Alpaca Shona Taco: Parallel Instruction-Response Data for Shona and English

A dataset following the TaCo paper format, created by saillab and last updated on September 20, 2024. It contains structured text entries with instructions and responses in both Shona and English. The data is intended for research in machine translation and instruction-following models.

Use Cases

• Train machine translation models based on parallel Shona-English response pairs.
• Fine-tune instruction-following language models based on the structured 'instruction', 'input', and 'output' format.
• Create evaluation benchmarks for cross-lingual natural language understanding tasks.
• Study the characteristics of the Shona language in a machine-readable, structured format.

Strengths

• Follows a documented, structured format from the TaCo paper.
• Provides parallel text in Shona and English, which is valuable for low-resource language tasks.
• Last updated on 2024-09-20, indicating recent maintenance.

Limitations

• Row count, file size, and specific column definitions are unknown.
• Column-level documentation is absent; field semantics must be inferred after download.
• License information is not provided, which may restrict usage.

Provenance

Source

saillab on Hugging Face

Collection Method

Created for the TaCo paper; specific collection method is not detailed in the provided description.

Freshness

Last updated 2024-09-20 22:08:59