The codes and related files to reproduce the results for Image Similarity Challenge Track 1.

Overview

ISC-Track1-Submission

The codes and related files to reproduce the results for Image Similarity Challenge Track 1.

Required dependencies

To begin with, you should install the following packages with the specified versions in Python, Anaconda. Other versions may work but please do NOT try. For instance, cuda 11.0 has some bugs which bring very bad results. The hardware chosen is Nvidia Tesla V100 and Intel CPU. Other hardware, such as A100, may work but please do NOT try. The stability is not guaranteed, for instance, the Ampere architecture is not suitable and some instability is observed. Please do NOT use AMD CPU, such as EPYC, we observe some instability on DGX server.

  • python 3.7.10
  • pytorch 1.7.1 with cuda 10.1
  • faiss-gpu 1.7.1 with cuda 10.1
  • h5py 3.4.0
  • pandas 1.3.3
  • sklearn 1.0
  • skimage 0.18.3
  • PIL 8.3.2
  • cv2 4.5.3.56
  • numpy 1.16.0
  • torchvision 0.8.2 with cuda 10.1
  • augly 0.1.4
  • selectivesearch 0.4
  • face-recognition 1.3.0 (with dlib of gpu-version)
  • tqdm 4.62.3
  • requests 2.26.0
  • seaborn 0.11.2
  • mkl 2.4.0
  • loguru 0.5.3

Note: Some unimportant packages may be missing, please install them using pip directly when an error occurs.

Pre-trained models

We use three pre-trained models. They are all pre-trained on ImageNet unsupervisedly. To be convenient, we first directly give the pre-trained models as follows, then also the training codes are given.

The first backbone: ResNet-50; The second backbone: ResNet-152; The third backbone: ResNet-50-IBN.

For ResNet-50, we do not pre-train it by ourselves. It is directly downloaded from here. It is supplied by Facebook Research, and the project is Barlow Twins. You should rename it to resnet50_bar.pth.

For ResNet-152 and ResNet-50-IBN, we use the official codes of Momentum2-teacher. We only change the backbone to ResNet-152 and ResNet-50-IBN. It takes about 2 weeks to pre-train the ResNet-152, and 1 week to pre-train the ResNet-50-IBN on 8 V100 GPUs. To be convenient, we supply the whole pre-training codes in the Pretrain folder. The related readme file is also given in that folder.

It should be noted that pre-training processing plays a very important role in our algorithm. Therefore, if you want to reproduce the pre-trained results, please do NOT change the number of GPUs, the batch size, and other related hyper-parameters.

Training

For training, we generate 11 datasets. For each dataset, 3 models with different backbones are trained. Each training takes about/less than 1 day on 4 V100 GPUs (bigger backbone takes longer and smaller backbone takes shorter). The whole training codes, including how to generate training datasets and the link to the generated datasets, are given in the Training folder. For more details, please refer to the readme file in that folder.

Test

To test the performance of the trained model, we perform multi-scale, multi-model, and multi-part testing and ensemble all the scores to get the final score. To be efficient, 33 V100 GPUs are suggested to use. The time for extracting all query images' features using 33 V100 GPUs is about 3 hours. Also extracting and storing training and reference images' features take a lot of time. Please be patient and prepare enough storage to reproduce the testing process. We give all the information to generate our final results in the Test folder. Please reproduce the results according to the readme file in that folder.

Owner
Wenhao Wang
I am a student from Beihang University. My research interests include person re-identification, unsupervised domain adaptation, and domain generalization.
Wenhao Wang
PyTorch implementation of Glow

glow-pytorch PyTorch implementation of Glow, Generative Flow with Invertible 1x1 Convolutions (https://arxiv.org/abs/1807.03039) Usage: python train.p

Kim Seonghyeon 433 Dec 27, 2022
Adjusting for Autocorrelated Errors in Neural Networks for Time Series

Adjusting for Autocorrelated Errors in Neural Networks for Time Series This repository is the official implementation of the paper "Adjusting for Auto

Fan-Keng Sun 51 Nov 05, 2022
Learning to Reconstruct 3D Non-Cuboid Room Layout from a Single RGB Image

NonCuboidRoom Paper Learning to Reconstruct 3D Non-Cuboid Room Layout from a Single RGB Image Cheng Yang*, Jia Zheng*, Xili Dai, Rui Tang, Yi Ma, Xiao

67 Dec 15, 2022
PyTorch Implementation of Region Similarity Representation Learning (ReSim)

ReSim This repository provides the PyTorch implementation of Region Similarity Representation Learning (ReSim) described in this paper: @Article{xiao2

Tete Xiao 74 Jan 03, 2023
This is a re-implementation of TransGAN: Two Pure Transformers Can Make One Strong GAN (CVPR 2021) in PyTorch.

TransGAN: Two Transformers Can Make One Strong GAN [YouTube Video] Paper Authors: Yifan Jiang, Shiyu Chang, Zhangyang Wang CVPR 2021 This is re-implem

Ahmet Sarigun 79 Jan 05, 2023
This code is an implementation for Singing TTS.

MLP Singer This code is an implementation for Singing TTS. The algorithm is based on the following papers: Tae, J., Kim, H., & Lee, Y. (2021). MLP Sin

Heejo You 22 Dec 23, 2022
All materials of Cassandra Event, Udyam'22

Cassandra 2022 Workspace Workshop Materials Workshop-1 Workshop-2 Workshop-3 Workshop-4 Assignments Assignment-1 Assignment-2 Assignment-3 Resources P

36 Dec 31, 2022
Randstad Artificial Intelligence Challenge (powered by VGEN). Soluzione proposta da Stefano Fiorucci (anakin87) - primo classificato

Randstad Artificial Intelligence Challenge (powered by VGEN) Soluzione proposta da Stefano Fiorucci (anakin87) - primo classificato Struttura director

Stefano Fiorucci 1 Nov 13, 2021
[NeurIPS'20] Multiscale Deep Equilibrium Models

Multiscale Deep Equilibrium Models 💥 💥 💥 💥 This repo is deprecated and we will soon stop actively maintaining it, as a more up-to-date (and simple

CMU Locus Lab 221 Dec 26, 2022
CLIPImageClassifier wraps clip image model from transformers

CLIPImageClassifier CLIPImageClassifier wraps clip image model from transformers. CLIPImageClassifier is initialized with the argument classes, these

Jina AI 6 Sep 12, 2022
This repository is the offical Pytorch implementation of ContextPose: Context Modeling in 3D Human Pose Estimation: A Unified Perspective (CVPR 2021).

Context Modeling in 3D Human Pose Estimation: A Unified Perspective (CVPR 2021) Introduction This repository is the offical Pytorch implementation of

37 Nov 21, 2022
Orchestrating Distributed Materials Acceleration Platform Tutorial

Orchestrating Distributed Materials Acceleration Platform Tutorial This tutorial for orchestrating distributed materials acceleration platform was pre

BIG-MAP 1 Jan 25, 2022
This is the official PyTorch implementation of the paper "TransFG: A Transformer Architecture for Fine-grained Recognition" (Ju He, Jie-Neng Chen, Shuai Liu, Adam Kortylewski, Cheng Yang, Yutong Bai, Changhu Wang, Alan Yuille).

TransFG: A Transformer Architecture for Fine-grained Recognition Official PyTorch code for the paper: TransFG: A Transformer Architecture for Fine-gra

Ju He 307 Jan 03, 2023
A Pytorch implementation of the multi agent deep deterministic policy gradients (MADDPG) algorithm

Multi-Agent-Deep-Deterministic-Policy-Gradients A Pytorch implementation of the multi agent deep deterministic policy gradients(MADDPG) algorithm This

Phil Tabor 159 Dec 28, 2022
Code for SyncTwin: Treatment Effect Estimation with Longitudinal Outcomes (NeurIPS 2021)

SyncTwin: Treatment Effect Estimation with Longitudinal Outcomes (NeurIPS 2021) SyncTwin is a treatment effect estimation method tailored for observat

Zhaozhi Qian 3 Nov 03, 2022
Industrial Image Anomaly Localization Based on Gaussian Clustering of Pre-trained Feature

Industrial Image Anomaly Localization Based on Gaussian Clustering of Pre-trained Feature Q. Wan, L. Gao, X. Li and L. Wen, "Industrial Image Anomaly

smiler 6 Dec 25, 2022
Unofficial pytorch implementation of paper "One-Shot Free-View Neural Talking-Head Synthesis for Video Conferencing"

One-Shot Free-View Neural Talking Head Synthesis Unofficial pytorch implementation of paper "One-Shot Free-View Neural Talking-Head Synthesis for Vide

ZLH 406 Dec 23, 2022
SE-MSCNN: A Lightweight Multi-scaled Fusion Network for Sleep Apnea Detection Using Single-Lead ECG Signals

SE-MSCNN: A Lightweight Multi-scaled Fusion Network for Sleep Apnea Detection Using Single-Lead ECG Signals Abstract Sleep apnea (SA) is a common slee

9 Dec 21, 2022
Experiments for Neural Flows paper

Neural Flows: Efficient Alternative to Neural ODEs [arxiv] TL;DR: We directly model the neural ODE solutions with neural flows, which is much faster a

54 Dec 07, 2022
Fast algorithms to compute an approximation of the minimal volume oriented bounding box of a point cloud in 3D.

ApproxMVBB Status Build UnitTests Homepage Fast algorithms to compute an approximation of the minimal volume oriented bounding box of a point cloud in

Gabriel Nützi 390 Dec 31, 2022