Skip to content

Latest commit

 

History

History

MobileOne

README.md

An Improved One millisecond Mobile Backbone arxiv

PaddlePaddle training/validation code and pretrained models for the model released in: MobileOne.

The official PyTorch implementation is here.

This implementation is developed by PaddleViT.

drawing

RepLKNet Model Overview

Update

  • Update (2022-08-29): Code is released and ported weights are uploaded.

Models Zoo

Model Acc@1 Acc@5 #Params FLOPs Image Size Crop_pct Interpolation Link Link(fused)
mobileone_s0 71.14 89.65 5.4M 1.1G 224 0.875 bicubic google/baidu google/baidu
mobileone_s1 75.65 92.66 4.9M 0.9G 224 0.875 bicubic google/baidu google/baidu
mobileone_s2 77.22 93.49 8.0M 1.3G 224 0.875 bicubic google/baidu google/baidu
mobileone_s3 77.74 93.76 10.3M 1.9G 224 0.875 bicubic google/baidu google/baidu
mobileone_s4 79.07 94.23 15.1M 3.0G 224 0.875 bicubic google/baidu 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 ./mobileone_s0.pdparams, to use the mobileone_s0 model in python:

from config import get_config
from mobileone import build_mobileone as build_model
# config files in ./configs/
config = get_config('./configs/mobileone_s0.yaml')
# build model
model = build_model(config)
# load pretrained weights
model_state_dict = paddle.load('./mobileone_s0.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/mobileone_s0.yaml' \
-dataset='imagenet2012' \
-batch_size=256 \
-data_path='/dataset/imagenet' \
-eval \
-pretrained='./mobileone_s0.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/mobileone_s0.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{vasu2022improved,
  title={An Improved One millisecond Mobile Backbone},
  author={Vasu, Pavan Kumar Anasosalu and Gabriel, James and Zhu, Jeff and Tuzel, Oncel and Ranjan, Anurag},
  journal={arXiv preprint arXiv:2206.04040},
  year={2022}
}