Instructions to use lmms-lab/llava-onevision-qwen2-72b-ov-sft with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use lmms-lab/llava-onevision-qwen2-72b-ov-sft with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="lmms-lab/llava-onevision-qwen2-72b-ov-sft")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoProcessor, AutoModelForCausalLM

processor = AutoProcessor.from_pretrained("lmms-lab/llava-onevision-qwen2-72b-ov-sft")
model = AutoModelForCausalLM.from_pretrained("lmms-lab/llava-onevision-qwen2-72b-ov-sft")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = processor.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use lmms-lab/llava-onevision-qwen2-72b-ov-sft with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "lmms-lab/llava-onevision-qwen2-72b-ov-sft"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "lmms-lab/llava-onevision-qwen2-72b-ov-sft",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/lmms-lab/llava-onevision-qwen2-72b-ov-sft

SGLang

How to use lmms-lab/llava-onevision-qwen2-72b-ov-sft with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "lmms-lab/llava-onevision-qwen2-72b-ov-sft" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "lmms-lab/llava-onevision-qwen2-72b-ov-sft",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "lmms-lab/llava-onevision-qwen2-72b-ov-sft" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "lmms-lab/llava-onevision-qwen2-72b-ov-sft",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use lmms-lab/llava-onevision-qwen2-72b-ov-sft with Docker Model Runner:
```
docker model run hf.co/lmms-lab/llava-onevision-qwen2-72b-ov-sft
```

LLaVA-OneVision

Play with the model on the LLaVA OneVision Chat.

Model Summary
Use
Limitations
Training
License
Citation

Model Summary

The LLaVA-OneVision models are 0.5/7/72B parameter models trained on LLaVA-OneVision, based on Qwen2 language model with a context window of 32K tokens.

Repository: LLaVA-VL/LLaVA-NeXT
Project Website: llava-onevision.lmms-lab.com
Paper: LLaVA-OneVision
Point of Contact: Bo Li
Languages: English, Chinese

Use

Intended use

The model was trained on LLaVA-OneVision Dataset and have the ability to interact with images, multi-image and videos.

Feel free to share your generations in the Community tab!

Generation

# pip install git+https://github.com/LLaVA-VL/LLaVA-NeXT.git
from llava.model.builder import load_pretrained_model
from llava.mm_utils import get_model_name_from_path, process_images, tokenizer_image_token
from llava.constants import IMAGE_TOKEN_INDEX, DEFAULT_IMAGE_TOKEN, DEFAULT_IM_START_TOKEN, DEFAULT_IM_END_TOKEN, IGNORE_INDEX
from llava.conversation import conv_templates, SeparatorStyle

from PIL import Image
import requests
import copy
import torch

import sys
import warnings

warnings.filterwarnings("ignore")
pretrained = "lmms-lab/llava-onevision-qwen2-0.5b-si"
model_name = "llava_qwen"
device = "cuda"
device_map = "auto"
tokenizer, model, image_processor, max_length = load_pretrained_model(pretrained, None, model_name, device_map=device_map)  # Add any other thing you want to pass in llava_model_args

model.eval()

url = "https://github.com/haotian-liu/LLaVA/blob/1a91fc274d7c35a9b50b3cb29c4247ae5837ce39/images/llava_v1_5_radar.jpg?raw=true"
image = Image.open(requests.get(url, stream=True).raw)
image_tensor = process_images([image], image_processor, model.config)
image_tensor = [_image.to(dtype=torch.float16, device=device) for _image in image_tensor]

conv_template = "qwen_1_5"  # Make sure you use correct chat template for different models
question = DEFAULT_IMAGE_TOKEN + "\nWhat is shown in this image?"
conv = copy.deepcopy(conv_templates[conv_template])
conv.append_message(conv.roles[0], question)
conv.append_message(conv.roles[1], None)
prompt_question = conv.get_prompt()

input_ids = tokenizer_image_token(prompt_question, tokenizer, IMAGE_TOKEN_INDEX, return_tensors="pt").unsqueeze(0).to(device)
image_sizes = [image.size]


cont = model.generate(
    input_ids,
    images=image_tensor,
    image_sizes=image_sizes,
    do_sample=False,
    temperature=0,
    max_new_tokens=4096,
)
text_outputs = tokenizer.batch_decode(cont, skip_special_tokens=True)
print(text_outputs)

Training

Model

Architecture: SO400M + Qwen2
Pretraining Stage: LCS-558K, 1 epoch, projector
Mid Stage: A mixture of 4.7M high-quality synthetic data, 1 epoch, full model
Final-Image Stage: A mixture of 3.6M single-image data, 1 epoch, full model
OneVision Stage: A mixture of 1.6M single-image/multi-image/video data, 1 epoch, full model
Precision: bfloat16

Hardware & Software

GPUs: 256 * Nvidia Tesla A100 (for whole model series training)
Orchestration: Huggingface Trainer
Neural networks: PyTorch

Citation

@article{li2024llavaonevision,
      title={LLaVA-OneVision},
}

Downloads last month: 236

Safetensors

Model size

73B params

Tensor type

BF16

Model tree for lmms-lab/llava-onevision-qwen2-72b-ov-sft

Finetunes

1 model

Dataset used to train lmms-lab/llava-onevision-qwen2-72b-ov-sft

Collection including lmms-lab/llava-onevision-qwen2-72b-ov-sft

LLaVA-OneVision

Collection

a model good at arbitrary types of visual input • 17 items • Updated Sep 17, 2025 • 32

Paper for lmms-lab/llava-onevision-qwen2-72b-ov-sft

LLaVA-OneVision: Easy Visual Task Transfer

Paper • 2408.03326 • Published Aug 6, 2024 • 61

lmms-lab
/

llava-onevision-qwen2-72b-ov-sft

LLaVA-OneVision

Table of Contents

Model Summary

Use

Intended use

Generation

Training

Model

Hardware & Software

Citation

Model tree for lmms-lab/llava-onevision-qwen2-72b-ov-sft

Dataset used to train lmms-lab/llava-onevision-qwen2-72b-ov-sft

Collection including lmms-lab/llava-onevision-qwen2-72b-ov-sft

LLaVA-OneVision

Paper for lmms-lab/llava-onevision-qwen2-72b-ov-sft

LLaVA-OneVision: Easy Visual Task Transfer