CycleMLP

CycleMLP: A MLP-like Architecture for Dense Prediction, arxiv

PaddlePaddle training/validation code and pretrained models for CycleMLP.

The official and 3rd party pytorch implementation are here.

This implementation is developed by PPViT.

CycleMLP Model Overview

Update

Update (2022-04-08): Code is updated. Update (2021-09-24): Code is released and ported weights are uploaded.

Models Zoo

Model	Acc@1	Acc@5	#Params	Image Size	Crop_pct	Interpolation	Link
cyclemlp_b1	78.85	94.60	15.1M	224	0.9	bicubic	google/baidu
cyclemlp_b2	81.58	95.81	26.8M	224	0.9	bicubic	google/baidu
cyclemlp_b3	82.42	96.07	38.3M	224	0.9	bicubic	google/baidu
cyclemlp_b4	82.96	96.33	51.8M	224	0.875	bicubic	google/baidu
cyclemlp_b5	83.25	96.44	75.7M	224	0.875	bicubic	google/baidu

*The results are evaluated on ImageNet2012 validation set.

Note: CycleMLP weights are ported from here

Data Preparation

ImageNet2012 dataset is used in the following file structure:

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

• train_list.txt: list of relative paths and labels of training images. You can download it from: google/baidu
• val_list.txt: list of relative paths and labels of validation images. You can download it from: google/baidu

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 weight file is downloaded in ./cyclemlp_b1.pdparams, to use the cyclemlp_b1 model in python:

from config import get_config
from cyclemlp import build_cyclemlp as build_model
# config files in ./configs/
config = get_config('./configs/cyclemlp_b1.yaml')
# build model
model = build_model(config)
# load pretrained weights
model_state_dict = paddle.load('./cyclemlp_b1.pdparams')
model.set_state_dict(model_state_dict)

Evaluation

To evaluate model performance on ImageNet2012, run the following script using command line:

sh run_eval_multi.sh

or

CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 \
python main_multi_gpu.py \
-cfg='./configs/cyclemlp_b1.yaml' \
-dataset='imagenet2012' \
-batch_size=256 \
-data_path='/dataset/imagenet' \
-eval \
-pretrained='./cyclemlp_b1.pdparams' \
-amp

Note: if you have only 1 GPU, change device number to CUDA_VISIBLE_DEVICES=0 would run the evaluation on single GPU.

Training

To train the model on ImageNet2012, run the following script using command line:

sh run_train_multi.sh

or

CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 \
python main_multi_gpu.py \
-cfg='./configs/cyclemlp_b1.yaml' \
-dataset='imagenet2012' \
-batch_size=256 \
-data_path='/dataset/imagenet' \
-amp

Note: it is highly recommanded to run the training using multiple GPUs / multi-node GPUs.

Reference

@article{chen2021cyclemlp,
  title={CycleMLP: A MLP-like Architecture for Dense Prediction},
  author={Chen, Shoufa and Xie, Enze and Ge, Chongjian and Liang, Ding and Luo, Ping},
  journal={arXiv preprint arXiv:2107.10224},
  year={2021}
}