Implementation of the paper All Labels Are Not Created Equal: Enhancing Semi-supervision via Label Grouping and Co-training

Related tags

Deep Learningsemco
Overview

SemCo

The official pytorch implementation of the paper All Labels Are Not Created Equal: Enhancing Semi-supervision via Label Grouping and Co-training (appearing in CVPR2021)

SemCo Conceptual Diagram

Install Dependencies

  • Create a new environment and install dependencies using pip install -r requirements.txt
  • Install apex to enable automatic mixed precision training (AMP).
git clone https://github.com/NVIDIA/apex
cd apex
python setup.py install --cpp_ext --cuda_ext

Note: Installing apex is optional, if you don't want to implement amp, you can simply pass --no_amp command line argument to the launcher.

Dataset

We use a standard directory structure for all our datasets to enable running the code on any dataset of choice without the need to edit the dataloaders. The datasets directory follow the below structure (only shown for cifar100 but is the same for all other datasets):

datasets
└───cifar100
   └───train
       │   <image1>
       │   <image2>
       │   ...
   └───test
       │   <image1-test>
       │   <image2-test>
       │   ...
   └───labels
       │   labels_train.feather
       │   labels_test.feather

An example of the above directory structure for cifar100 can be found here.

To preprocess a generic dataset into the above format, you can refer to utils/utils.py for several examples.

To configure the datasets directory path, you can either set the environment variable SEMCO_DATA_PATH or pass a command line argument --dataset-path to the launcher. (e.g. export SEMCO_DATA_PATH=/home/data). Note that this path references the parent datasets directory which contains the different sub directories for the individual datasets (e.g. cifar100, mini-imagenet, etc.)

Label Semantics Embeddings

SemCo expects a prior representation of all class labels via a semantic embedding for each class name. In our experiments, we use embeddings obtained from ConceptNet knowledge graph which contains a total of ~550K term embeddings. SemCo uses a matching criteria to find the best embedding for each of the class labels. Alternatively, you can use class attributes as the prior (like we did for CUB200 dataset), so you can build your own semantic dictionary.

To run experiments, please download the semantic embedding file here and set the path to the downloaded file either via SEMCO_WV_PATH environment variable or --word-vec-path command line argument. (e.g. export SEMCO_WV_PATH=/home/inas0003/data/numberbatch-en-19.08_128D.dict.pkl

Defining the Splits

For each of the experiments, you will need to specify to the launcher 4 command line arguments:

  • --dataset-name: denoting the dataset directory name (e.g. cifar100)
  • --train-split-pickle: path to pickle file with training split
  • --valid-split-pickle: (optional) path to pickle file with validation/test split (by default contains all the files in the test folder)
  • --classes-pickle: (optional) path to pickle file with list of class names

To obtain the three pickle files for any dataset, you can use generate_tst_pkls.py script specifying the dataset name and the number of instances per label and optionally a random seed. Example as follows:

python generate_tst_pkls.py --dataset-name cifar100 --instances-per-label 10 --random-seed 000 --output-path splits

The above will generate a train split with 10 images per class using a random seed of 000 together with the class names and the validation split containing all the files placed in the test folder. This can be tweaked by editing the python script.

Training the model

To train the model on cifar100 with 40 labeled samples, you can run the script:

    $ python launch_semco.py --dataset-name cifar100 --train-split-pickle splits/cifar100_labelled_data_40_seed123.pkl --model_backbone=wres --wres-k=2

or without amp

    $ python launch_semco.py --dataset-name cifar100 --train-split-pickle splits/cifar100_labelled_data_40_seed123.pkl --model_backbone=wres --wres-k=2 --no_amp

Similary to train the model on mini_imagenet with 400 labeled samples, you can run the script:

    $  python launch_semco.py --dataset-name mini_imagenet --train-split-pickle testing/mini_imagenet_labelled_data_40_seed456.pkl --model_backbone=resnet18 --im-size=84 --cropsize=84 
Animation of solving the traveling salesman problem to optimality using mixed-integer programming and iteratively eliminating sub tours

tsp-streamlit Animation of solving the traveling salesman problem to optimality using mixed-integer programming and iteratively eliminating sub tours.

4 Nov 05, 2022
Implementation of "Deep Implicit Templates for 3D Shape Representation"

Deep Implicit Templates for 3D Shape Representation Zerong Zheng, Tao Yu, Qionghai Dai, Yebin Liu. arXiv 2020. This repository is an implementation fo

Zerong Zheng 144 Dec 07, 2022
Boosted neural network for tabular data

XBNet - Xtremely Boosted Network Boosted neural network for tabular data XBNet is an open source project which is built with PyTorch which tries to co

Tushar Sarkar 175 Jan 04, 2023
EssentialMC2 Video Understanding

EssentialMC2 Introduction EssentialMC2 is a complete system to solve video understanding tasks including MHRL(representation learning), MECR2( relatio

Alibaba 106 Dec 11, 2022
🤗 Push your spaCy pipelines to the Hugging Face Hub

spacy-huggingface-hub: Push your spaCy pipelines to the Hugging Face Hub This package provides a CLI command for uploading any trained spaCy pipeline

Explosion 30 Oct 09, 2022
This repository is an implementation of our NeurIPS 2021 paper (Stylized Dialogue Generation with Multi-Pass Dual Learning) in PyTorch.

MPDL---TODO This repository is an implementation of our NeurIPS 2021 paper (Stylized Dialogue Generation with Multi-Pass Dual Learning) in PyTorch. Ci

CodebaseLi 3 Nov 27, 2022
Differentiable Prompt Makes Pre-trained Language Models Better Few-shot Learners

DART Implementation for ICLR2022 paper Differentiable Prompt Makes Pre-trained Language Models Better Few-shot Learners. Environment

ZJUNLP 83 Dec 27, 2022
Simple node deletion tool for onnx.

snd4onnx Simple node deletion tool for onnx. I only test very miscellaneous and limited patterns as a hobby. There are probably a large number of bugs

Katsuya Hyodo 6 May 15, 2022
A repository for benchmarking neural vocoders by their quality and speed.

License The majority of VocBench is licensed under CC-BY-NC, however portions of the project are available under separate license terms: Wavenet, Para

Meta Research 177 Dec 12, 2022
[NeurIPS 2021] “Improving Contrastive Learning on Imbalanced Data via Open-World Sampling”,

Improving Contrastive Learning on Imbalanced Data via Open-World Sampling Introduction Contrastive learning approaches have achieved great success in

VITA 24 Dec 17, 2022
AI Summer's complete catalog of articles

Learn Deep Learning with AI Summer A collection of all articles (almost 100) written for the AI Summer blog organized by topic. Deep Learning Theory M

AI Summer 95 Dec 29, 2022
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

g-parki 7 Jul 15, 2022
NuPIC Studio is an all­-in-­one tool that allows users create a HTM neural network from scratch

NuPIC Studio is an all­-in-­one tool that allows users create a HTM neural network from scratch, train it, collect statistics, and share it among the members of the community. It is not just a visual

HTM Community 93 Sep 30, 2022
The AWS Certified SysOps Administrator

The AWS Certified SysOps Administrator – Associate (SOA-C02) exam is intended for system administrators in a cloud operations role who have at least 1 year of hands-on experience with deployment, man

Aiden Pearce 32 Dec 11, 2022
A parametric soroban written with CADQuery.

A parametric soroban written in CADQuery The purpose of this project is to demonstrate how "code CAD" can be intuitive to learn. See soroban.py for a

Lee 4 Aug 13, 2022
RepVGG: Making VGG-style ConvNets Great Again

This repository is the code that needs to be submitted for OpenMMLab Algorithm Ecological Challenge,the paper is RepVGG: Making VGG-style ConvNets Great Again

Ty Feng 62 May 21, 2022
The project of phase's key role in complex and real NN

Phase-in-NN This is the code for our project at Princeton (co-authors: Yuqi Nie, Hui Yuan). The paper title is: "Neural Network is heterogeneous: Phas

YuqiNie-lab 1 Nov 04, 2021
CondenseNet V2: Sparse Feature Reactivation for Deep Networks

CondenseNetV2 This repository is the official Pytorch implementation for "CondenseNet V2: Sparse Feature Reactivation for Deep Networks" paper by Le Y

Haojun Jiang 74 Dec 12, 2022
HistoKT: Cross Knowledge Transfer in Computational Pathology

HistoKT: Cross Knowledge Transfer in Computational Pathology Exciting News! HistoKT has been accepted to ICASSP 2022. HistoKT: Cross Knowledge Transfe

Mahdi S. Hosseini 5 Jan 05, 2023
Tensors and Dynamic neural networks in Python with strong GPU acceleration

PyTorch is a Python package that provides two high-level features: Tensor computation (like NumPy) with strong GPU acceleration Deep neural networks b

61.4k Jan 04, 2023