Taiwanese Mandarin Speech Corpus for ASR Pre-training

Offering audio and text data for pre-training Automatic Speech Recognition models, specifically for Taiwanese Mandarin. It is structured into training and test subsets using the WebDataset format for direct integration with PyTorch and Hugging Face tools. The dataset is published by the author 'adi-gov-tw' and was last updated in December 2025.

Use Cases

• Pre-train an ASR model on Taiwanese Mandarin audio files paired with text transcriptions.
• Fine-tune a speech recognition model using the provided training and test WebDataset tar files.
• Benchmark ASR model performance on the dedicated test set for Taiwanese Mandarin.

Strengths

• Data is formatted for direct use with popular machine learning frameworks like PyTorch and Hugging Face datasets.
• The dataset was updated in December 2025, indicating recent maintenance.

Limitations

• The dataset size, row count, and specific audio/text features are not provided.
• The license terms are unknown, which may restrict commercial use.

Provenance

Source

huggingface

Collection Method

Data is packaged in WebDataset tar format for machine learning workflows.

Freshness

Last updated on 2025-12-22.

Geography

Taiwan