Pay Attention to MLPs, arxiv

PaddlePaddle training/validation code and pretrained models for gMLP.

The 3rd party pytorch implementation is here.

This implementation is developed by PaddleViT.

gMLP Model Overview

Update

• Update (2021-09-27): Model FLOPs and # params are uploaded.
• Update (2021-08-11): Code is released and ported weights are uploaded.

Models Zoo

Model	Acc@1	Acc@5	#Params	FLOPs	Image Size	Crop_pct	Interpolation	Link
gmlp_s16_224	79.64	94.63	19.4M	4.5G	224	0.875	bicubic	google/baidu(bcth)

*The results are evaluated on ImageNet2012 validation set.

Note: gMLP weights are ported from timm

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 ./gmlp_s16_224.pdparams, to use the gmlp_s16_224 model in python:

from config import get_config
from gmlp import build_gated_mlp as build_model
# config files in ./configs/
config = get_config('./configs/gmlp_s16_224.yaml')
# build model
model = build_model(config)
# load pretrained weights, .pdparams is NOT needed
model_state_dict = paddle.load('./gmlp_s16_224')
model.set_dict(model_state_dict)

Evaluation

To evaluate gMLP 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/gmlp_s16_224.yaml' \
    -dataset='imagenet2012' \
    -batch_size=16 \
    -data_path='/dataset/imagenet' \
    -eval \
    -pretrained='./gmlp_s16_224'

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/gmlp_s16_224.yaml' \
    -dataset='imagenet2012' \
    -batch_size=16 \
    -data_path='/dataset/imagenet' \
    -eval \
    -pretrained='./gmlp_s16_224'

Training

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

sh run_train.sh

or

CUDA_VISIBLE_DEVICES=0 \
python main_single_gpu.py \
  -cfg='./configs/gmlp_s16_224.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/gmlp_s16_224.yaml' \
    -dataset='imagenet2012' \
    -batch_size=16 \
    -data_path='/dataset/imagenet' \

Visualization Attention Map

(coming soon)

Reference

@article{zhang2021gmlp,
  title={GMLP: Building Scalable and Flexible Graph Neural Networks with Feature-Message Passing},
  author={Zhang, Wentao and Shen, Yu and Lin, Zheyu and Li, Yang and Li, Xiaosen and Ouyang, Wen and Tao, Yangyu and Yang, Zhi and Cui, Bin},
  journal={arXiv preprint arXiv:2104.09880},
  year={2021}
}