---
license: llama3.2
base_model:
- meta-llama/Llama-3.2-11B-Vision-Instruct
language:
- en
- ko
tags:
- vlm-ko
- meta
- llama-3.2
- llama-3.2-ko
datasets:
- maum-ai/General-Evol-VQA
---
<p align="left">
  <img src="https://cdn-uploads.huggingface.co/production/uploads/646484cfb90150b2706df03b/BEOyMpnnY9VY2KXlc3V2F.png" width="20%"/>
<p>

# Llama-3.2-MAAL-11B-Vision-v0.1
**Llama-3.2-MAAL-11B-Vision-v0.1** is bilingual multimodal model trained for text and visual understanding across Korean and English languages. We are releasing a [model](https://huggingface.co/maum-ai/Llama-3.2-MAAL-11B-Vision-v0.1), a subset of the [training dataset](https://huggingface.co/datasets/maum-ai/General-Evol-VQA), and a [leaderboard](https://huggingface.co/spaces/maum-ai/KOFFVQA-Leaderboard) to promote and accelerate the development of Korean Vision-Language Models (VLMs).

- **Developed by:** [maum.ai Brain NLP](https://maum-ai.github.io). Jaeyoon Jung, Yoonshik Kim, Yekyung Nah
- **Language(s) (NLP):** Korean, English (currently, bilingual)


## Model Description

Version 0.1 is fine-tuned by English and Korean VQA datasets with other datasets (OCR, Math, etc)...

- We trained this model on 8 H100-80G for 2 days with image-text pair multimodal fine-tuning dataset
- [maum-ai/General-Evol-VQA](https://huggingface.co/datasets/maum-ai/General-Evol-VQA) is one of the datasets that we used for fine-tuning.

## sample inference code (GPU)
Starting with transformers >= 4.45.0 onward, you can run inference to generate text based on an image and a starting prompt you supply.

```
import requests
import torch
from PIL import Image
from transformers import MllamaForConditionalGeneration, AutoProcessor

model_id = "maum-ai/Llama-3.2-MAAL-11B-Vision-v0.1"

model = MllamaForConditionalGeneration.from_pretrained(
    model_id,
    torch_dtype=torch.bfloat16,
    device_map="auto",
)
processor = AutoProcessor.from_pretrained(model_id)

url = "https://huggingface.co/datasets/huggingface/documentation-images/resolve/0052a70beed5bf71b92610a43a52df6d286cd5f3/diffusers/rabbit.jpg"
image = Image.open(requests.get(url, stream=True).raw)

messages = [
    {"role": "user", "content": [
        {"type": "image"},
        {"type": "text", "text": "이 이미지에 대해서 시를 써줘"}
    ]}
]
input_text = processor.apply_chat_template(messages, add_generation_prompt=True)
inputs = processor(
    image,
    input_text,
    add_special_tokens=False,
    return_tensors="pt"
).to(model.device)

output = model.generate(**inputs, max_new_tokens=200)
print(processor.decode(output[0]))
```

## Evaluation Results
As the main goal of version 0.1 is **leveraging Korean VQA and OCR capabilities tailored to real-world business use cases**, we select [**KOFFVQA**](https://huggingface.co/spaces/maum-ai/KOFFVQA-Leaderboard) as our evaluation method to assess the Korean instruction-following skills. 

|Model|Params (B)|average(↑)|
|-|-|-|
|NCSOFT/VARCO-VISION-14B|15.2b|66.69|
|Qwen/Qwen2-VL-7B-Instruct|8.3b|63.53|
|**maum-ai/Llama-3.2-MAAL-11B-Vision-v0.1**|10.7b|61.13|
|meta-llama/Llama-3.2-11B-Vision-Instruct|10.7b|50.36|
|mistralai/Pixtral-12B-2409|12.7b|44.62|
|llava-onevision-qwen2-7b-ov|8b|43.78|
|InternVL2-8b|8.1b|32.76|
|MiniCPM-V-2_6|8.1b|32.69|

Our model has achieved a **20%** performance improvement compared to the previous base model.
You can check more results in [this Leaderboard](https://huggingface.co/spaces/maum-ai/KOFFVQA-Leaderboard)

### We will release enhanced model, v0.2 soon