Vision Permutator: A Permutable MLP-Like Architecture for Visual Recognition, arxiv

PaddlePaddle training/validation code and pretrained models for ViP.

The official and 3rd party pytorch implementation are here.

This implementation is developed by PPViT.

ViP Model Overview

Update

• Update (2021-11-03): Code and weights are updated.
• Update (2021-09-23): Code is released and ported weights are uploaded.

Models Zoo

Model	Acc@1	Acc@5	#Params	FLOPs	Image Size	Crop_pct	Interpolation	Link
vip_s7	81.50	95.76	25.1M	7.0G	224	0.875	bicubic	google/baidu(mh9b)
vip_m7	82.75	96.05	55.3M	16.4G	224	0.875	bicubic	google/baidu(hvm8)
vip_l7	83.18	96.37	87.8M	24.5G	224	0.875	bicubic	google/baidu(tjvh)

*The results are evaluated on ImageNet2012 validation set.

Note: ViP weights are ported from here

Notebooks

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

*(coming soon)*

Requirements

• Python>=3.6
• yaml>=0.2.5
• PaddlePaddle>=2.1.0
• yacs>=0.1.8

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 ./vip_s7.pdparams, to use the vip_s7 model in python:

from config import get_config
from vip import build_vip as build_model
# config files in ./configs/
config = get_config('./configs/vip_s7.yaml')
# build model
model = build_model(config)
# load pretrained weights, .pdparams is NOT needed
model_state_dict = paddle.load('./vip_s7')
model.set_dict(model_state_dict)

Evaluation

To evaluate ViP 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/vip_s7.yaml' \
    -dataset='imagenet2012' \
    -batch_size=16 \
    -data_path='/dataset/imagenet' \
    -eval \
    -pretrained='./vip_s7'

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/vip_s7.yaml' \
    -dataset='imagenet2012' \
    -batch_size=16 \
    -data_path='/dataset/imagenet' \
    -eval \
    -pretrained='./vip_s7'

Training

To train the ViP 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/vip_s7.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/vip_s7.yaml' \
    -dataset='imagenet2012' \
    -batch_size=16 \
    -data_path='/dataset/imagenet' \ 

Visualization Attention Map

(coming soon)

Reference

@misc{hou2021vision,
    title={Vision Permutator: A Permutable MLP-Like Architecture for Visual Recognition},
    author={Qibin Hou and Zihang Jiang and Li Yuan and Ming-Ming Cheng and Shuicheng Yan and Jiashi Feng},
    year={2021},
    eprint={2106.12368},
    archivePrefix={arXiv},
    primaryClass={cs.CV}
}