RaMLP: Vision MLP via Region-aware Mixing

You could reproduce the model by the code.

nohup python -u -m torch.distributed.run --nproc_per_node=8 main.py --model tiny --drop_path 0.2 --epochs 300 --batch_size 128 --lr 4.0e-3 --update_freq 4 --model_ema false --model_ema_eval false --use_amp true --data_path /INPUT/dataset/imagenet --output_dir ./checkpoint &

nohup python -u -m torch.distributed.run --nproc_per_node=8 main.py --model small --drop_path 0.3 --epochs 300 --batch_size 128 --lr 4.0e-3 --update_freq 4 --model_ema false --model_ema_eval false --use_amp true --data_path /INPUT/dataset/imagenet --output_dir ./checkpoint &

nohup python -u -m torch.distributed.run --nproc_per_node=8 main.py --model base --drop_path 0.4 --epochs 300 --batch_size 128 --lr 4.0e-3 --update_freq 4 --model_ema false --model_ema_eval false --use_amp true --data_path /INPUT/dataset/imagenet --output_dir ./checkpoint &

Results and Pre-trained Models

ImageNet-1K trained models

name	resolution	acc@1	#params	FLOPs	model
RaMLP-T	224x224	82.9	25M	4.2G	model
RaMLP-S	224x224	83.8	38M	7.8G	model
RaMLP-B	224x224	84.1	58M	12.0G	model

Citation

If you find this repository helpful, please consider citing:

@Article{liu2022convnet,
  author  = {Shenqi Lai, Xi Du and Jia Guo and Kaipeng Zhang},
  title   = {RaMLP: Vision MLP via Region-aware Mixing},
  journal = {IJCAI},
  year    = {2023},
}