Improving Transferability of Representations via Augmentation-Aware Self-Supervision

Last update: Sep 16, 2022

Related tags

Overview

Improving Transferability of Representations via Augmentation-Aware Self-Supervision

Accepted to NeurIPS 2021

TL;DR: Learning augmentation-aware information by predicting the difference between two augmented samples improves the transferability of representations.

Dependencies

conda create -n AugSelf python=3.8 pytorch=1.7.1 torchvision=0.8.2 cudatoolkit=10.1 ignite -c pytorch
conda activate AugSelf
pip install scipy tensorboard kornia==0.4.1 sklearn

Checkpoints

We provide ImageNet100-pretrained models in this Dropbox link.

Pretraining

We here provide SimSiam+AugSelf pretraining scripts. For training the baseline (i.e., no AugSelf), remove --ss-crop and --ss-color options. For using other frameworks like SimCLR, use the --framework option.

STL-10

CUDA_VISIBLE_DEVICES=0 python pretrain.py \
    --logdir ./logs/stl10/simsiam/aug_self \
    --framework simsiam \
    --dataset stl10 \
    --datadir DATADIR \
    --model resnet18 \
    --batch-size 256 \
    --max-epochs 200 \
    --ss-color 1.0 --ss-crop 1.0

ImageNet100

python pretrain.py \
    --logdir ./logs/imagenet100/simsiam/aug_self \
    --framework simsiam \
    --dataset imagenet100 \
    --datadir DATADIR \
    --batch-size 256 \
    --max-epochs 500 \
    --model resnet50 \
    --base-lr 0.05 --wd 1e-4 \
    --ckpt-freq 50 --eval-freq 50 \
    --ss-crop 0.5 --ss-color 0.5 \
    --num-workers 16 --distributed

Evaluation

Our main evaluation setups are linear evaluation on fine-grained classification datasets (Table 1) and few-shot benchmarks (Table 2).

linear evaluation

CUDA_VISIBLE_DEVICES=0 python transfer_linear_eval.py \
    --pretrain-data imagenet100 \
    --ckpt CKPT \
    --model resnet50 \
    --dataset cifar10 \
    --datadir DATADIR \
    --metric top1

few-shot

CUDA_VISIBLE_DEVICES=0 python transfer_few_shot.py \
    --pretrain-data imagenet100 \
    --ckpt CKPT \
    --model resnet50 \
    --dataset cub200 \
    --datadir DATADIR

Improving Transferability of Representations via Augmentation-Aware Self-Supervision

Related tags

Overview

Improving Transferability of Representations via Augmentation-Aware Self-Supervision

Dependencies

Checkpoints

Pretraining

STL-10

ImageNet100

Evaluation

linear evaluation

few-shot

Owner

hankook

Deep Q-network learning to play flappybird.

TensorFlow CNN for fast style transfer

PyTorch implementation for Convolutional Networks with Adaptive Inference Graphs

Chainer implementation of recent GAN variants

pix2pix in tensorflow.js

Cave Generation using metaballs in Blender. Originally created by sdfgeoff, Edited by Myself (Archie Jaskowicz).

kullanışlı ve işinizi kolaylaştıracak bir araç

The implementation of PEMP in paper "Prior-Enhanced Few-Shot Segmentation with Meta-Prototypes"

A Python training and inference implementation of Yolov5 helmet detection in Jetson Xavier nx and Jetson nano

Deep Learning Package based on TensorFlow

blind SQLIpy sebuah alat injeksi sql yang menggunakan waktu sql untuk mendapatkan sebuah server database.

Unoffical implementation about Image Super-Resolution via Iterative Refinement by Pytorch

This is the code for our paper "Iconary: A Pictionary-Based Game for Testing Multimodal Communication with Drawings and Text"

Generative Adversarial Networks(GANs)

Learning Features with Parameter-Free Layers (ICLR 2022)

Code for the paper "Reinforcement Learning as One Big Sequence Modeling Problem"

URIE: Universal Image Enhancementfor Visual Recognition in the Wild

Package for working with hypernetworks in PyTorch.

A C implementation for creating 2D voronoi diagrams

ML models and internal tensors 3D visualizer