EASY - Ensemble Augmented-Shot Y-shaped Learning: State-Of-The-Art Few-Shot Classification with Simple Ingredients.

Related tags

Deep Learningeasy
Overview

EASY - Ensemble Augmented-Shot Y-shaped Learning: State-Of-The-Art Few-Shot Classification with Simple Ingredients.

This repository is the official implementation of EASY - Ensemble Augmented-Shot Y-shaped Learning: State-Of-The-Art Few-Shot Classification with Simple Ingredients.

EASY proposes a simple methodology, that reaches or even beats state of the art performance on multiple standardized benchmarks of the field, while adding almost no hyperparameters or parameters to those used for training the initial deep learning models on the generic dataset.

Downloads

Please click the Google Drive link for downloading the features, backbones and datasets.

Each of the files (backbones and features) have the following prefixes depending on the backbone:

Backbone prefix Number of parameters
ResNet12 12M
ResNet12(1/sqrt(2)) small 6M
ResNet12(1/2) tiny 3M

Each of the features file is named as follow :

  • if not AS : " features .pt11"
  • if AS : " featuresAS .pt11"

Testing scripts for EASY

Run scripts to evaluate the features on FSL tasks for Y and ASY. For EY and EASY use the corresponding features.

Inductive setup using NCM

Test features on miniimagenet using Y (Resnet12)

" --dataset miniimagenet --model resnet12 --test-features ' /minifeatures1.pt11' --preprocessing ME"> 
$ python main.py --dataset-path "
     
      " --dataset miniimagenet --model resnet12 --test-features '
      
       /minifeatures1.pt11' --preprocessing ME

Test features on miniimagenet using ASY (Resnet12)

" --dataset miniimagenet --model resnet12 --test-features ' /minifeaturesAS1.pt11' --preprocessing ME"> 
$ python main.py --dataset-path "
     
      " --dataset miniimagenet --model resnet12 --test-features '
      
       /minifeaturesAS1.pt11' --preprocessing ME

Test features on miniimagenet using EY (3xResNet12)

" --dataset miniimagenet --model resnet12 --test-features "[ /minifeatures1.pt11, /minifeatures2.pt11, /minifeatures3.pt11]" --preprocessing ME"> 
$ python main.py --dataset-path "
       
        " --dataset miniimagenet --model resnet12 --test-features "[
        
         /minifeatures1.pt11, 
         
          /minifeatures2.pt11, 
          
           /minifeatures3.pt11]" --preprocessing ME

Test features on miniimagenet using EASY (3xResNet12)

" --dataset miniimagenet --model resnet12 --test-features "[ /minifeaturesAS1.pt11, /minifeaturesAS2.pt11, /minifeaturesAS3.pt11]" --preprocessing ME "> 
$ python main.py --dataset-path "
       
        " --dataset miniimagenet --model resnet12 --test-features "[
        
         /minifeaturesAS1.pt11, 
         
          /minifeaturesAS2.pt11, 
          
           /minifeaturesAS3.pt11]" --preprocessing ME

Transductive setup using Soft k-means

Test features on miniimagenet using Y (ResNet12)

" --dataset miniimagenet --model resnet12 --test-features ' /minifeatures1.pt11'--postprocessing ME --transductive --transductive-softkmeans --transductive-temperature-softkmeans 20"> 
$ python main.py --dataset-path "
     
      " --dataset miniimagenet --model resnet12 --test-features '
      
       /minifeatures1.pt11'--postprocessing ME --transductive --transductive-softkmeans --transductive-temperature-softkmeans 20

Test features on miniimagenet using ASY (ResNet12)

" --dataset miniimagenet --model resnet12 --test-features ' /minifeaturesAS1.pt11' --postprocessing ME --transductive --transductive-softkmeans --transductive-temperature-softkmeans 20"> 
$ python main.py --dataset-path "
     
      " --dataset miniimagenet --model resnet12 --test-features '
      
       /minifeaturesAS1.pt11' --postprocessing ME --transductive --transductive-softkmeans --transductive-temperature-softkmeans 20

Test features on miniimagenet using EY (3xResNet12)

" --dataset miniimagenet --model resnet12 --test-features "[ /minifeatures1.pt11, /minifeatures2.pt11, /minifeatures3.pt11]" --postrocessing ME --transductive --transductive-softkmeans --transductive-temperature-softkmeans 20"> 
$ python main.py --dataset-path "
       
        " --dataset miniimagenet --model resnet12 --test-features "[
        
         /minifeatures1.pt11, 
         
          /minifeatures2.pt11, 
          
           /minifeatures3.pt11]" --postrocessing ME  --transductive --transductive-softkmeans --transductive-temperature-softkmeans 20

Test features on miniimagenet using EASY (3xResNet12)

" --dataset miniimagenet --model resnet12 --test-features "[ /minifeaturesAS1.pt11, /minifeaturesAS2.pt11, /minifeaturesAS3.pt11]" --postrocessing ME --transductive --transductive-softkmeans --transductive-temperature-softkmeans 20"> 
$ python main.py --dataset-path "
       
        " --dataset miniimagenet --model resnet12 --test-features "[
        
         /minifeaturesAS1.pt11, 
         
          /minifeaturesAS2.pt11, 
          
           /minifeaturesAS3.pt11]" --postrocessing ME  --transductive --transductive-softkmeans --transductive-temperature-softkmeans 20

Training scripts for Y

Train a model on miniimagenet using manifold mixup, self-supervision and cosine scheduler

" --dataset miniimagenet --model resnet12 --epochs 0 --manifold-mixup 500 --rotations --cosine --gamma 0.9 --milestones 100 --batch-size 128 --preprocessing ME "> 
$ python main.py --dataset-path "
    
     " --dataset miniimagenet --model resnet12 --epochs 0 --manifold-mixup 500 --rotations --cosine --gamma 0.9 --milestones 100 --batch-size 128 --preprocessing ME

Important Arguments

Some important arguments for our code.

Training arguments

  • dataset: choices=['miniimagenet', 'cubfs','tieredimagenet', 'fc100', 'cifarfs']
  • model: choices=['resnet12', 'resnet18', 'resnet20', 'wideresnet', 's2m2r']
  • dataset-path: path of the datasets folder which contains folders of all the datasets.

Few-shot Classification

  • preprocessing: preprocessing sequence for few shot given as a string, can contain R:relu P:sqrt E:sphering and M:centering using the base data.
  • postprocessing: postprocessing sequence for few shot given as a string, can contain R:relu P:sqrt E:sphering and M:centering on the few-shot data, used for transductive setting.

Few-shot classification Results

Experimental results on few-shot learning datasets with ResNet-12 backbone. We report our average results with 10000 randomly sampled episodes for both 1-shot and 5-shot evaluations.

MiniImageNet Dataset (inductive)

Methods 1-Shot 5-Way 5-Shot 5-Way
SimpleShot [29] 62.85 ± 0.20 80.02 ± 0.14
Baseline++ [30] 53.97 ± 0.79 75.90 ± 0.61
TADAM [35] 58.50 ± 0.30 76.70 ± 0.30
ProtoNet [10] 60.37 ± 0.83 78.02 ± 0.57
R2-D2 (+ens) [20] 64.79 ± 0.45 81.08 ± 0.32
FEAT [36] 66.78 82.05
CNL [37] 67.96 ± 0.98 83.36 ± 0.51
MERL [38] 67.40 ± 0.43 83.40 ± 0.28
Deep EMD v2 [13] 68.77 ± 0.29 84.13 ± 0.53
PAL [8] 69.37 ± 0.64 84.40 ± 0.44
inv-equ [39] 67.28 ± 0.80 84.78 ± 0.50
CSEI [40] 68.94 ± 0.28 85.07 ± 0.50
COSOC [9] 69.28 ± 0.49 85.16 ± 0.42
EASY 2×ResNet12 1/√2 (ours) 70.63 ± 0.20 86.28 ± 0.12
above <=12M nb of parameters below 36M
3S2M2R [12] 64.93 ± 0.18 83.18 ± 0.11
LR + DC [17] 68.55 ± 0.55 82.88 ± 0.42
EASY 3×ResNet12 (ours) 71.75 ± 0.19 87.15 ± 0.12

TieredImageNet Dataset (inductive)

Methods 1-Shot 5-Way 5-Shot 5-Way
SimpleShot [29] 69.09 ± 0.22 84.58 ± 0.16
ProtoNet [10] 65.65 ± 0.92 83.40 ± 0.65
FEAT [36] 70.80 ± 0.23 84.79 ± 0.16
PAL [8] 72.25 ± 0.72 86.95 ± 0.47
DeepEMD v2 [13] 74.29 ± 0.32 86.98 ± 0.60
MERL [38] 72.14 ± 0.51 87.01 ± 0.35
COSOC [9] 73.57 ± 0.43 87.57 ± 0.10
CNL [37] 73.42 ± 0.95 87.72 ± 0.75
invariance-equivariance [39] 72.21 ± 0.90 87.08 ± 0.58
CSEI [40] 73.76 ± 0.32 87.83 ± 0.59
ASY ResNet12 (ours) 74.31 ± 0.22 87.86 ± 0.15
above <=12M nb of parameters below 36M
S2M2R [12] 73.71 ± 0.22 88.52 ± 0.14
EASY 3×ResNet12 (ours) 74.71 ± 0.22 88.33 ± 0.14

CUBFS Dataset (inductive)

Methods 1-Shot 5-Way 5-Shot 5-Way
FEAT [36] 68.87 ± 0.22 82.90 ± 0.10
LaplacianShot [41] 80.96 88.68
ProtoNet [10] 66.09 ± 0.92 82.50 ± 0.58
DeepEMD v2 [13] 79.27 ± 0.29 89.80 ± 0.51
EASY 4×ResNet12 1/sqrt(2) 77.97 ± 0.20 91.59 ± 0.10
above <=12M nb of parameters below 36M
S2M2R [12] 80.68 ± 0.81 90.85 ± 0.44
EASY 3×ResNet12 (ours) 78.56 ± 0.19 91.93 ± 0.10

CIFAR-FS Dataset (inductive)

Methods 1-Shot 5-Way 5-Shot 5-Way
S2M2R [12] 63.66 ± 0.17 76.07 ± 0.19
R2-D2 (+ens) [20] 76.51 ± 0.47 87.63 ± 0.34
invariance-equivariance [39] 77.87 ± 0.85 89.74 ± 0.57
EASY 2×ResNet12 1/sqrt(2) (ours) 75.24 ± 0.20 88.38 ± 0.14
above <=12M nb of parameters below 36M
S2M2R [12] 74.81 ± 0.19 87.47 ± 0.13
EASY 3×ResNet12 (ours) 76.20 ± 0.20 89.00 ± 0.14

FC-100 Dataset (inductive)

Methods 1-Shot 5-Way 5-Shot 5-Way
DeepEMD v2 [13] 46.60 ± 0.26 63.22 ± 0.71
TADAM [35] 40.10 ± 0.40 56.10 ± 0.40
ProtoNet [10] 41.54 ± 0.76 57.08 ± 0.76
invariance-equivariance [39] 47.76 ± 0.77 65.30 ± 0.76
R2-D2 (+ens) [20] 44.75 ± 0.43 59.94 ± 0.41
EASY 2×ResNet12 1/sqrt(2) (ours) 47.94 ± 0.19 64.14 ± 0.19
above <=12M nb of parameters below 36M
EASY 3×ResNet12 (ours) 48.07 ± 0.19 64.74 ± 0.19

Minimagenet (transductive)

Methods 1-Shot 5-Way 5-Shot 5-Way
TIM-GD [42] 73.90 85.00
ODC [43] 77.20 ± 0.36 87.11 ± 0.42
PEMnE-BMS∗ [32] 80.56 ± 0.27 87.98 ± 0.14
SSR [44] 68.10 ± 0.60 76.90 ± 0.40
iLPC [45] 69.79 ± 0.99 79.82 ± 0.55
EPNet [31] 66.50 ± 0.89 81.60 ± 0.60
DPGN [46] 67.77 ± 0.32 84.60 ± 0.43
ECKPN [47] 70.48 ± 0.38 85.42 ± 0.46
Rot+KD+POODLE [48] 77.56 85.81
EASY 2×ResNet12( 1√2) (ours) 81.70 ±0.25 88.29 ±0.13
above <=12M nb of parameters below 36M
SSR [44] 72.40 ± 0.60 80.20 ± 0.40
fine-tuning(train+val) [49] 68.11 ± 0.69 80.36 ± 0.50
SIB+E3BM [50] 71.40 81.20
LR+DC [17] 68.57 ± 0.55 82.88 ± 0.42
EPNet [31] 70.74 ± 0.85 84.34 ± 0.53
TIM-GD [42] 77.80 87.40
PT+MAP [51] 82.92 ± 0.26 88.82 ± 0.13
iLPC [45] 83.05 ± 0.79 88.82 ± 0.42
ODC [43] 80.64 ± 0.34 89.39 ± 0.39
PEMnE-BMS∗ [32] 83.35 ± 0.25 89.53 ± 0.13
EASY 3×ResNet12 (ours) 82.75 ±0.25 88.93 ±0.12

CUB-FS (transductive)

Methods 1-Shot 5-Way 5-Shot 5-Way
TIM-GD [42] 82.20 90.80
ODC [43] 85.87 94.97
DPGN [46] 75.71 ± 0.47 91.48 ± 0.33
ECKPN [47] 77.43 ± 0.54 92.21 ± 0.41
iLPC [45] 89.00 ± 0.70 92.74 ± 0.35
Rot+KD+POODLE [48] 89.93 93.78
EASY 4×ResNet12( 1/2) (ours) 90.41 ± 0.19 93.58 ± 0.10
above <=12M nb of parameters below 36M
LR+DC [17] 79.56 ± 0.87 90.67 ± 0.35
PT+MAP [51] 91.55 ± 0.19 93.99 ± 0.10
iLPC [45] 91.03 ± 0.63 94.11 ± 0.30
EASY 3×ResNet12 (ours) 90.76 ± 0.19 93.90 ± 0.09

CIFAR-FS (transductive)

Methods 1-Shot 5-Way 5-Shot 5-Way
SSR [44] 76.80 ± 0.60 83.70 ± 0.40
iLPC [45] 77.14 ± 0.95 85.23 ± 0.55
DPGN [46] 77.90 ± 0.50 90.02 ± 0.40
ECKPN [47] 79.20 ± 0.40 91.00 ± 0.50
EASY 2×ResNet12 (1/sqrt(2)) (ours) 86.40 ± 0.23 89.75 ± 0.15
above <=12M nb of parameters below 36M
SSR [44] 81.60 ± 0.60 86.00 ± 0.40
fine-tuning (train+val) [49] 78.36 ± 0.70 87.54 ± 0.49
iLPC [45] 86.51 ± 0.75 90.60 ± 0.48
PT+MAP [51] 87.69 ± 0.23 90.68 ± 0.15
EASY 3×ResNet12 (ours) 86.96 ± 0.22 90.30 ± 0.15

FC-100 (transductive)

Methods 1-Shot 5-Way 5-Shot 5-Way
EASY 2×ResNet12( 1√2)(ours) 54.68 ± 0.25 66.19 ± 0.20
above <=12M nb of parameters below 36M
SIB+E3BM [50] 46.00 57.10
fine-tuning (train) [49] 43.16 ± 0.59 57.57 ± 0.55
ODC [43] 47.18 ± 0.30 59.21 ± 0.56
fine-tuning (train+val) [49] 50.44 ± 0.68 65.74 ± 0.60
EASY 3×ResNet12 (ours) 55.11 ± 0.25 67.09 ± 0.20

Tiered Imagenet (transducive)

Methods 1-Shot 5-Way 5-Shot 5-Way
PT+MAP [51] 85.67 ± 0.26 90.45 ± 0.14
TIM-GD [42] 79.90 88.50
ODC [43] 83.73 ± 0.36 90.46 ± 0.46
SSR [44] 81.20 ± 0.60 85.70 ± 0.40
Rot+KD+POODLE [48] 79.67 86.96
DPGN [46] 72.45 ± 0.51 87.24 ± 0.39
EPNet [31] 76.53 ± 0.87 87.32 ± 0.64
ECKPN [47] 73.59 ± 0.45 88.13 ± 0.28
iLPC [45] 83.49 ± 0.88 89.48 ± 0.47
ASY ResNet12 (ours) 82.66 ± 0.27 88.60 ± 0.14
above <=12M nb of parameters below 36M
SIB+E3BM [50] 75.60 84.30
SSR [44] 79.50 ± 0.60 84.80 ± 0.40
fine-tuning (train+val) [49] 72.87 ± 0.71 86.15 ± 0.50
TIM-GD [42] 82.10 89.80
LR+DC [17] 78.19 ± 0.25 89.90 ± 0.41
EPNet [31] 78.50 ± 0.91 88.36 ± 0.57
ODC [43] 85.22 ± 0.34 91.35 ± 0.42
iLPC [45] 88.50 ± 0.75 92.46 ± 0.42
PEMnE-BMS∗ [32] 86.07 ± 0.25 91.09 ± 0.14
EASY 3×ResNet12 (ours) 84.48 ± 0.27 89.71 ± 0.14
Owner
Yassir BENDOU
Ph.D student working on Few-shot learning problems. I enjoy maths and coding.
Yassir BENDOU
GitHub Repository
modelvshuman is a Python library to benchmark the gap between human and machine vision

modelvshuman is a Python library to benchmark the gap between human and machine vision. Using this library, both PyTorch and TensorFlow models can be evaluated on 17 out-of-distribution datasets with

Bethge Lab 244 Jan 03, 2023
code for our paper "Source Data-absent Unsupervised Domain Adaptation through Hypothesis Transfer and Labeling Transfer"

SHOT++ Code for our TPAMI submission "Source Data-absent Unsupervised Domain Adaptation through Hypothesis Transfer and Labeling Transfer" that is ext

75 Dec 16, 2022
PyTorch Implementation of CycleGAN and SSGAN for Domain Transfer (Minimal)

MNIST-to-SVHN and SVHN-to-MNIST PyTorch Implementation of CycleGAN and Semi-Supervised GAN for Domain Transfer. Prerequites Python 3.5 PyTorch 0.1.12

Yunjey Choi 401 Dec 30, 2022
CoTr: Efficiently Bridging CNN and Transformer for 3D Medical Image Segmentation

CoTr: Efficient 3D Medical Image Segmentation by bridging CNN and Transformer This is the official pytorch implementation of the CoTr: Paper: CoTr: Ef

218 Dec 25, 2022
MaRS - a recursive filtering framework that allows for truly modular multi-sensor integration

The Modular and Robust State-Estimation Framework, or short, MaRS, is a recursive filtering framework that allows for truly modular multi-sensor integration

Control of Networked Systems - University of Klagenfurt 143 Dec 29, 2022
PyTorch implementations of deep reinforcement learning algorithms and environments

Deep Reinforcement Learning Algorithms with PyTorch This repository contains PyTorch implementations of deep reinforcement learning algorithms and env

Petros Christodoulou 4.7k Jan 04, 2023
DVG-Face: Dual Variational Generation for Heterogeneous Face Recognition, TPAMI 2021

DVG-Face: Dual Variational Generation for HFR This repo is a PyTorch implementation of DVG-Face: Dual Variational Generation for Heterogeneous Face Re

52 Dec 30, 2022
PyTorch implementation of Weak-shot Fine-grained Classification via Similarity Transfer

SimTrans-Weak-Shot-Classification This repository contains the official PyTorch implementation of the following paper: Weak-shot Fine-grained Classifi

BCMI 60 Dec 02, 2022
AEI: Actors-Environment Interaction with Adaptive Attention for Temporal Action Proposals Generation

AEI: Actors-Environment Interaction with Adaptive Attention for Temporal Action Proposals Generation A pytorch-version implementation codes of paper:

11 Dec 13, 2022
[NeurIPS 2021] Official implementation of paper "Learning to Simulate Self-driven Particles System with Coordinated Policy Optimization".

Code for Coordinated Policy Optimization Webpage | Code | Paper | Talk (English) | Talk (Chinese) Hi there! This is the source code of the paper “Lear

DeciForce: Crossroads of Machine Perception and Autonomy 81 Dec 19, 2022
tmm_fast is a lightweight package to speed up optical planar multilayer thin-film device computation.

tmm_fast tmm_fast or transfer-matrix-method_fast is a lightweight package to speed up optical planar multilayer thin-film device computation. It is es

26 Dec 11, 2022
A framework to train language models to learn invariant representations.

Invariant Language Modeling Implementation of the training for invariant language models. Motivation Modern pretrained language models are critical co

6 Nov 16, 2022
This is the codebase for the ICLR 2021 paper Trajectory Prediction using Equivariant Continuous Convolution

Trajectory Prediction using Equivariant Continuous Convolution (ECCO) This is the codebase for the ICLR 2021 paper Trajectory Prediction using Equivar

Spatiotemporal Machine Learning 45 Jul 22, 2022
[ArXiv 2021] Data-Efficient Instance Generation from Instance Discrimination

InsGen - Data-Efficient Instance Generation from Instance Discrimination Data-Efficient Instance Generation from Instance Discrimination Ceyuan Yang,

GenForce: May Generative Force Be with You 93 Dec 25, 2022
This repo includes the CUB-GHA (Gaze-based Human Attention) dataset and code of the paper "Human Attention in Fine-grained Classification".

HA-in-Fine-Grained-Classification This repo includes the CUB-GHA (Gaze-based Human Attention) dataset and code of the paper "Human Attention in Fine-g

16 Oct 29, 2022
Pytorch Implementation of PointNet and PointNet++++

Pytorch Implementation of PointNet and PointNet++ This repo is implementation for PointNet and PointNet++ in pytorch. Update 2021/03/27: (1) Release p

Luigi Ariano 1 Nov 11, 2021
MinkLoc++: Lidar and Monocular Image Fusion for Place Recognition

MinkLoc++: Lidar and Monocular Image Fusion for Place Recognition Paper: MinkLoc++: Lidar and Monocular Image Fusion for Place Recognition accepted fo

64 Dec 18, 2022
Image processing in Python

scikit-image: Image processing in Python Website (including documentation): https://scikit-image.org/ Mailing list: https://mail.python.org/mailman3/l

Image Processing Toolbox for SciPy 5.2k Dec 31, 2022
Angular & Electron desktop UI framework. Angular components for native looking and behaving macOS desktop UI (Electron/Web)

Angular Desktop UI This is a collection for native desktop like user interface components in Angular, especially useful for Electron apps. It starts w

Marc J. Schmidt 49 Dec 22, 2022
A distributed, plug-n-play algorithm for multi-robot applications with a priori non-computable objective functions

A distributed, plug-n-play algorithm for multi-robot applications with a priori non-computable objective functions Kapoutsis, A.C., Chatzichristofis,

Athanasios Ch. Kapoutsis 5 Oct 15, 2022