Official implementation of our paper "Learning to Bootstrap for Combating Label Noise"

Related tags

Deep Learning L2B
Overview

Learning to Bootstrap for Combating Label Noise

This repo is the official implementation of our paper "Learning to Bootstrap for Combating Label Noise".

Citation

If you use this code for your research, please cite our paper "Learning to Bootstrap for Combating Label Noise".

@misc{zhou2022learning,
      title={Learning to Bootstrap for Combating Label Noise}, 
      author={Yuyin Zhou and Xianhang Li and Fengze Liu and Xuxi Chen and Lequan Yu and Cihang Xie and Matthew P. Lungren and Lei Xing},
      year={2022},
      eprint={2202.04291},
      archivePrefix={arXiv},
      primaryClass={cs.CV}
}

Requirements

Python >= 3.6.4
Pytorch >= 1.6.0
Higher = 0.2.1
Tensorboardx = 2.4.1

Training

First, please create a folder to store checkpoints by using the following command.

mkdir checkpoint

CIFAR-10

To reproduce the results on CIFAR dataset from our paper, please follow the command and our hyper-parameters.

First, you can adjust the corruption_prob and corruption_type to obtain different noise rates and noise type.

Second, the reweight_label indicates you are using the our L2B method. You can change it to baseline or mixup.

python  main.py  --arch res18 --dataset cifar10 --num_classes 10 --exp L2B --train_batch_size  512 \
 --corruption_prob 0.2 --reweight_label  --lr 0.15  -clipping_norm 0.25  --num_epochs 300  --scheduler cos \
 --corruption_type unif  --warm_up 10  --seed 0  

CIFAR-100

Most of settings are the same as CIFAR-10. To reproduce the results, please follow the command.

python  main.py  --arch res18 --dataset cifar100 --num_classes 100 --exp L2B --train_batch_size  256  \
--corruption_prob 0.2 --reweight_label  --lr 0.15  --clipping_norm 0.80  --num_epochs 300  --scheduler cos \
--corruption_type unif  --warm_up 10  --seed 0 \ 

ISIC2019

On the ISIC dataset, first you should download the dataset by following command.

Download ISIC dataset as follows:
wget https://isic-challenge-data.s3.amazonaws.com/2019/ISIC_2019_Training_Input.zip
wget https://isic-challenge-data.s3.amazonaws.com/2019/ISIC_2019_Training_GroundTruth.csv \

Then you can reproduce the results by following the command.

python main.py  --arch res50  --dataset ISIC --data_path isic_data/ISIC_2019_Training_Input --num_classes 8 
--exp L2B  --train_batch_size 64  --corruption_prob 0.2 --lr 0.01 --clipping_norm 0.80 --num_epochs 30 
--temperature 10.0  --wd 5e-4  --scheduler cos --reweight_label --norm_type softmax --warm_up 1 

Clothing-1M

First, the num_batch and train_batch_size indicates how many training images you want to use (we sample a balanced training data for each epoch).

Second, you can adjust the num_meta to sample different numbers of validation images to form the metaset. We use the whole validation set as metaset by default.

The data_path is where you store the data and key-label lists. And also change the data_path in the line 20 of main.py. If you have issue for downloading the dataset, please feel free to contact us.

Then you can reproduce the results by following the command.

python main.py --arch res18_224 --num_batch 250 --dataset clothing1m \
--exp L2B_clothing1m_one_stage_multi_runs  --train_batch_size 256  --lr 0.005  \
--num_epochs 300  --reweight_label  --wd 5e-4 --scheduler cos   --warm_up 0 \
--data_path /data1/data/clothing1m/clothing1M  --norm_type org  --num_classes 14 \ 
--multi_runs 3 --num_meta 14313

Contact

Yuyin Zhou

Xianhang Li

If you have any question about the code and data, please contact us directly.

You might also like...
This is the official pytorch implementation for our ICCV 2021 paper
This is the official pytorch implementation for our ICCV 2021 paper "TRAR: Routing the Attention Spans in Transformers for Visual Question Answering" on VQA Task

🌈 ERASOR (RA-L'21 with ICRA Option) Official page of "ERASOR: Egocentric Ratio of Pseudo Occupancy-based Dynamic Object Removal for Static 3D Point C

Convolutional neural network web app trained to track our infant’s sleep schedule using our Google Nest camera.
Convolutional neural network web app trained to track our infant’s sleep schedule using our Google Nest camera.

Machine Learning Sleep Schedule Tracker What is it? Convolutional neural network web app trained to track our infant’s sleep schedule using our Google

This is the official code of our paper
This is the official code of our paper "Diversity-based Trajectory and Goal Selection with Hindsight Experience Relay" (PRICAI 2021)

Diversity-based Trajectory and Goal Selection with Hindsight Experience Replay This is the official implementation of our paper "Diversity-based Traje

Official code for our ICCV paper: "From Continuity to Editability: Inverting GANs with Consecutive Images"

GANInversion_with_ConsecutiveImgs Official code for our ICCV paper: "From Continuity to Editability: Inverting GANs with Consecutive Images" https://a

Official PyTorch implemention of our paper "Learning to Rectify for Robust Learning with Noisy Labels".

WarPI The official PyTorch implemention of our paper "Learning to Rectify for Robust Learning with Noisy Labels". Run python main.py --corruption_type

The official codes of our CVPR2022 paper: A Differentiable Two-stage Alignment Scheme for Burst Image Reconstruction with Large Shift
The official codes of our CVPR2022 paper: A Differentiable Two-stage Alignment Scheme for Burst Image Reconstruction with Large Shift

TwoStageAlign The official codes of our CVPR2022 paper: A Differentiable Two-stage Alignment Scheme for Burst Image Reconstruction with Large Shift Pa

Official code for our CVPR '22 paper
Official code for our CVPR '22 paper "Dataset Distillation by Matching Training Trajectories"

Dataset Distillation by Matching Training Trajectories Project Page | Paper This repo contains code for training expert trajectories and distilling sy

Implementation of our paper 'RESA: Recurrent Feature-Shift Aggregator for Lane Detection' in AAAI2021.
Implementation of our paper 'RESA: Recurrent Feature-Shift Aggregator for Lane Detection' in AAAI2021.

RESA PyTorch implementation of the paper "RESA: Recurrent Feature-Shift Aggregator for Lane Detection". Our paper has been accepted by AAAI2021. Intro

Comments
  • The relationship between the KD and L2B

    The relationship between the KD and L2B

    Thanks for your idea of label reweighting. I am curious about the theoretical foundation. The loss designed contain the label loss and pseudo ones. The latter one seemingly plays a role of teacher model in the knowledge distilling and teach the current batch to train. I think there is a sub-field of KD related, self-teaching.

    Moreover, the alpha and beta are both updated during training, it is new in KD where the weight is controlled by a constant or temperature.

    opened by JiaxiangBU 1
Owner
null
The project is an official implementation of our CVPR2019 paper "Deep High-Resolution Representation Learning for Human Pose Estimation"

Deep High-Resolution Representation Learning for Human Pose Estimation (CVPR 2019) News [2020/07/05] A very nice blog from Towards Data Science introd

Leo Xiao 3.9k Jan 5, 2023
Official implementation of our paper "LLA: Loss-aware Label Assignment for Dense Pedestrian Detection" in Pytorch.

LLA: Loss-aware Label Assignment for Dense Pedestrian Detection This project provides an implementation for "LLA: Loss-aware Label Assignment for Dens

null 35 Dec 6, 2022
This project is the official implementation of our accepted ICLR 2021 paper BiPointNet: Binary Neural Network for Point Clouds.

BiPointNet: Binary Neural Network for Point Clouds Created by Haotong Qin, Zhongang Cai, Mingyuan Zhang, Yifu Ding, Haiyu Zhao, Shuai Yi, Xianglong Li

Haotong Qin 59 Dec 17, 2022
Official implementation of our CVPR2021 paper "OTA: Optimal Transport Assignment for Object Detection" in Pytorch.

OTA: Optimal Transport Assignment for Object Detection This project provides an implementation for our CVPR2021 paper "OTA: Optimal Transport Assignme

null 217 Jan 3, 2023
This is an official implementation of our CVPR 2021 paper "Bottom-Up Human Pose Estimation Via Disentangled Keypoint Regression" (https://arxiv.org/abs/2104.02300)

Bottom-Up Human Pose Estimation Via Disentangled Keypoint Regression Introduction In this paper, we are interested in the bottom-up paradigm of estima

HRNet 367 Dec 27, 2022
The official pytorch implementation of our paper "Is Space-Time Attention All You Need for Video Understanding?"

TimeSformer This is an official pytorch implementation of Is Space-Time Attention All You Need for Video Understanding?. In this repository, we provid

Facebook Research 1k Dec 31, 2022
Official implementation of GraphMask as presented in our paper Interpreting Graph Neural Networks for NLP With Differentiable Edge Masking.

GraphMask This repository contains an implementation of GraphMask, the interpretability technique for graph neural networks presented in our ICLR 2021

Michael Schlichtkrull 29 Sep 2, 2022
The official implementation of our CVPR 2021 paper - Hybrid Rotation Averaging: A Fast and Robust Rotation Averaging Approach

Graph Optimizer This repo contains the official implementation of our CVPR 2021 paper - Hybrid Rotation Averaging: A Fast and Robust Rotation Averagin

Chenyu 109 Dec 23, 2022
The repository offers the official implementation of our paper in PyTorch.

Cloth Interactive Transformer (CIT) Cloth Interactive Transformer for Virtual Try-On Bin Ren1, Hao Tang1, Fanyang Meng2, Runwei Ding3, Ling Shao4, Phi

Bingoren 49 Dec 1, 2022
The project is an official implementation of our paper "3D Human Pose Estimation with Spatial and Temporal Transformers".

3D Human Pose Estimation with Spatial and Temporal Transformers This repo is the official implementation for 3D Human Pose Estimation with Spatial and

Ce Zheng 363 Dec 28, 2022