---
title: MiniCPM-Llama3-V-2_5-int4
canonical_url: "https://www.modelscope.ai/models/OpenBMB/MiniCPM-Llama3-V-2_5-int4"
md_url: "https://www.modelscope.ai/models/OpenBMB/MiniCPM-Llama3-V-2_5-int4.md"
repository: OpenBMB/MiniCPM-Llama3-V-2_5-int4
last_updated: 2025-09-19
pipeline_tag: visual-question-answering
tasks:
  - visual-question-answering
model_type:
  - minicpmv
base_model:
  - openbmb/MiniCPM-Llama3-V-2_5
base_model_relation: quantized
library_name:
  - pytorch
  - transformer
  - safetensors
frameworks:
  - Pytorch
downloads: 108
stars: 0
---

# MiniCPM-Llama3-V-2_5-int4

> MiniCPM-Llama3-V-2_5-int4 - An open-source model by OpenBMB on ModelScope. MiniCPM-Llama3-V 2.5 int4 This is the int4 quantized version of MiniCPM-Llama3-V 2.5. Running with int4 version would use lower GPU memory (about 9GB).

OpenBMB/MiniCPM-Llama3-V-2_5-int4 is a visual-question-answering model on ModelScope. derived from openbmb/MiniCPM-Llama3-V-2_5.

- **Repository**: OpenBMB/MiniCPM-Llama3-V-2_5-int4
- **Tasks**: visual-question-answering
- **Base model**: openbmb/MiniCPM-Llama3-V-2_5
- **Downloads**: 108
- **Stars**: 0
- **Last updated**: 2025-09-19

Source: https://www.modelscope.ai/models/OpenBMB/MiniCPM-Llama3-V-2_5-int4

---

## MiniCPM-Llama3-V 2.5 int4
This is the int4 quantized version of [MiniCPM-Llama3-V 2.5](https://huggingface.co/openbmb/MiniCPM-Llama3-V-2_5).   
Running with int4 version would use lower GPU memory (about 9GB).


## Usage
Inference using Huggingface transformers on NVIDIA GPUs. Requirements tested on python 3.10：
```
Pillow==10.1.0
torch==2.1.2
torchvision==0.16.2
transformers==4.40.0
sentencepiece==0.1.99
accelerate==0.30.1
bitsandbytes==0.43.1
```

```python
# test.py
import torch
from PIL import Image
from transformers import AutoModel, AutoTokenizer

model = AutoModel.from_pretrained('openbmb/MiniCPM-Llama3-V-2_5-int4', trust_remote_code=True)
tokenizer = AutoTokenizer.from_pretrained('openbmb/MiniCPM-Llama3-V-2_5-int4', trust_remote_code=True)
model.eval()

image = Image.open('xx.jpg').convert('RGB')
question = 'What is in the image?'
msgs = [{'role': 'user', 'content': question}]

res = model.chat(
    image=image,
    msgs=msgs,
    tokenizer=tokenizer,
    sampling=True, # if sampling=False, beam_search will be used by default
    temperature=0.7,
    # system_prompt='' # pass system_prompt if needed
)
print(res)

## if you want to use streaming, please make sure sampling=True and stream=True
## the model.chat will return a generator
res = model.chat(
    image=image,
    msgs=msgs,
    tokenizer=tokenizer,
    sampling=True,
    temperature=0.7,
    stream=True
)

generated_text = ""
for new_text in res:
    generated_text += new_text
    print(new_text, flush=True, end='')
```