BaltiVoice ASR: Speech Recognition Data for the Balti Language

BaltiVoice is the first publicly available Automatic Speech Recognition dataset for the critically low-resource Balti language (ISO 639-3: bft). The dataset was collected, validated, and processed by mohdali1 to build the first open-source Balti ASR system using OpenAI Whisper fine-tuning. It was last updated on the Hugging Face platform on June 11, 2026.

Use Cases

• Fine-tuning speech recognition models like OpenAI Whisper based on the described audio-text pairs.
• Benchmarking ASR system performance for the critically low-resource Balti language.
• Researching linguistic features and acoustic models for Tibetic languages.
• Developing language technology applications for speakers in the Gilgit-Baltistan and Ladakh regions.

Strengths

• First publicly available ASR dataset for the critically low-resource Balti language.
• Designed for building the first open-source Balti ASR system using OpenAI Whisper.
• Data was collected, validated, and processed specifically for ASR model fine-tuning.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count, file formats, and dataset size are unknown, which may limit suitability assessment.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

Hugging Face dataset uploaded by author mohdali1.

Collection Method

Collected, validated, and processed for ASR system development.

Freshness

Last updated 2026-06-11 12:41:02; freshness should be verified.

Geography

Primarily the Gilgit-Baltistan region of Pakistan and parts of Ladakh, India.