Llava En Zh 300K: 300,000 Multimodal Instruction Examples

300,000 examples of visual instruction data for training multimodal large language models. The dataset combines 150,000 English examples from the LLaVA project and 150,000 from the openbmb project. Author BUAADreamer uploaded this collection to Hugging Face on September 2, 2024.

Use Cases

• Fine-tuning vision-language models based on the described multimodal instruction examples.
• Training models for visual question answering based on the instruction data.
• Benchmarking model performance on multimodal instruction-following tasks.
• Conducting research on instruction tuning for multimodal AI systems.

Strengths

• Contains 300,000 total examples, providing a substantial volume of training data.
• Combines data from two established sources: LLaVA and openbmb.
• Specifically formatted for use with the LLaMA Factory training toolkit.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Data may reflect source bias inherent to the contributing projects LLaVA and openbmb.

Provenance

Source

Combined from the LLaVA and openbmb projects.

Collection Method

Likely curated and aggregated from existing visual instruction datasets.

Time Range

null

Freshness

Last updated 2024-09-02 14:20:59; freshness should be verified.

Geography

null