metadata
configs:
- config_name: th
data_files:
- split: train
path: corpus/th.train.clean.jsonl
- config_name: audio-aadonis
data_files:
- split: train
path: audio/aadonis/*/part-*.parquet
- config_name: audio-common_voice16
data_files:
- split: train
path: audio/common_voice16/*/part-*.parquet
- config_name: audio-fleurs
data_files:
- split: train
path: audio/fleurs/*/part-*.parquet
- config_name: audio-gigaspeech2
data_files:
- split: train
path: audio/gigaspeech2/*/part-*.parquet
- config_name: audio-worldspeech
data_files:
- split: train
path: audio/worldspeech/*/part-*.parquet
th high-quality speech corpus
This dataset contains the complete cleaned Thai corpus and its source-partitioned audio.
All audio is FLAC PCM16. Source sample rate and channel count are preserved. Every packaged item passed a source-to-FLAC bit-identical PCM round trip and a separate full-release decode/hash validation.
Corpus rows use portable locators of the form audio/<subset>/<group>/part-N.parquet#row=N.
Download only selected subsets
from huggingface_hub import snapshot_download
snapshot_download(
repo_id="ACCOUNT/th",
repo_type="dataset",
allow_patterns=[
"corpus/*",
"audio/aadonis/*/part-*.parquet",
],
)
Or load an individual configuration:
from datasets import load_dataset
corpus = load_dataset("ACCOUNT/th", "th")
audio = load_dataset("ACCOUNT/th", "audio-aadonis")
Release totals
- Unique audio: 11,762,414
- Audio hours: 13,430.474
- Parquet size: 0.914450 TB (0.831687 TiB)
See release.json for exact per-subset counts, hours, bytes, and corpus checksums.