Skip to content

Latest commit

 

History

History

RepLKNet

README.md

RepLKNet: Scaling Up Your Kernels to 31x31: Revisiting Large Kernel Design in CNNs arxiv

PaddlePaddle training/validation code and pretrained models for the model released in CVPR2022: RepLKNet (classification backbone).

The official PyTorch implementation is here.

This implementation is developed by PaddleViT.

drawing

RepLKNet Model Overview

Update

  • Update (2022-07-15): Code is released and ported weights are uploaded.

Models Zoo

Model Acc@1 Acc@5 #Params FLOPs Image Size Crop_pct Interpolation Link
replknet-31b-1k-224 83.48 96.57 79M 15.3G 224 0.875 bicubic google/baidu
replknet-31b-1k-384 84.58 97.23 79M 45.1G 384 1.0 bicubic google/baidu
replknet-31b-22k1k-224 85.20 97.56 79M 15.3G 224 0.875 bicubic google/baidu
replknet-31b-22k1k-384 85.77 97.68 79M 45.1G 384 1.0 bicubic google/baidu
replknet-31l-22k1k-384 86.38 97.88 172M 96.0G 384 1.0 bicubic google/baidu
replknet-xl-73m1k-320 - - 335M 128.7G 320 1.0 bicubic google/baidu
replknet-31b-22k - - 79M 15.3G 224 0.875 bicubic google/baidu
replknet-31l-22k - - 172M 96.0G 384 1.0 bicubic google/baidu

*The results are above are ported from official implemetation and 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 ./replknet_31b.pdparams, to use the replknet_31b model in python:

from config import get_config
from replknet import build_replknet as build_model
# config files in ./configs/
config = get_config('./configs/replknet_31b.yaml')
# build model
model = build_model(config)
# load pretrained weights
model_state_dict = paddle.load('./replknet_31b.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/replknet_31b.yaml' \
-dataset='imagenet2012' \
-batch_size=256 \
-data_path='/dataset/imagenet' \
-eval \
-pretrained='./replknet_31b.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/replknet_31b.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{replknet,
title={Scaling Up Your Kernels to 31x31: Revisiting Large Kernel Design in CNNs},
author={Ding, Xiaohan and Zhang, Xiangyu and Zhou, Yizhuang and Han, Jungong and Ding, Guiguang and Sun, Jian},
journal={arXiv preprint arXiv:2203.06717},
year={2022}
}