Skip to content

Latest commit

 

History

History
 
 

CvT

README.md

CvT: Introducing Convolutions to Vision Transformers, arxiv

PaddlePaddle training/validation code and pretrained models for CvT.

The official pytorch implementation is here.

This implementation is developed by PaddleViT.

drawing

CvT Model Overview

Update

  • Update (2021-12-24): Code is released and ported weights are uploaded.

Models Zoo

Model Acc@1 Acc@5 #Params FLOPs Image Size Crop_pct Interpolation Link
CvT-13-224 81.59 95.67 20M 4.5G 224 0.875 bicubic google/baidu(vev9)
CvT-21-224 82.46 96.00 32M 7.1G 224 0.875 bicubic google/baidu(t2rv)
CvT-13-384 83.00 96.36 20M 16.3G 384 1.0 bicubic google/baidu(wswt)
CvT-21-384 83.27 96.16 32M 24.9G 384 1.0 bicubic google/baidu(hcem)
CvT-13-384-22k 83.26 97.09 20M 16.3G 384 1.0 bicubic google/baidu(c7m9)
CvT-21-384-22k 84.91 97.62 32M 24.9G 384 1.0 bicubic google/baidu(9jxe)
CvT-w24-384-22k 87.58 98.47 277M 193.2G 384 1.0 bicubic google/baidu(bbj2)

*The results are evaluated on ImageNet2012 validation set.

Notebooks

We provide a few notebooks in aistudio to help you get started:

*(coming soon)*

Requirements

Data

ImageNet2012 dataset is used in the following folder structure:

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

Usage

To use the model with pretrained weights, download the .pdparam weight file and change related file paths in the following python scripts. The model config files are located in ./configs/.

For example, assume the downloaded weight file is stored in ./CvT-13-224x224-IN-1k.pdparams, to use the CvT-13-224x224-IN-1k model in python:

from config import get_config
from cvt import build_cvt as build_model
# config files in ./configs/
config = get_config('./configs/cvt-13-224x224.yaml')
# build model
model = build_model(config)
# load pretrained weights, .pdparams is NOT needed
model_state_dict = paddle.load('./CvT-13-224x224-IN-1k')
model.set_dict(model_state_dict)

Evaluation

To evaluate CvT model performance on ImageNet2012 with a single GPU, run the following script using command line:

sh run_eval.sh

or

CUDA_VISIBLE_DEVICES=0 \
python main_single_gpu.py \
    -cfg='./configs/cvt-13-224x224.yaml' \
    -dataset='imagenet2012' \
    -batch_size=16 \
    -data_path='/dataset/imagenet' \
    -eval \
    -pretrained='./CvT-13-224x224-IN-1k'
Run evaluation using multi-GPUs: 
sh run_eval_multi.sh

or

CUDA_VISIBLE_DEVICES=0,1,2,3 \
python main_multi_gpu.py \
    -cfg='./configs/cvt-13-224x224.yaml' \
    -dataset='imagenet2012' \
    -batch_size=16 \
    -data_path='/dataset/imagenet' \
    -eval \
    -pretrained='./CvT-13-224x224-IN-1k'

Training

To train the CvT Transformer model on ImageNet2012 with single GPUs, run the following script using command line:

sh run_train.sh

or

CUDA_VISIBLE_DEVICES=0 \
python main_single_gpu.py \
  -cfg='./configs/cvt-13-224x224.yaml' \
  -dataset='imagenet2012' \
  -batch_size=32 \
  -data_path='/dataset/imagenet' \
Run training using multi-GPUs: 
sh run_train_multi.sh

or

CUDA_VISIBLE_DEVICES=0,1,2,3 \
python main_multi_gpu.py \
    -cfg='./configs/cvt-13-224x224.yaml' \
    -dataset='imagenet2012' \
    -batch_size=16 \
    -data_path='/dataset/imagenet' \

Visualization Attention Map

(coming soon)

Reference

@article{wu2021cvt,
title={CvT: Introducing Convolutions to Vision Transformers},
author={Haiping Wu, Bin Xiao, Noel Codella, Mengchen Liu, Xiyang Dai, Lu Yuan, Lei Zhang},
journal={arXiv preprint arXiv:2103.15808},
year={2021}
}