DynamicViT: Efficient Vision Transformers with Dynamic Token Sparsification

Related tags

Deep Learningcomputer-visiondeep-learningimage-classificationvision-transformers
Overview

DynamicViT: Efficient Vision Transformers with Dynamic Token Sparsification

Created by Yongming Rao, Wenliang Zhao, Benlin Liu, Jiwen Lu, Jie Zhou, Cho-Jui Hsieh

This repository contains PyTorch implementation for DynamicViT.

We introduce a dynamic token sparsification framework to prune redundant tokens in vision transformers progressively and dynamically based on the input:

intro

Our code is based on pytorch-image-models, DeiT and LV-ViT

[Project Page] [arXiv]

Model Zoo

We provide our DynamicViT models pretrained on ImageNet:

name arch rho [email protected] [email protected] FLOPs url
DynamicViT-256/0.7 deit_256 0.7 76.532 93.118 1.3G Google Drive / Tsinghua Cloud
DynamicViT-384/0.7 deit_small 0.7 79.316 94.676 2.9G Google Drive / Tsinghua Cloud
DynamicViT-LV-S/0.5 lvvit_s 0.5 81.970 95.756 3.7G Google Drive / Tsinghua Cloud
DynamicViT-LV-S/0.7 lvvit_s 0.7 83.076 96.252 4.6G Google Drive / Tsinghua Cloud
DynamicViT-LV-M/0.7 lvvit_m 0.7 83.816 96.584 8.5G Google Drive / Tsinghua Cloud

Usage

Requirements

  • torch>=1.7.0
  • torchvision>=0.8.1
  • timm==0.4.5

Data preparation: download and extract ImageNet images from http://image-net.org/. The directory structure should be

│ILSVRC2012/
├──train/
│  ├── n01440764
│  │   ├── n01440764_10026.JPEG
│  │   ├── n01440764_10027.JPEG
│  │   ├── ......
│  ├── ......
├──val/
│  ├── n01440764
│  │   ├── ILSVRC2012_val_00000293.JPEG
│  │   ├── ILSVRC2012_val_00002138.JPEG
│  │   ├── ......
│  ├── ......

Model preparation: download pre-trained DeiT and LV-ViT models for training DynamicViT:

sh download_pretrain.sh

Demo

We provide a Jupyter notebook where you can run the visualization of DynamicViT.

To run the demo, you need to install matplotlib.

demo

Evaluation

To evaluate a pre-trained DynamicViT model on ImageNet val with a single GPU, run:

python infer.py --data-path /path/to/ILSVRC2012/ --arch arch_name --model-path /path/to/model --base_rate 0.7

Training

To train DynamicViT models on ImageNet, run:

DeiT-small

python -m torch.distributed.launch --nproc_per_node=8 --use_env main_dynamic_vit.py  --output_dir logs/dynamic-vit_deit-small --arch deit_small --input-size 224 --batch-size 96 --data-path /path/to/ILSVRC2012/ --epochs 30 --dist-eval --distill --base_rate 0.7

LV-ViT-S

python -m torch.distributed.launch --nproc_per_node=8 --use_env main_dynamic_vit.py  --output_dir logs/dynamic-vit_lvvit-s --arch lvvit_s --input-size 224 --batch-size 64 --data-path /path/to/ILSVRC2012/ --epochs 30 --dist-eval --distill --base_rate 0.7

LV-ViT-M

python -m torch.distributed.launch --nproc_per_node=8 --use_env main_dynamic_vit.py  --output_dir logs/dynamic-vit_lvvit-m --arch lvvit_m --input-size 224 --batch-size 48 --data-path /path/to/ILSVRC2012/ --epochs 30 --dist-eval --distill --base_rate 0.7

You can train models with different keeping ratio by adjusting base_rate. DynamicViT can also achieve comparable performance with only 15 epochs training (around 0.1% lower accuracy).

License

MIT License

Citation

If you find our work useful in your research, please consider citing:

@article{rao2021dynamicvit,
  title={DynamicViT: Efficient Vision Transformers with Dynamic Token Sparsification},
  author={Rao, Yongming and Zhao, Wenliang and Liu, Benlin and Lu, Jiwen and Zhou, Jie and Hsieh, Cho-Jui},
  journal={arXiv preprint arXiv:2106.02034},
  year={2021}
}
Owner
Yongming Rao
Yongming Rao
GitHub Repository https://dynamicvit.ivg-research.xyz/
learning and feeling SLAM together with hands-on-experiments

modern-slam-tutorial-python Learning and feeling SLAM together with hands-on-experiments 😀 😃 😆 Dependencies Most of the examples are based on GTSAM

Giseop Kim 59 Dec 22, 2022
Realtime segmentation with ENet, the fast and accurate segmentation net.

Enet This is a realtime segmentation net with almost 22 fps on GTX1080 ti, and the model size is very small with only 28M. This repo contains the infe

JinTian 14 Aug 30, 2022
Over9000 optimizer

Optimizers and tests Every result is avg of 20 runs. Dataset LR Schedule Imagenette size 128, 5 epoch Imagewoof size 128, 5 epoch Adam - baseline OneC

Mikhail Grankin 405 Nov 27, 2022
Detector for Log4Shell exploitation attempts

log4shell-detector Detector for Log4Shell exploitation attempts Idea The problem with the log4j CVE-2021-44228 exploitation is that the string can be

Florian Roth 729 Dec 25, 2022
This is the pytorch implementation for the paper: Generalizable Mixed-Precision Quantization via Attribution Rank Preservation, which is accepted to ICCV2021.

GMPQ: Generalizable Mixed-Precision Quantization via Attribution Rank Preservation This is the pytorch implementation for the paper: Generalizable Mix

18 Sep 02, 2022
Pytorch implementation of CVPR2021 paper "MUST-GAN: Multi-level Statistics Transfer for Self-driven Person Image Generation"

MUST-GAN Code | paper The Pytorch implementation of our CVPR2021 paper "MUST-GAN: Multi-level Statistics Transfer for Self-driven Person Image Generat

TianxiangMa 46 Dec 26, 2022
Mixed Transformer UNet for Medical Image Segmentation

MT-UNet Update 2022/01/05 By another round of training based on previous weights, our model also achieved a better performance on ACDC (91.61% DSC). W

dotman 92 Dec 25, 2022
A Research-oriented Federated Learning Library and Benchmark Platform for Graph Neural Networks. Accepted to ICLR'2021 - DPML and MLSys'21 - GNNSys workshops.

FedGraphNN: A Federated Learning System and Benchmark for Graph Neural Networks A Research-oriented Federated Learning Library and Benchmark Platform

FedML-AI 175 Dec 01, 2022
A vanilla 3D face modeling on pose-invariant and multi-lightning image data

3D-Face-Modeling A vanilla 3D face modeling on pose-invariant and multi-lightning image data Table of Contents Background Install Usage Contributing B

Haochen Zhang 1 Mar 12, 2022
Generate saved_model, tfjs, tf-trt, EdgeTPU, CoreML, quantized tflite and .pb from .tflite.

tflite2tensorflow Generate saved_model, tfjs, tf-trt, EdgeTPU, CoreML, quantized tflite and .pb from .tflite. 1. Supported Layers No. TFLite Layer TF

Katsuya Hyodo 214 Dec 29, 2022
Advanced Signal Processing Notebooks and Tutorials

Advanced Digital Signal Processing Notebooks and Tutorials Prof. Dr. -Ing. Gerald Schuller Jupyter Notebooks and Videos: Renato Profeta Applied Media

Guitars.AI 115 Dec 13, 2022
Automatic detection and classification of Covid severity degree in LUS (lung ultrasound) scans

Final-Project Final project in the Technion, Biomedical faculty, by Mor Ventura, Dekel Brav & Omri Magen. Subproject 1: Automatic Detection of LUS Cha

Mor Ventura 1 Dec 18, 2021
BLEND: A Fast, Memory-Efficient, and Accurate Mechanism to Find Fuzzy Seed Matches

BLEND is a mechanism that can efficiently find fuzzy seed matches between sequences to significantly improve the performance and accuracy while reducing the memory space usage of two important applic

SAFARI Research Group at ETH Zurich and Carnegie Mellon University 19 Dec 26, 2022
Graph Transformer Architecture. Source code for

Graph Transformer Architecture Source code for the paper "A Generalization of Transformer Networks to Graphs" by Vijay Prakash Dwivedi and Xavier Bres

NTU Graph Deep Learning Lab 561 Jan 08, 2023
🥈78th place in Riiid Solution🥈

Riiid Answer Correctness Prediction Introduction This repository is the code that placed 78th in Riiid Answer Correctness Prediction competition. Requ

ds wook 14 Apr 26, 2022
Vrcwatch - Supply the local time to VRChat as Avatar Parameters through OSC

English: README-EN.md VRCWatch VRCWatch は、VRChat 内のアバター向けに現在時刻を送信するためのプログラムです。 使

Kosaki Mezumona 17 Nov 30, 2022
Pipeline code for Sequential-GAM(Genome Architecture Mapping).

Sequential-GAM Pipeline code for Sequential-GAM(Genome Architecture Mapping). mapping whole_preprocess.sh include the whole processing of mapping. usa

3 Nov 03, 2022
[ICCV2021] Learning to Track Objects from Unlabeled Videos

Unsupervised Single Object Tracking (USOT) 🌿 Learning to Track Objects from Unlabeled Videos Jilai Zheng, Chao Ma, Houwen Peng and Xiaokang Yang 2021

53 Dec 28, 2022
Pytorch implementation of "Training a 85.4% Top-1 Accuracy Vision Transformer with 56M Parameters on ImageNet"

Token Labeling: Training an 85.4% Top-1 Accuracy Vision Transformer with 56M Parameters on ImageNet (arxiv) This is a Pytorch implementation of our te

蒋子航 383 Dec 27, 2022
A python library for self-supervised learning on images.

Lightly is a computer vision framework for self-supervised learning. We, at Lightly, are passionate engineers who want to make deep learning more effi

Lightly 2k Jan 08, 2023