 Bytes

metadata

license: apache-2.0
language:
  - pt
task_categories:
  - text-to-speech

pt-br-tts-synth

98813 frases PT-BR sintetizadas com Kokoro-82M (vozes pf_dora/pm_alex/pm_santa, speeds 0.9-1.1x), 16kHz mono WAV em 32 tar shards (WebDataset). Texto gerado por LLM (3 tiers de complexidade x 40 topicos); transcricao, tier, topico, voz e speed em metadata.jsonl.

from datasets import load_dataset
ds = load_dataset("webdataset", data_files="hf://datasets/marcosremar2/pt-br-tts-synth/shard_*.tar", split="train")