Deep-VLM/LLaVA-OneVision-1.5-4B-Instruct-hf

LLaVA-OneVision-1.5: Fully Open-Source State-of-the-Art VLM Model

✨ Key Features

LLaVA-OneVision-1.5 introduces a novel family of fully open-source Large Multimodal Models (LMMs) that achieves state-of-the-art performance with substantially lower cost through training on native resolution images.

1. Superior Performance A family of fully open-source large multimodal models demonstrating superior performance across multiple multimodal benchmarks, outperforming Qwen2.5-VL in most evaluation tasks.

2. High-Quality Data at Scale Meticulously curated mid-training and SFT data with rigorous filtering and quality control.

• Concept-balanced, highly diverse, high-quality caption data
• Comprehensive instruction fine-tuning data covering a wide range of tasks

3. Ultra-Efficient Training Framework Complete end-to-end training framework designed for maximum efficiency:

• $16K total budget for full model training
• 45% HFU efficiency on A100 GPUs ($0.6 per GPU/Hour)
• Built on MegatronLM with support for MoE, FP8, and long sequence parallelization
• Optimized codebase for cost-effective scaling

4. Fully Open Framework for community access and reproducibility:

• ✅ High-quality mid-training & SFT data
• ✅ Complete training framework & code
• ✅ Training recipes & configurations
• ✅ Base & instruct model checkpoints
• ✅ Comprehensive training logs & metrics

Code

This model is trained using a fully open-source, end-to-end training framework, with all code available at EvolvingLMMs-Lab/LLaVA-OneVision-1.5.

Models

Model	HF Link	Training Log
LLaVA-OV-1.5-4B-Instruct	🤗 HF / 4B-Instruct	📈 Tensorboard
LLaVA-OV-1.5-8B-Instruct	🤗 HF / 8B-Instruct	📈 Tensorboard

Dataset

Description	Link
Mid-training data for LLaVA-OneVision-1.5	🤗 Download (Uploading!)
SFT data for LLaVA-OneVision-1.5	🤗 Download (Uploading!)

Evaluation Results

All evaluations were conducted using lmms_eval.

Using 🤗 Transformers to Chat

Here we show a code snippet to show you how to use the chat model with transformers and qwen_vl_utils:

from transformers import AutoTokenizer, AutoProcessor, AutoModelForCausalLM
from qwen_vl_utils import process_vision_info
model_path = "lmms-lab/LLaVA-One-Vision-1.5-8B-Instruct"

# default: Load the model on the available device(s)
model = AutoModelForCausalLM.from_pretrained(
    model_path, torch_dtype="auto", device_map="auto", trust_remote_code=True
)

# default processer
processor = AutoProcessor.from_pretrained(model_path, trust_remote_code=True)

messages = [
    {
        "role": "user",
        "content": [
            {
                "type": "image",
                "image": "https://qianwen-res.oss-cn-beijing.aliyuncs.com/Qwen-VL/assets/demo.jpeg",
            },
            {"type": "text", "text": "Describe this image."},
        ],
    }
]

# Preparation for inference
text = processor.apply_chat_template(
    messages, tokenize=False, add_generation_prompt=True
)
image_inputs, video_inputs = process_vision_info(messages)
inputs = processor(
    text=[text],
    images=image_inputs,
    videos=video_inputs,
    padding=True,
    return_tensors="pt",
)
inputs = inputs.to("cuda")

# Inference: Generation of the output
generated_ids = model.generate(**inputs, max_new_tokens=1024)
generated_ids_trimmed = [
    out_ids[len(in_ids) :] for in_ids, out_ids in zip(inputs.input_ids, generated_ids)
]
output_text = processor.batch_decode(
    generated_ids_trimmed, skip_special_tokens=True, clean_up_tokenization_spaces=False
)
print(output_text)

Citation

If you find LLaVA-OneVision-1.5 useful in your research, please consider to cite the following related papers:

@inproceedings{LLaVA-OneVision-1.5,
  title={LLaVA-OneVision-1.5: Fully Open Framework for Democratized Multimodal Training},
  author={An, Xiang and Xie, Yin and Yang, Kaicheng and Zhang, Wenkang and Zhao, Xiuwei and Cheng, Zheng and Wang, Yirui and Xu, Songcen and Chen, Changrui and Wu, Chunsheng and Tan, Huajie and Li, Chunyuan and Yang, Jing and Yu, Jie and Wang, Xiyao and Qin, Bin and Wang, Yumeng and Yan, Zizhen and Feng, Ziyong and Liu, Ziwei and Li, Bo and Deng, Jiankang},
  booktitle={arxiv},  
  year={2025}
 }

@inproceedings{xie2025region,
  title={Region-based Cluster Discrimination for Visual Representation Learning},
  author={Xie, Yin and Yang, Kaicheng and An, Xiang and Wu, Kun and Zhao, Yongle and Deng, Weimo and Ran, Zimin and Wang, Yumeng and Feng, Ziyong and Miles, Roy and Elezi, Ismail and Deng, Jiankang},
  booktitle={ICCV},
  year={2025}
}

@article{lillava,
  title={LLaVA-OneVision: Easy Visual Task Transfer},
  author={Li, Bo and Zhang, Yuanhan and Guo, Dong and Zhang, Renrui and Li, Feng and Zhang, Hao and Zhang, Kaichen and Zhang, Peiyuan and Li, Yanwei and Liu, Ziwei and Li, Chunyuan},
  journal={Transactions on Machine Learning Research}
  year={2024}
}