---
license: apache-2.0
---
PaddleOCR-VL-1.5: Towards a Multi-Task 0.9B VLM for Robust In-the-Wild Document Parsing
[](https://github.com/PaddlePaddle/PaddleOCR)
[](https://huggingface.co/PaddlePaddle/PaddleOCR-VL-1.5)
[](https://modelscope.cn/models/PaddlePaddle/PaddleOCR-VL-1.5)
[](https://huggingface.co/spaces/PaddlePaddle/PaddleOCR-VL-1.5_Online_Demo)
[](https://modelscope.cn/studios/PaddlePaddle/PaddleOCR-VL-1.5_Online_Demo/summary)
[](https://discord.gg/JPmZXDsEEK)
[](https://x.com/PaddlePaddle)
[](./LICENSE)
**🔥 [Official Website](https://www.paddleocr.com)**
**đź“ť [Technical Report](https://arxiv.org/abs/2601.21957)**
## Introduction
**PaddleOCR-VL-1.5 is an advanced next-generation model of PaddleOCR-VL, achieving a new state-of-the-art accuracy of 94.5% on OmniDocBench v1.5.** To rigorously evaluate robustness against real-world physical distortions—including scanning artifacts, skew, warping, screen photography, and illumination—we propose the Real5-OmniDocBench benchmark. Experimental results demonstrate that this enhanced model attains SOTA performance on the newly curated benchmark. Furthermore, we extend the model’s capabilities by incorporating seal recognition and text spotting tasks, while remaining a 0.9B ultra-compact VLM with high efficiency.
### **Model Architecture**
## PaddleOCR-VL-1.5 Usage with llama.cpp
### Install Dependencies
Install [PaddlePaddle](https://www.paddlepaddle.org.cn/install/quick) and [PaddleOCR](https://github.com/PaddlePaddle/PaddleOCR):
```bash
# The following command installs the PaddlePaddle version for CUDA 12.6. For other CUDA versions and the CPU version, please refer to https://www.paddlepaddle.org.cn/en/install/quick?docurl=/documentation/docs/en/develop/install/pip/linux-pip_en.html
python -m pip install paddlepaddle-gpu==3.2.1 -i https://www.paddlepaddle.org.cn/packages/stable/cu126/
python -m pip install -U "paddleocr[doc-parser]"
```
> **Please ensure that you install PaddlePaddle framework version 3.2.1 or above, along with the special version of safetensors.** For macOS users, please use Docker to set up the environment.
### Basic Usage
1. Start the VLM inference server:
```
llama-server \
-m /path/to/PaddleOCR-VL-1.5-GGUF.gguf \
--mmproj /path/to/PaddleOCR-VL-1.5-GGUF-mmproj.gguf \
--port 8080 \
--host 0.0.0.0 \
--temp 0
```
2. Call the PaddleOCR CLI or Python API:
```bash
paddleocr doc_parser \
-i https://paddle-model-ecology.bj.bcebos.com/paddlex/imgs/demo_image/paddleocr_vl_demo.png \
--vl_rec_backend llama-cpp-server \
--vl_rec_server_url http://127.0.0.1:8080/v1
```
```python
from paddleocr import PaddleOCRVL
pipeline = PaddleOCRVL(vl_rec_backend="llama-cpp-server", vl_rec_server_url="http://127.0.0.1:8080/v1")
output = pipeline.predict("https://paddle-model-ecology.bj.bcebos.com/paddlex/imgs/demo_image/paddleocr_vl_demo.png")
for res in output:
res.print()
res.save_to_json(save_path="output")
res.save_to_markdown(save_path="output")
```
**For more usage details and parameter explanations, see the [documentation](https://www.paddleocr.ai/latest/en/version3.x/pipeline_usage/PaddleOCR-VL.html).**
## PaddleOCR-VL-1.5-0.9B Usage with llama.cpp
Currently, the PaddleOCR-VL-1.5-0.9B model facilitates seamless inference via the transformers library, supporting comprehensive text spotting and the recognition of complex elements including formulas, tables, charts, and seals. Below is a simple script we provide to support inference using the PaddleOCR-VL-1.5-0.9B model with llama.cpp.
#### Usage Tips
We have six types of element-level recognition:
- Text recognition, indicated by the prompt `OCR:`.
- Formula recognition, indicated by the prompt `Formula Recognition:`.
- Table recognition, indicated by the prompt `Table Recognition:`.
- Chart recognition, indicated by the prompt `Chart Recognition:`.
- Seal recognition, indicated by the prompt `Seal Recognition:`.
- Spotting, indicated by the prompt `Spotting:`, and need to set `image_max_pixels` to `1605632`:
```
git clone https://github.com/ggml-org/llama.cpp.git
cd llama.cpp
python -m pip install gguf
python ./gguf-py/gguf/scripts/gguf_set_metadata.py /path/to/PaddleOCR-VL-1.5-GGUF/PaddleOCR-VL-1.5-mmproj.gguf clip.vision.image_max_pixels 1605632 --force
# back to default value (1003520):
# python ./gguf-py/gguf/scripts/gguf_set_metadata.py /path/to/PaddleOCR-VL-1.5-GGUF/PaddleOCR-VL-1.5-mmproj.gguf clip.vision.image_max_pixels 1003520 --force
```
#### llama-cli Usage
```shell
llama-cli \
-m /path/to/PaddleOCR-VL-1.5-GGUF/PaddleOCR-VL-1.5.gguf \
--mmproj /path/to/PaddleOCR-VL-1.5-GGUF/PaddleOCR-VL-1.5-mmproj.gguf \
-p 'OCR:' \
--image 'test_image.jpg'
```
#### llama-server Usage
```
llama-server -m /path/to/PaddleOCR-VL-1.5.gguf --mmproj /path/to/PaddleOCR-VL-1.5-GGUF/PaddleOCR-VL-1.5-mmproj.gguf --temp 0
```
# Acknowledgements
We extend our gratitude to [@megemini](https://github.com/megemini) for their significant [pull request](https://github.com/ggml-org/llama.cpp/pull/18825), which adds support for [PaddleOCR-VL-1.5-0.9B](https://huggingface.co/PaddlePaddle/PaddleOCR-VL-1.5) and [PaddleOCR-VL-0.9B](https://huggingface.co/PaddlePaddle/PaddleOCR-VL) models to llama.cpp.