---
dataset_info:
  features:
  - name: audio
    dtype:
      audio:
        sampling_rate: 16000
  - name: text
    dtype: string
  splits:
  - name: train
    num_bytes: 11189157343.599
    num_examples: 382731
  - name: test
    num_bytes: 1117996019.704
    num_examples: 12078
  download_size: 11973404473
  dataset_size: 12307153363.303001
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: test
    path: data/test-*
language:
- th
---

# Thai Fine Tuning Dataset

Dataset for Finetuning Audio models for Thai. Built from a combination of:

- mort666/cv_corpus_v22
- shunyalabs/thai-speech-dataset
- CMKL/Porjai-Thai-voice-dataset-central

Features reduced to required base for use with openai/whisper-large-v3 and Qwen/Qwen3-ASR model fine tuning.