University-1652_triplet-loss

Baseline Code (with bottleneck) for University-1652 (pytorch).

Any suggestion is welcomed.

Model Structure

You may learn more from model.py. We use the L2-norm 2048-dim feature as the input.

Tips

• I did not optimize the code. I strongly suggest use fp16 and use with torch.no_grad(). I will update the code later.
• Larger margin may lead to a worse local minimum. (margin = 0.1-0.3 may provide a better result.)
• Per-class sampler (Satisfied sampler)is not neccessary.
• Adam optimizer is not neccessary.

Prerequisites

• Python 3.6
• GPU Memory >= 6G
• Numpy
• Pytorch 0.3+

(Some reports found that updating numpy can arrive the right accuracy. If you only get 50~80 Top1 Accuracy, just try it.) We have successfully run the code based on numpy 1.12.1 and 1.13.1 .

Getting started

Installation

• Install Pytorch from http://pytorch.org/
• Install Torchvision from the source

git clone https://github.com/pytorch/vision
cd vision
python setup.py install

Because pytorch and torchvision are ongoing projects.

Here we noted that our code is tested based on Pytorch 0.3.0/0.4.0 and Torchvision 0.2.0.

Dataset & Preparation

Download Market1501 Dataset

Preparation: Put the images with the same id in one folder. You may use

python prepare.py

Remember to change the dataset path to your own path.

Futhermore, you also can test our code on DukeMTMC-reID Dataset. Our baseline code is not such high on DukeMTMC-reID Rank@1=64.23%, mAP=43.92%. Hyperparameters are need to be tuned.

To save trained model, we make a dir.

mkdir model 

Train

Train a model by

python train_new.py --gpu_ids 0 --name ft_ResNet50 --train_all --batchsize 32  --data_dir your_data_path

--gpu_ids which gpu to run.

--name the name of model.

--data_dir the path of the training data.

--train_all using all images to train.

--batchsize batch size.

--erasing_p random erasing probability.

Train a model with random erasing by

python train_new.py --gpu_ids 0 --name ft_ResNet50 --train_all --batchsize 32  --data_dir your_data_path --erasing_p 0.5

Test

Use trained model to extract feature by

python test.py --gpu_ids 0 --name ft_ResNet50 --test_dir your_data_path  --which_epoch 59

--gpu_ids which gpu to run.

--name the dir name of trained model.

--which_epoch select the i-th model.

--data_dir the path of the testing data.

Evaluation

python evaluate.py

It will output Rank@1, Rank@5, Rank@10 and mAP results. You may also try evaluate_gpu.py to conduct a faster evaluation with GPU.

For mAP calculation, you also can refer to the C++ code for Oxford Building. We use the triangle mAP calculation (consistent with the Market1501 original code).