WTForbes/th download history
WTForbes/th is a dataset on the Hugging Face Hub. In the last 30 days it was downloaded 33 times (0 in the last 7 days), and 65 times in total. It ranks #254,112 among datasets by monthly downloads.
th cleaned speech corpus This refresh provides a converged full train manifest, a frozen eval manifest, and a token-coverage subset of approximately 5000 hours. All three manifests reuse the existing validated FLAC-in-Parquet audio pool; no audio was copied, repacked, or transcoded fo