Llava Instruct 9K: A Multimodal Instruction Dataset for Vision and Voice Tasks

A multimodal dataset derived from the LLaVA-Instruct-150K source, containing synthetic annotations for tasks involving text, images, and speech. It is licensed under CC-BY-4.0 and was uploaded by author dreyn74. The dataset's size is indicated to be between 10,000 and 100,000 samples.

Use Cases

• Fine-tuning models for visual question answering based on image-text pairs.
• Training text-to-speech or speech-to-speech systems using synthetic speech data.
• Developing multimodal assistants capable of image-to-text and image-to-speech generation.
• Benchmarking automatic speech recognition models on synthetic instruction data.

Strengths

• Covers multiple multimodal task categories, including text-to-text, image-to-text, and speech-related tasks.
• Built upon the established LLaVA-Instruct-150K source dataset.
• Uses a permissive CC-BY-4.0 license for open use and redistribution.

Limitations

• Row count is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• Data is synthetically generated, which may not reflect real-world distributions or complexities.

Provenance

Source

liuhaotian/LLaVA-Instruct-150K on Hugging Face

Collection Method

Synthetic annotations derived from a source dataset.

Freshness

Last updated 2026-06-03 01:22:12; freshness should be verified.