ConvMixer

ConvMixer: Patches Are All You Need? arxiv

PaddlePaddle training/validation code and pretrained models for ConvMixer.

The official pytorch implementation is here.

This implementation is developed by PaddleViT.

ConvMixer Model Overview

Update

• Update (2022-04-08): Code is updated.
• Update (2021-11-04): Model weights are updated.
• Update (2021-10-13): Code is released and ported weights are uploaded.

Models Zoo

Model	Acc@1	Acc@5	#Params	FLOPs	Image Size	Crop_pct	Interpolation	Link
convmixer_1024_20	76.94	93.35	24.5M	9.5G	224	0.96	bicubic	google/baidu
convmixer_768_32	80.16	95.08	21.2M	20.8G	224	0.96	bicubic	google/baidu
convmixer_1536_20	81.37	95.62	51.8M	72.4G	224	0.96	bicubic	google/baidu

*The results are evaluated on ImageNet2012 validation set.

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 ./convmixer_768_32.pdparams, to use the convmixer_768_32 model in python:

from config import get_config
from convmixer import build_convmixer as build_model
# config files in ./configs/
config = get_config('./configs/convmixer_768_32.yaml')
# build model
model = build_model(config)
# load pretrained weights
model_state_dict = paddle.load('./convmixer_768_32.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/convmixer_768_32.yaml' \
-dataset='imagenet2012' \
-batch_size=256 \
-data_path='/dataset/imagenet' \
-eval \
-pretrained='./convmixer_768_32.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/convmixer_768_32.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{trockman2022patches,
  title={Patches Are All You Need?},
  author={Trockman, Asher and Kolter, J Zico},
  journal={arXiv preprint arXiv:2201.09792},
  year={2022}
}