INSTALL.md

Installation

Requirements

• Linux (Windows is not officially supported)
• Python 3.5+
• PyTorch 1.1 or higher
• CUDA 9.0 or higher

We have tested the following versions of OS and softwares:

• OS: Ubuntu 16.04
• Python: 3.6/3.7
• PyTorch: 1.1/1.5/1.6
• CUDA: 9.0/11.0

Install OpenUnReID

a. Create a conda virtual environment and activate it.

conda create -n open-mmlab python=3.7 -y
conda activate open-mmlab

b. Install PyTorch and torchvision following the official instructions, e.g.,

conda install pytorch torchvision -c pytorch

c. Clone the this repository.

git clone https://github.com/open-mmlab/OpenUnReID.git
cd OpenUnReID

d. Install the dependent libraries.

pip install -r requirements.txt

e. Install openunreid library.

python setup.py develop

Prepare datasets

It is recommended to symlink your dataset root to OpenUnReID/datasets. If your folder structure is different, you may need to change the corresponding paths (namely DATA_ROOT) in config files.

Download the datasets from

• DukeMTMC-reID: [Google Drive] [Baidu Pan] (password: bhbh)
• Market-1501-v15.09.15: [Google Drive] [Baidu Pan]
• subset1 (PersonX): [Google Drive]
• MSMT17_V1: [Home Page] (request link by email the holder)
• VehicleID_V1.0: [Home Page] (request link by email the holder)
• AIC20_ReID_Simulation (VehicleX): [Home Page] (request password by email the holder)
• VeRi_with_plate: [Home Page] (request link by email the holder)

Save them under

OpenUnReID
└── datasets
    ├── dukemtmcreid
    │   └── DukeMTMC-reID
    ├── market1501
    │   └── Market-1501-v15.09.15
    ├── msmt17
    │   └── MSMT17_V1
    ├── personx
    │   └── subset1
    ├── vehicleid
    │   └── VehicleID_V1.0
    ├── vehiclex
    │   └── AIC20_ReID_Simulation
    └── veri
        └── VeRi_with_plate