---
license: mit
---
This is the official repository for the XTrack Base model.
It includes our foundation models (OSTrack-Base for XTrack-Base and DropMAE for XTrack-Large) as well as the pre-trained weights for XTrack-Base.

Papers：  
[DropMAE: Masked Autoencoders With Spatial-Attention Dropout for Tracking Tasks, CVPR 2023](https://openaccess.thecvf.com/content/CVPR2023/papers/Wu_DropMAE_Masked_Autoencoders_With_Spatial-Attention_Dropout_for_Tracking_Tasks_CVPR_2023_paper.pdf)

[OSTrack: Towards Robust and Accurate Visual Tracking via Transformer, ECCV 2022](https://arxiv.org/abs/2203.11991)