Transformer Tracking (CVPR2021)

Related tags

Deep Learning TransT
Overview

TransT - Transformer Tracking [CVPR2021]

Official implementation of the TransT (CVPR2021) , including training code and trained models.

We are revising the paper and will upload it in the next week

Results

Model LaSOT
AUC (%)
TrackingNet
AUC (%)
GOT-10k
AO (%)
OTB100
AUC (%)
NFS
AUC (%)
UAV123
AUC (%)
Speed
Params
TransT-N2 64.2 80.9 69.9 69.3 65.4 66.0 65.6fps 16.7M
TransT-N4 64.9 81.4 72.3 69.0 65.3 68.1 47.3fps 23.0M

Installation

This document contains detailed instructions for installing the necessary dependencied for TransT. The instructions have been tested on Ubuntu 18.04 system.

Install dependencies

  • Create and activate a conda environment

    conda create -n transt python=3.7
    conda activate transt
  • Install PyTorch

    conda install -c pytorch pytorch=1.5 torchvision=0.6.1 cudatoolkit=10.2
  • Install other packages

    conda install matplotlib pandas tqdm
    pip install opencv-python tb-nightly visdom scikit-image tikzplotlib gdown
    conda install cython scipy
    pip install pycocotools jpeg4py
    pip install wget yacs
    pip install shapely==1.6.4.post2
  • Setup the environment
    Create the default environment setting files.

    # Change directory to <PATH_of_TransT>
    cd TransT
    
    # Environment settings for pytracking. Saved at pytracking/evaluation/local.py
    python -c "from pytracking.evaluation.environment import create_default_local_file; create_default_local_file()"
    
    # Environment settings for ltr. Saved at ltr/admin/local.py
    python -c "from ltr.admin.environment import create_default_local_file; create_default_local_file()"

You can modify these files to set the paths to datasets, results paths etc.

  • Add the project path to environment variables
    Open ~/.bashrc, and add the following line to the end. Note to change <path_of_TransT> to your real path.
    export PYTHONPATH=<path_of_TransT>:$PYTHONPATH
    
  • Download the pre-trained networks
    Download the network for TransT and put it in the directory set by "network_path" in "pytracking/evaluation/local.py". By default, it is set to pytracking/networks.

Quick Start

Traning

  • Modify local.py to set the paths to datasets, results paths etc.
  • Runing the following commands to train the TransT. You can customize some parameters by modifying transt.py
    conda activate transt
    cd TransT/ltr
    python run_training.py transt transt

Evaluation

  • We integrated PySOT for evaluation.

    You need to specify the path of the model and dataset in the test.py.

    net_path = '/path_to_model' #Absolute path of the model
    dataset_root= '/path_to_datasets' #Absolute path of the datasets

    Then run the following commands.

    conda activate TransT
    cd TransT
    python -u pysot_toolkit/test.py --dataset <name of dataset> --name 'transt' #test tracker #test tracker
    python pysot_toolkit/eval.py --tracker_path results/ --dataset <name of dataset> --num 1 --tracker_prefix 'transt' #eval tracker

    The testing results will in the current directory(results/dataset/transt/)

  • You can also use pytracking to test and evaluate tracker. The results might be slightly different with PySOT due to the slight difference in implementation (pytracking saves results as integers, pysot toolkit saves the results as decimals).

Acknowledgement

This is a modified version of the python framework PyTracking based on Pytorch, also borrowing from PySOT and detr. We would like to thank their authors for providing great frameworks and toolkits.

Contact

  • Xin Chen (email:[email protected])

    Feel free to contact me if you have additional questions.

Comments
  • About training with 8 gpu

    About training with 8 gpu

    Hi,

    I found that the training time of the final 500 epochs does not change much whether it is 8 cards or 2 cards. In theory, shouldn’t it be accelerated by 4 times with 8 cards?

    Thanks!

    opened by memoiry 9
  • 关于attention map的问题

    关于attention map的问题

    作者你好,我看了之前您回答其他人关于如何画attention map的问题,我尝试把代码加了进去`
    src12, attn_map1 = self.self_attn1(q1, k1, value=src1, attn_mask=src1_mask, key_padding_mask=src1_key_padding_mask) src22, attn_map2 = self.self_attn2(q2, k2, value=src2, attn_mask=src2_mask, key_padding_mask=src2_key_padding_mask) attn_map1 = attn_map1.cpu().data.numpy()[0].reshape(256,16,16)[0] attn_map2 = attn_map2.cpu().data.numpy()[0].reshape(1024,32,32)[0]

        def pltshow(pred_map, name):
           import matplotlib.pyplot as plt
           plt.figure(2)
           pred_frame = plt.gca()
           plt.imshow(pred_map, 'jet')
           pred_frame.axes.get_yaxis().set_visible(False)
           pred_frame.axes.get_xaxis().set_visible(False)
           pred_frame.spines['top'].set_visible(False)
           pred_frame.spines['bottom'].set_visible(False)
           pred_frame.spines['left'].set_visible(False)
           pred_frame.spines['right'].set_visible(False)
           pred_name = '/home/sun/桌面/TransT/heatmap/' + name + '.png'
           plt.savefig(pred_name, bbox_inches='tight', pad_inches=0, dpi=150)
           plt.close(2)
    
        pltshow(attn_map1, 'aaa')
        pltshow(attn_map2, 'bbb')`
        src12, attn_map11 = self.multihead_attn1(query=self.with_pos_embed(src1, pos_src1),
                                   key=self.with_pos_embed(src2, pos_src2),
                                   value=src2, attn_mask=src2_mask,
                                   key_padding_mask=src2_key_padding_mask)
        src22, attn_map12 = self.multihead_attn2(query=self.with_pos_embed(src2, pos_src2),
                                   key=self.with_pos_embed(src1, pos_src1),
                                   value=src1, attn_mask=src1_mask,
                                   key_padding_mask=src1_key_padding_mask)
        attn_map11 = attn_map11.cpu().data.numpy()[0].reshape(256,32,32)[0]
        attn_map12 = attn_map12.cpu().data.numpy()[0].reshape(1024,16,16)[0]
        pltshow(attn_map11, 'ccc')
        pltshow(attn_map12, 'ddd') 
    

    我加进去之后,发现第二个图bbb一直在左上角有很高的响应值,无论什么序列,还有第四副图ddd,他的高响应值都在四个角上,和论文中的不一样,我想问下为什么会这样,是我的代码有问题吗?期待您的回复!

    opened by sunmengyu11 7
  • Transt Checkpoint file

    Transt Checkpoint file

    Hi,

    I wanna work on your tracker and train on my own dateset, unfortunately there is no checkpoint file available in your GitHub, can you upload the checkpoint file?

    Obviously, I will cite your paper after getting appropriate results.

    opened by amirhamidihd 6
  • [Errno 13] Permission denied: '/tensorboard'

    [Errno 13] Permission denied: '/tensorboard'

    作者您好,我在运行python run_training.py transt transt时,出现了如标题所示的错误,请问该如何解决? Traceback (most recent call last): File "run_training.py", line 55, in main() File "run_training.py", line 50, in main run_training(args.train_module, args.train_name, args.cudnn_benchmark) File "run_training.py", line 39, in run_training expr_func(settings) File "../ltr/train_settings/transt/transt.py", line 98, in run trainer = LTRTrainer(actor, [loader_train], optimizer, settings, lr_scheduler) File "../ltr/trainers/ltr_trainer.py", line 30, in init self.tensorboard_writer = TensorboardWriter(tensorboard_writer_dir, [l.name for l in loaders]) File "../ltr/admin/tensorboard.py", line 13, in init self.writer = OrderedDict({name: SummaryWriter(os.path.join(self.directory, name)) for name in loader_names}) File "../ltr/admin/tensorboard.py", line 13, in self.writer = OrderedDict({name: SummaryWriter(os.path.join(self.directory, name)) for name in loader_names}) File "/home/lab318/anaconda3/envs/transt/lib/python3.7/site-packages/torch/utils/tensorboard/writer.py", line 225, in init self._get_file_writer() File "/home/lab318/anaconda3/envs/transt/lib/python3.7/site-packages/torch/utils/tensorboard/writer.py", line 256, in _get_file_writer self.flush_secs, self.filename_suffix) File "/home/lab318/anaconda3/envs/transt/lib/python3.7/site-packages/torch/utils/tensorboard/writer.py", line 66, in init log_dir, max_queue, flush_secs, filename_suffix) File "/home/lab318/anaconda3/envs/transt/lib/python3.7/site-packages/tensorboard/summary/writer/event_file_writer.py", line 72, in init tf.io.gfile.makedirs(logdir) File "/home/lab318/anaconda3/envs/transt/lib/python3.7/site-packages/tensorboard/compat/tensorflow_stub/io/gfile.py", line 653, in makedirs return get_filesystem(path).makedirs(path) File "/home/lab318/anaconda3/envs/transt/lib/python3.7/site-packages/tensorboard/compat/tensorflow_stub/io/gfile.py", line 191, in makedirs os.makedirs(path, exist_ok=True) File "/home/lab318/anaconda3/envs/transt/lib/python3.7/os.py", line 213, in makedirs makedirs(head, exist_ok=exist_ok) File "/home/lab318/anaconda3/envs/transt/lib/python3.7/os.py", line 213, in makedirs makedirs(head, exist_ok=exist_ok) File "/home/lab318/anaconda3/envs/transt/lib/python3.7/os.py", line 213, in makedirs makedirs(head, exist_ok=exist_ok) [Previous line repeated 1 more time] File "/home/lab318/anaconda3/envs/transt/lib/python3.7/os.py", line 223, in makedirs mkdir(name, mode) PermissionError: [Errno 13] Permission denied: '/tensorboard'

    opened by LeungWaiHo 6
  • Also about the inference speed

    Also about the inference speed

    hello, I have test the model(trans.pth) in GOT10k-test, but I also got the lower inference speed (around 30 fps) than the paper had mentioned(47 fps). Need to note that I use the same GPU: RTX-TITAN with you and I'm sure no other programs occupy the computing resources. Will the inference speed be affected by other factors? PS my CPU: Intel(R) Xeon(R) Silver 4114 CPU @ 2.20GHz Snipaste_2021-05-19_22-01-19

    opened by DonDominic 4
  • 关于缩短训练时间的参数设置

    关于缩短训练时间的参数设置

    能否提供训练500个epoch时的训练设置TransT/ltr/train_settings/transt/transt.py,(学习率,weight_decay等参数)。我发现我自己修改的使用500个epoch,学习率变为原来的两倍,然后在第250个epoch进行学习率衰减的策略性能下降的比较多,不清楚是哪方面原因造成的,谢谢。

    opened by RaymondCover 4
  • trackingnet训练出现 KeyError: 'Nf1aqv5Fg5o_0'

    trackingnet训练出现 KeyError: 'Nf1aqv5Fg5o_0'

    您好,我发现在训练trackingnet的时候会出现KeyError: 'Nf1aqv5Fg5o_0',参照原始的pytracking issue156 似乎需要adding "Nf1aqv5Fg5o_0 airplane" line in trackingnet_classmap.txt file

    opened by kongbia 4
  • Performance on UAV123

    Performance on UAV123

    Hi! Thank for your great work!

    I used the transt.pth model and tested on the UAV123 dataset. I got 66.1 AUC score with pytracking toolkit which is lower than the given raw results(68.1 AUC score). Does this performance normal?

    Thank you for your time.

    opened by huanmx 3
  • 关于在ECA,CFA中用到的参数key_padding_mask

    关于在ECA,CFA中用到的参数key_padding_mask

    您好!在代码中看到您设置的attn_mask的值都为None,在encoder的ECA、CFA和decoder的CFA中都设置了key_padding_mask参数,在ECA中key_padding_mask与q值对应,在CFA中key_padding_mask与q值”相反“,想问一下您key_padding_mask这个参数起到了什么作用呢?

    opened by ANdong-star 3
  • Question in featurefusion_network.py

    Question in featurefusion_network.py

    Dear Chen:

    "hs = self.decoder(memory_search, memory_temp, tgt_key_padding_mask=mask_search, memory_key_padding_mask=mask_temp, pos_enc=pos_temp, pos_dec=pos_search)"

    I think that the order of input in decoder is firstly memory_temp( i think it is template branch) and secondly memory_search( i think it is search image branch) , because i read your paper I find that k,v of the decode CFA is from above branch and q is from below branch. Thank you and look forward to your reply.

    opened by LeungWaiHo 3
  • GOT-10k训练问题

    GOT-10k训练问题

    您好,我是一名深度学习的初学者,有一些关于TransT在仅使用GOT-10k训练的问题想请教。

    1. 请问您在仅使用GOT-10K时训练的参数是怎么设置的呢?epoch还是1k么?
    2. 得到TransT_GOT.pth时的训练数据是完整的GOT-10k训练集么?
    3. 在TransT/train_settings/transt/transt.py中进行如下修改时我的samples_per_epoch以及max_gap是否需要进行调整呢,还是就如下代码中设置即可? dataset_train = sampler.TransTSampler([got10k_train], [1], samples_per_epoch=1000 * settings.batch_size, max_gap=100, processing=data_processing_train)

    期待您的解答😀

    opened by Nightwatch-Fox11 2
  • Same Inference time in GPU and CPU

    Same Inference time in GPU and CPU

    I tried running the code in CPU and GPU from parameter setting. However, I am getting the same 11 FPS on both devices. CPU = Intel i7, 11th Gen and GPU = GTX 1050 Ti. May I know your suggestion on it? Thank you.

    opened by ghimireadarsh 0
  • TransT-GOT Train Setting?

    TransT-GOT Train Setting?

    I want to know the learning settings of TransT-Got.

    1. batch size
    2. samples per epoch
    3. epoch
    4. learning rate & weight decay

    And I want to know if I can use the 'sampler for training' code like this.

    The sampler for training

    dataset_train = sampler.TransTSampler([got10k_train], None, samples_per_epoch=1000*settings.batch_size, max_gap=100, processing=data_processing_train)

    opened by SkiddieAhn 1
  • 训练问题,训练进行几次后,会出错,出现nan值,导致AssertionError!!!

    训练问题,训练进行几次后,会出错,出现nan值,导致AssertionError!!!

    C:\Users\bxf\anaconda3\envs\transt\python.exe C:/PyCharmProjects/TransT-main/ltr/run_training.py Training: transt transt WARNING: You are using tensorboardX instead sis you have a too old pytorch version. loading annotations into memory... Done (t=13.20s) creating index... index created! number of params: 23016006 No matching checkpoint file found [train: 1, 1 / 1000] FPS: 0.0 (0.0) , Loss/total: 12.99988 , Loss/ce: 0.69430 , Loss/bbox: 0.97997 , Loss/giou: 1.15687 , iou: 0.03106 [train: 1, 2 / 1000] FPS: 0.0 (5.1) , Loss/total: 13.18990 , Loss/ce: 0.67882 , Loss/bbox: 1.01086 , Loss/giou: 1.23913 , iou: 0.01553 [train: 1, 3 / 1000] FPS: 0.0 (5.1) , Loss/total: 13.00681 , Loss/ce: 0.69773 , Loss/bbox: 0.93112 , Loss/giou: 1.26818 , iou: 0.01083 [train: 1, 4 / 1000] FPS: 0.0 (5.3) , Loss/total: 12.93164 , Loss/ce: 0.69913 , Loss/bbox: 0.91258 , Loss/giou: 1.27109 , iou: 0.01094 [train: 1, 5 / 1000] FPS: 0.0 (4.9) , Loss/total: 12.94410 , Loss/ce: 0.69589 , Loss/bbox: 0.91288 , Loss/giou: 1.29008 , iou: 0.00936 [train: 1, 6 / 1000] FPS: 0.0 (5.1) , Loss/total: 12.90344 , Loss/ce: 0.69371 , Loss/bbox: 0.90170 , Loss/giou: 1.30679 , iou: 0.00780 Training crashed at epoch 1 Traceback for the error! Traceback (most recent call last): File "C:\PyCharmProjects\TransT-main\ltr\trainers\base_trainer.py", line 70, in train self.train_epoch() # 调用ltr/trainers/ltr_trainer.py写的train_epoch方法 File "C:\PyCharmProjects\TransT-main\ltr\trainers\ltr_trainer.py", line 79, in train_epoch self.cycle_dataset(loader) # 调用自己写的cycle_dataset方法 File "C:\PyCharmProjects\TransT-main\ltr\trainers\ltr_trainer.py", line 60, in cycle_dataset loss, stats = self.actor(data) # 跳转到ltr/actors/tracking.py里面 File "C:\PyCharmProjects\TransT-main\ltr\actors\tracking.py", line 44, in call loss_dict = self.objective(outputs, targets) # 跳转到ltr/models/tracking/transt.py的182行的forward方法,用于计算损失 File "C:\Users\bxf\anaconda3\envs\transt\lib\site-packages\torch\nn\modules\module.py", line 550, in call result = self.forward(*input, **kwargs) File "C:\PyCharmProjects\TransT-main\ltr\models\tracking\transt.py", line 204, in forward losses.update(self.get_loss(loss, outputs, targets, indices, num_boxes_pos)) File "C:\PyCharmProjects\TransT-main\ltr\models\tracking\transt.py", line 180, in get_loss return loss_map[loss](outputs, targets, indices, num_boxes) File "C:\PyCharmProjects\TransT-main\ltr\models\tracking\transt.py", line 153, in loss_boxes box_ops.box_cxcywh_to_xyxy(target_boxes)) File "C:\PyCharmProjects\TransT-main\util\box_ops.py", line 52, in generalized_box_iou assert (boxes1[:, 2:] >= boxes1[:, :2]).all() AssertionError

    opened by xiaofengBian 0
  • why show “No matching checkpoint file found”

    why show “No matching checkpoint file found”

    when i run "run_training.py", it shows "No matching checkpoint file found" Restarting training from last epoch ... No matching checkpoint file found Training crashed at epoch 1 Traceback for the error! Traceback (most recent call last): File "/ai/lu/TransT-main/ltr/../ltr/trainers/base_trainer.py", line 70, in train self.train_epoch() File "/ai/lu/TransT-main/ltr/../ltr/trainers/ltr_trainer.py", line 79, in train_epoch self.cycle_dataset(loader) File "/ai/lu/TransT-main/ltr/../ltr/trainers/ltr_trainer.py", line 52, in cycle_dataset for i, data in enumerate(loader, 1): File "/root/anaconda3/envs/transt/lib/python3.7/site-packages/torch/utils/data/dataloader.py", line 345, in next data = self._next_data() File "/root/anaconda3/envs/transt/lib/python3.7/site-packages/torch/utils/data/dataloader.py", line 856, in _next_data return self._process_data(data) File "/root/anaconda3/envs/transt/lib/python3.7/site-packages/torch/utils/data/dataloader.py", line 881, in _process_data data.reraise() File "/root/anaconda3/envs/transt/lib/python3.7/site-packages/torch/_utils.py", line 395, in reraise raise self.exc_type(msg) ValueError: Caught ValueError in DataLoader worker process 0. Original Traceback (most recent call last): File "/root/anaconda3/envs/transt/lib/python3.7/site-packages/torch/utils/data/_utils/worker.py", line 178, in _worker_loop data = fetcher.fetch(index) File "/root/anaconda3/envs/transt/lib/python3.7/site-packages/torch/utils/data/_utils/fetch.py", line 44, in fetch data = [self.dataset[idx] for idx in possibly_batched_index] File "/root/anaconda3/envs/transt/lib/python3.7/site-packages/torch/utils/data/_utils/fetch.py", line 44, in data = [self.dataset[idx] for idx in possibly_batched_index] File "/ai/lu/TransT-main/ltr/../ltr/data/sampler.py", line 92, in getitem dataset = random.choices(self.datasets, self.p_datasets)[0] File "/root/anaconda3/envs/transt/lib/python3.7/random.py", line 361, in choices raise ValueError('The number of weights does not match the population')

    opened by Mango1218 12
Owner
chenxin
Master Student of Dalian University of Technology
chenxin
Code for the paper "Graph Attention Tracking". (CVPR2021)

SiamGAT 1. Environment setup This code has been tested on Ubuntu 16.04, Python 3.5, Pytorch 1.2.0, CUDA 9.0. Please install related libraries before r

null 122 Dec 24, 2022
Joint detection and tracking model named DEFT, or ``Detection Embeddings for Tracking.

DEFT: Detection Embeddings for Tracking DEFT: Detection Embeddings for Tracking, Mohamed Chaabane, Peter Zhang, J. Ross Beveridge, Stephen O'Hara

Mohamed Chaabane 253 Dec 18, 2022
Tracking code for the winner of track 1 in the MMP-Tracking Challenge at ICCV 2021 Workshop.

Tracking Code for the winner of track1 in MMP-Trakcing challenge This repository contains our tracking code for the Multi-camera Multiple People Track

DamoCV 29 Nov 13, 2022
Tracking Pipeline helps you to solve the tracking problem more easily

Tracking_Pipeline Tracking_Pipeline helps you to solve the tracking problem more easily I integrate detection algorithms like: Yolov5, Yolov4, YoloX,

VNOpenAI 32 Dec 21, 2022
Quadruped-command-tracking-controller - Quadruped command tracking controller (flat terrain)

Quadruped command tracking controller (flat terrain) Prepare Install RAISIM link

Yunho Kim 4 Oct 20, 2022
Python package for multiple object tracking research with focus on laboratory animals tracking.

motutils is a Python package for multiple object tracking research with focus on laboratory animals tracking. Features loads: MOTChallenge CSV, sleap

Matěj Šmíd 2 Sep 5, 2022
VSR-Transformer - This paper proposes a new Transformer for video super-resolution (called VSR-Transformer).

VSR-Transformer By Jiezhang Cao, Yawei Li, Kai Zhang, Luc Van Gool This paper proposes a new Transformer for video super-resolution (called VSR-Transf

Jiezhang Cao 225 Nov 13, 2022
Multiple-Object Tracking with Transformer

TransTrack: Multiple-Object Tracking with Transformer Introduction TransTrack: Multiple-Object Tracking with Transformer Models Training data Training

Peize Sun 537 Jan 4, 2023
git git《Transformer Meets Tracker: Exploiting Temporal Context for Robust Visual Tracking》(CVPR 2021) GitHub:git2] 《Masksembles for Uncertainty Estimation》(CVPR 2021) GitHub:git3]

Transformer Meets Tracker: Exploiting Temporal Context for Robust Visual Tracking Ning Wang, Wengang Zhou, Jie Wang, and Houqiang Li Accepted by CVPR

NingWang 236 Dec 22, 2022
Learning Spatio-Temporal Transformer for Visual Tracking

STARK The official implementation of the paper Learning Spatio-Temporal Transformer for Visual Tracking Hiring research interns for visual transformer

Multimedia Research 484 Dec 29, 2022
TrTr: Visual Tracking with Transformer

TrTr: Visual Tracking with Transformer We propose a novel tracker network based on a powerful attention mechanism called Transformer encoder-decoder a

趙 漠居(Zhao, Moju) 66 Dec 27, 2022
HiFT: Hierarchical Feature Transformer for Aerial Tracking (ICCV2021)

HiFT: Hierarchical Feature Transformer for Aerial Tracking Ziang Cao, Changhong Fu, Junjie Ye, Bowen Li, and Yiming Li Our paper is Accepted by ICCV 2

Intelligent Vision for Robotics in Complex Environment 55 Nov 23, 2022
This repository is an official implementation of the paper MOTR: End-to-End Multiple-Object Tracking with TRansformer.

MOTR: End-to-End Multiple-Object Tracking with TRansformer This repository is an official implementation of the paper MOTR: End-to-End Multiple-Object

null 348 Jan 7, 2023
This is the official code for the paper "Tracker Meets Night: A Transformer Enhancer for UAV Tracking".

SCT This is the official code for the paper "Tracker Meets Night: A Transformer Enhancer for UAV Tracking" The spatial-channel Transformer (SCT) enhan

Intelligent Vision for Robotics in Complex Environment 27 Nov 23, 2022
The official implementation of paper Siamese Transformer Pyramid Networks for Real-Time UAV Tracking, accepted by WACV22

SiamTPN Introduction This is the official implementation of the SiamTPN (WACV2022). The tracker intergrates pyramid feature network and transformer in

Robotics and Intelligent Systems Control @ NYUAD 28 Nov 25, 2022
(CVPR2021) ClassSR: A General Framework to Accelerate Super-Resolution Networks by Data Characteristic

ClassSR (CVPR2021) ClassSR: A General Framework to Accelerate Super-Resolution Networks by Data Characteristic Paper Authors: Xiangtao Kong, Hengyuan

Xiangtao Kong 308 Jan 5, 2023
[CVPR2021 Oral] UP-DETR: Unsupervised Pre-training for Object Detection with Transformers

UP-DETR: Unsupervised Pre-training for Object Detection with Transformers This is the official PyTorch implementation and models for UP-DETR paper: @a

dddzg 430 Dec 23, 2022
Code for our CVPR2021 paper coordinate attention

Coordinate Attention for Efficient Mobile Network Design (preprint) This repository is a PyTorch implementation of our coordinate attention (will appe

Qibin (Andrew) Hou 726 Jan 5, 2023
[CVPR2021] The source code for our paper 《Removing the Background by Adding the Background: Towards Background Robust Self-supervised Video Representation Learning》.

TBE The source code for our paper "Removing the Background by Adding the Background: Towards Background Robust Self-supervised Video Representation Le

Jinpeng Wang 150 Dec 28, 2022