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

Localized representation learning from Vision and Text (LoVT)

Aws-machine-learning-university-accelerated-tab - Machine Learning University: Accelerated Tabular Data Class

Unimodal Face Classification with Multimodal Training

Large-scale language modeling tutorials with PyTorch

Neural Cellular Automata + CLIP

PyMove is a Python library to simplify queries and visualization of trajectories and other spatial-temporal data

This is the code for Compressing BERT: Studying the Effects of Weight Pruning on Transfer Learning

NeurIPS workshop paper 'Counter-Strike Deathmatch with Large-Scale Behavioural Cloning'

Group R-CNN for Point-based Weakly Semi-supervised Object Detection (CVPR2022)

FusionNet: A deep fully residual convolutional neural network for image segmentation in connectomics

Final project code: Implementing BicycleGAN, for CIS680 FA21 at University of Pennsylvania

Final project for Intro to CS class.

A torch.Tensor-like DataFrame library supporting multiple execution runtimes and Arrow as a common memory format

A spatial genome aligner for analyzing multiplexed DNA-FISH imaging data.

A small demonstration of using WebDataset with ImageNet and PyTorch Lightning

Code for Pose-Controllable Talking Face Generation by Implicitly Modularized Audio-Visual Representation (CVPR 2021)

PyTorch implementation of an end-to-end Handwritten Text Recognition (HTR) system based on attention encoder-decoder networks

StyleGAN2 with adaptive discriminator augmentation (ADA) - Official TensorFlow implementation

https://sites.google.com/cornell.edu/recsys2021tutorial

Some tentative models that incorporate label propagation to graph neural networks for graph representation learning in nodes, links or graphs.