STARK-LT

The STARK tracker for the VOT2021-LT challenge

Install the environment

Option1: Use the Anaconda

conda create -n stark python=3.6
conda activate stark
bash install.sh
pip install lap
pip install cython_bbox
pip install shapely
pip install sklearn
pip install mmcv==0.4.0
pip install terminaltables
cd ./global_track/_submodules/mmdetection
(if there exists 'build' package, you need to delete it and re-build.)
pip install -r requirements/build.txt
python setup.py develop

重点重点！安装 vot-toolkit-python

pip install git+https://github.com/votchallenge/vot-toolkit-python

Set paths

Run the following command to set paths for this project

python tracking/create_default_local_file.py --workspace_dir . --data_dir ./data --save_dir .

After running this command, you can also modify paths by editing these two files

lib/train/admin/local.py  # paths about training
lib/test/evaluation/local.py  # paths about testing

Download the model checkpoints

Download STARKST_ep0050.pth.tar and put it under checkpoints/train/stark_st2/baseline
Download STARKST_ep0500.pth.tar and put it under checkpoints/train/stark_ref/baseline

Test and evaluate STARK on benchmarks

VOT2021-LT

• Modify the <PATH_OF_STARK> in trackers.ini to the absolute path of the STARK project on your local machine.
• Modify the base_path in vot_path.py to the absolute path of the STARK project on your local machine.
• VOT2021-LT dataset should be put in under stark_st50_ref_baseline_R0 dir, called sequences.

cd VOT21/LT/stark_st50_ref_baseline_R0
bash exp.sh

问题

trax 编译

在 trax 目录下执行：

mkdir build
cd build
cmake ..
make

参考链接：trax 官网 https://trax.readthedocs.io/en/latest/tutorial_compiling.html

运行

bash exp.sh

• vot connt found: 没有安装 vot-toolkit-python，解决方法：

pip install git+https://github.com/votchallenge/vot-toolkit-python

参考链接1：https://www.votchallenge.net/howto/tutorial_python.html 参考链接2：votchallenge/toolkit#21

STARK 执行命令

Train STARK

Training with multiple GPUs using DDP

# STARK-S50
python tracking/train.py --script stark_s --config baseline --save_dir . --mode multiple --nproc_per_node 8  # STARK-S50
# STARK-ST50
python tracking/train.py --script stark_st1 --config baseline --save_dir . --mode multiple --nproc_per_node 8  # STARK-ST50 Stage1
python tracking/train.py --script stark_st2 --config baseline --save_dir . --mode multiple --nproc_per_node 8 --script_prv stark_st1 --config_prv baseline  # STARK-ST50 Stage2
# STARK-ST101
python tracking/train.py --script stark_st1 --config baseline_R101 --save_dir . --mode multiple --nproc_per_node 8  # STARK-ST101 Stage1
python tracking/train.py --script stark_st2 --config baseline_R101 --save_dir . --mode multiple --nproc_per_node 8 --script_prv stark_st1 --config_prv baseline_R101  # STARK-ST101 Stage2

(Optionally) Debugging training with a single GPU

python tracking/train.py --script stark_s --config baseline --save_dir . --mode single

Test and evaluate STARK on benchmarks

• LaSOT

python tracking/test.py stark_st baseline --dataset lasot --threads 16
python tracking/analysis_results.py # need to modify tracker configs and names

• GOT10K-test

python tracking/test.py stark_st baseline_got10k_only --dataset got10k_test --threads 16
python lib/test/utils/transform_got10k.py --tracker_name stark_st --cfg_name baseline_got10k_only

• TrackingNet

python tracking/test.py stark_st baseline --dataset trackingnet --threads 16
python lib/test/utils/transform_trackingnet.py --tracker_name stark_st --cfg_name baseline

• VOT2020
  Before evaluating "STARK+AR" on VOT2020, please install some extra packages following external/AR/README.md

cd external/vot20/<workspace_dir>
export PYTHONPATH=<path to the stark project>:$PYTHONPATH
bash exp.sh

• VOT2020-LT

cd external/vot20_lt/<workspace_dir>
export PYTHONPATH=<path to the stark project>:$PYTHONPATH
bash exp.sh

Test FLOPs, Params, and Speed

# Profiling STARK-S50 model
python tracking/profile_model.py --script stark_s --config baseline
# Profiling STARK-ST50 model
python tracking/profile_model.py --script stark_st2 --config baseline
# Profiling STARK-ST101 model
python tracking/profile_model.py --script stark_st2 --config baseline_R101
# Profiling STARK-Lightning-X-trt
python tracking/profile_model_lightning_X_trt.py

Model Zoo

The trained models, the training logs, and the raw tracking results are provided in the model zoo