---
dataset_info:
  features:
  - name: audio_path
    dtype: string
  - name: transcript
    dtype: string
  - name: ipa_transcript_word_batched_split
    dtype: string
  splits:
  - name: train
    num_bytes: 61812813
    num_examples: 157691
  download_size: 24193861
  dataset_size: 61812813
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
license: apache-2.0
task_categories:
- translation
- text2text-generation
language:
- th
---
# Optional card data
cardData:
  tag: audio # Or speech-recognition, speech-processing, etc.
  language: en # Or the appropriate language code
  license: apache-2.0 # Or your chosen license

# --- Crucial Dataset Loading Info ---
datasets:
- dataset_info:
    # Configuration name
    config_name: default
    # How to load the data file(s)
    data_files:
    - split: train # Assuming this is your main/training data
      path: dataset.csv # Path to your CSV file in the repo
    # Specify the CSV loading parameters
    # IMPORTANT: Tell it the delimiter is '|'
    download_config:
      separator: "|" # Specify the pipe delimiter
    # Define the columns and their types (recommended)
    features:
    - name: audio_path           # Column 1
      dtype: string              # Will display the path as text
    - name: transcript           # Column 2
      dtype: string
    - name: ipa_transcript_word_batched_split # Column 3
      dtype: string
---

## My Audio Transcription Dataset

Describe your dataset here.

*   **audio_path:** Path or identifier for the audio file.
*   **transcript:** The orthographic transcription.
*   **ipa_transcript_word_batched_split:** IPA transcription... (add more detail).

(Add any other relevant information about data collection, preprocessing, etc.)