Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,61 @@
|
|
| 1 |
-
|
| 2 |
-
|
| 3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# VPTracker: Global Vision-Language Tracking via Visual Prompt and MLLM
|
| 2 |
+
|
| 3 |
+
[](https://arxiv.org/abs/2512.22799)
|
| 4 |
+
[](https://www.python.org/downloads/)
|
| 5 |
+
[](https://pytorch.org/)
|
| 6 |
+
[](https://huggingface.co/docs/transformers/)
|
| 7 |
+
|
| 8 |
+
<img src="assets/VPTracker.jpg" width="800">
|
| 9 |
+
|
| 10 |
+
## 🚀 Quick Start
|
| 11 |
+
|
| 12 |
+
### Installation
|
| 13 |
+
|
| 14 |
+
```bash
|
| 15 |
+
conda create -n gltrack python==3.10
|
| 16 |
+
conda activate gltrack
|
| 17 |
+
|
| 18 |
+
cd ms-swift
|
| 19 |
+
conda install -c conda-forge pyarrow sentencepiece
|
| 20 |
+
pip install -e .
|
| 21 |
+
pip install "sglang[all]" -U
|
| 22 |
+
pip install "vllm>=0.5.1" "transformers<4.55" "trl<0.21" -U
|
| 23 |
+
pip install "lmdeploy>=0.5" -U
|
| 24 |
+
pip install autoawq -U --no-deps
|
| 25 |
+
pip install auto_gptq optimum bitsandbytes "gradio<5.33" -U
|
| 26 |
+
pip install git+https://github.com/modelscope/ms-swift.git
|
| 27 |
+
pip install timm -U
|
| 28 |
+
pip install "deepspeed" -U
|
| 29 |
+
pip install flash-attn==2.7.4.post1 --no-build-isolation
|
| 30 |
+
|
| 31 |
+
conda install av -c conda-forge
|
| 32 |
+
pip install qwen_vl_utils qwen_omni_utils decord librosa icecream soundfile -U
|
| 33 |
+
pip install liger_kernel nvitop pre-commit math_verify py-spy -U
|
| 34 |
+
|
| 35 |
+
```
|
| 36 |
+
|
| 37 |
+
## 👀 Visualization
|
| 38 |
+
<img src="assets/Results.jpg" width="800">
|
| 39 |
+
|
| 40 |
+
## 🙏 Acknowledgments
|
| 41 |
+
This code is developed on the top of [ms-swift](https://github.com/modelscope/ms-swift)
|
| 42 |
+
|
| 43 |
+
## ✉️ Contact
|
| 44 |
+
|
| 45 |
+
Email: jcwang@stu.ecnu.edu.cn. Any kind discussions are welcomed!
|
| 46 |
+
|
| 47 |
+
---
|
| 48 |
+
|
| 49 |
+
## 📖 Citation
|
| 50 |
+
If our work is useful for your research, please consider cite:
|
| 51 |
+
```
|
| 52 |
+
@misc{wang2025vptrackerglobalvisionlanguagetracking,
|
| 53 |
+
title={VPTracker: Global Vision-Language Tracking via Visual Prompt and MLLM},
|
| 54 |
+
author={Jingchao Wang and Kaiwen Zhou and Zhijian Wu and Kunhua Ji and Dingjiang Huang and Yefeng Zheng},
|
| 55 |
+
year={2025},
|
| 56 |
+
eprint={2512.22799},
|
| 57 |
+
archivePrefix={arXiv},
|
| 58 |
+
primaryClass={cs.CV},
|
| 59 |
+
url={https://arxiv.org/abs/2512.22799},
|
| 60 |
+
}
|
| 61 |
+
```
|