open-llava-next-llama3-8b Model Card

Model details

Model type: open-llava-next-llama3-8b is an open-source implementation chatbot trained by fine-tuning the entire model on open-source Open-LLaVA-Next-mix1M data.

Model date: open-llava-next-llama3-8b was trained in May 2024.

Paper or resources for more information: [Code]

Name	ViT	LLM	Weights	MME	SEED	SQA	MMB	MMB-CN	TextVQA	GQA
llava-next-vicuna-7b	CLIP-L-336	Vicuna-7B	HF	1519	70.2	70.1	67.4	60.6	64.9	64.2
open-llava-next-vicuna-7b	CLIP-L-336	Vicuna-7B	HF	1540	71.1	70.7	68.5	60.7	67.2	64.3
open-llava-next-llama3-8b	CLIP-L-336	LLaMA3-8B	HF	1552	74.4	77.3	74.4	70.4	69.8	65.9

Usage

You can utilize this model as we provide in our [repository]. Moreover, you can direct load this model and use it in the [LLaVA repository].

Training dataset

All training data are open-sourced in our repository.

• ShareGPT4V Mix665K
• 200K GPT4V-generated instruction data (ALLaVA)
• 200K various VQA data

Intended use

Primary intended uses: The primary use of open-llava-next-llama3-8b is research on large multimodal models and chatbots.

Primary intended users: The primary intended users of the model are researchers and hobbyists in computer vision, natural language processing, machine learning, and artificial intelligence.