Unified Model Records

LLaVa MM-Instruct

Type: dataset

Tags: api Publisher: OpenAI Released: None v.1.0.0

Metadata

General information.

name
LLaVa MM-Instruct
version
1.0.0
publisher
OpenAI
model type
Multi-Modal LLM
description
LLaVA utilized GPT-4 to generate multimodal instruction-following data, resulting in 158,000 unique samples. This dataset includes 58,000 conversational samples, 23,000 detailed descriptions, and 77,000 complex reasoning samples.

Relations

Relationship Graph

Relationship Graph for mm-instruct-data

Intended Use

No intended use specified.

Factors

No factors specified.

Evaluation Data

No evaluation data specified.

Training Data

No training data specified.

Additional Information

No additional information specified.

Recommendations

No recommendations specified.