xtuner/llava-phi-3-mini-gguf

Model

llava-phi-3-mini is a LLaVA model fine-tuned from microsoft/Phi-3-mini-4k-instruct and CLIP-ViT-Large-patch14-336 with ShareGPT4V-PT and InternVL-SFT by XTuner.

Note: This model is in GGUF format.

Resources:

• GitHub: xtuner
• Official LLaVA format model: xtuner/llava-phi-3-mini
• HuggingFace LLaVA format model: xtuner/llava-phi-3-mini-hf
• XTuner LLaVA format model: xtuner/llava-phi-3-mini-xtuner

Details

Model	Visual Encoder	Projector	Resolution	Pretraining Strategy	Fine-tuning Strategy	Pretrain Dataset	Fine-tune Dataset	Pretrain Epoch	Fine-tune Epoch
LLaVA-v1.5-7B	CLIP-L	MLP	336	Frozen LLM, Frozen ViT	Full LLM, Frozen ViT	LLaVA-PT (558K)	LLaVA-Mix (665K)	1	1
LLaVA-Llama-3-8B	CLIP-L	MLP	336	Frozen LLM, Frozen ViT	Full LLM, LoRA ViT	LLaVA-PT (558K)	LLaVA-Mix (665K)	1	1
LLaVA-Llama-3-8B-v1.1	CLIP-L	MLP	336	Frozen LLM, Frozen ViT	Full LLM, LoRA ViT	ShareGPT4V-PT (1246K)	InternVL-SFT (1268K)	1	1
LLaVA-Phi-3-mini	CLIP-L	MLP	336	Frozen LLM, Frozen ViT	Full LLM, Full ViT	ShareGPT4V-PT (1246K)	InternVL-SFT (1268K)	1	2

Results

Model	MMBench Test (EN)	MMMU Val	SEED-IMG	AI2D Test	ScienceQA Test	HallusionBench aAcc	POPE	GQA	TextVQA	MME	MMStar
LLaVA-v1.5-7B	66.5	35.3	60.5	54.8	70.4	44.9	85.9	62.0	58.2	1511/348	30.3
LLaVA-Llama-3-8B	68.9	36.8	69.8	60.9	73.3	47.3	87.2	63.5	58.0	1506/295	38.2
LLaVA-Llama-3-8B-v1.1	72.3	37.1	70.1	70.0	72.9	47.7	86.4	62.6	59.0	1469/349	45.1
LLaVA-Phi-3-mini	69.2	41.4	70.0	69.3	73.7	49.8	87.3	61.5	57.8	1477/313	43.7

Quickstart

Download models

# mmproj
wget https://huggingface.co/xtuner/llava-phi-3-mini-gguf/resolve/main/llava-phi-3-mini-mmproj-f16.gguf

# fp16 llm
wget https://huggingface.co/xtuner/llava-phi-3-mini-gguf/resolve/main/llava-phi-3-mini-f16.gguf

# int4 llm
wget https://huggingface.co/xtuner/llava-phi-3-mini-gguf/resolve/main/llava-phi-3-mini-int4.gguf

# (optional) ollama fp16 modelfile
wget https://huggingface.co/xtuner/llava-phi-3-mini-gguf/resolve/main/OLLAMA_MODELFILE_F16

# (optional) ollama int4 modelfile
wget https://huggingface.co/xtuner/llava-phi-3-mini-gguf/resolve/main/OLLAMA_MODELFILE_INT4

Chat by ollama

Note: llava-phi-3-mini uses the Phi-3-instruct chat template.

# fp16
ollama create llava-phi3-f16 -f ./OLLAMA_MODELFILE_F16
ollama run llava-phi3-f16 "xx.png Describe this image"

# int4
ollama create llava-phi3-int4 -f ./OLLAMA_MODELFILE_INT4
ollama run llava-phi3-int4 "xx.png Describe this image"

Chat by ./llava-cli

1. Build llama.cpp (docs) .
2. Build ./llava-cli (docs).

Note: llava-phi-3-mini uses the Phi-3-instruct chat template.

# fp16
./llava-cli -m ./llava-phi-3-mini-f16.gguf --mmproj ./llava-phi-3-mini-mmproj-f16.gguf --image YOUR_IMAGE.jpg -c 4096

# int4
./llava-cli -m ./llava-phi-3-mini-int4.gguf --mmproj ./llava-phi-3-mini-mmproj-f16.gguf --image YOUR_IMAGE.jpg -c 4096

Reproduce

Please refer to docs.

Citation

@misc{2023xtuner,
    title={XTuner: A Toolkit for Efficiently Fine-tuning LLM},
    author={XTuner Contributors},
    howpublished = {\url{https://github.com/InternLM/xtuner}},
    year={2023}
}