1. Paper download address
STMTrack: Template-free Visual Tracking with Space-time Memory Networks CVPR (2021). [paper][code]
2. Code download address
https://github.com/fzh0917/STMTrack
3. Create a virtual environment and activate it
conda create -n STMTrack python=3.7 -y
source activate STMTrack
4. Install torch and torchvision
pip install torch===1.4.0 -f https://download.pytorch.org/whl/torch_stable.html
pip install torchvision===0.5.0 -f https://download.pytorch.org/whl/torch_stable.html
5. Install dependent libraries
pip install -r requirements.txt
6. Download the pre-trained model
Link: https://pan.baidu.com/s/16aqJy4eMFTDjEXuWFxvlLg
Extraction code: z4aj
Place the downloaded pre-training model into the newly created pretrain_model path in the project directory
7. Experimental settings
打开~/STMTrack-main/experiments/stmtrack/test/otb/stmtrack-googlenet-otb.yaml
1) Change the path of the pre-trained model, which is the path where we stored it in step 6
2) Change device_num
3) Add the path where the data set is located
data_root: "/home1/publicData/Datasets/OTB100"
8. Run the code
python main/test.py --config experiments/stmtrack/test/otb/stmtrack-googlenet-otb.yaml
You may encounter the following errors
File "pycocotools/_mask.pyx", line 1, in init pycocotools._mask
ValueError: numpy.ndarray size changed, may indicate binary incompatibility. Expected 88 from C header, got 80 from PyObject
Solution
The numpy version is too low, so you need to upgrade the numpy version!
pip install --upgrade numpy
9. Run successfully