Skip to content

Dataset · Porameht

Cleaned Thai Speech Dataset for ASR

Porameht/processed-voice-th-169k

169k Thai audio recordings with transcriptions for ASR fine-tuning.

What's inside

The dataset contains 169,550 audio clips with Thai text transcriptions (split into training, validation, and test sets). The dataset is prepared for automatic speech recognition (ASR) tasks and has a volume between 100K and 1M items.

What you can do with it

  • Fine-tuning ASR models, such as Whisper, for the Thai language
Tasks
Automatic speech recognition
Languages
Thai
Size
100K–1M rows
Downloads this month
318
Likes
18

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("Porameht/processed-voice-th-169k")