Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/shoufachen/cyclemlp
[ICLR'22 Oral] Implementation of "CycleMLP: A MLP-like Architecture for Dense Prediction"
https://github.com/shoufachen/cyclemlp
backbone-models mlp pytorch
Last synced: 19 days ago
JSON representation
[ICLR'22 Oral] Implementation of "CycleMLP: A MLP-like Architecture for Dense Prediction"
- Host: GitHub
- URL: https://github.com/shoufachen/cyclemlp
- Owner: ShoufaChen
- License: mit
- Created: 2021-07-21T14:16:51.000Z (over 3 years ago)
- Default Branch: main
- Last Pushed: 2022-04-25T10:36:03.000Z (over 2 years ago)
- Last Synced: 2024-10-11T18:09:50.488Z (about 1 month ago)
- Topics: backbone-models, mlp, pytorch
- Language: Python
- Homepage:
- Size: 554 KB
- Stars: 280
- Watchers: 4
- Forks: 29
- Open Issues: 4
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
## CycleMLP: A MLP-like Architecture for Dense Prediction (ICLR 2022 Oral)
[](https://opensource.org/licenses/MIT)
This is a PyTorch implementation of the paper [CycleMLP: A MLP-like Architecture for Dense Prediction](https://arxiv.org/abs/2107.10224).
## Updates
- (29/01/2022) CycleMLP is accepted by ICLR 2022 as an **oral presentation**.:fire::fire::fire:
- (22/07/2021) Initial release.
## Model Zoo
We provide CycleMLP models pretrained on ImageNet 2012.
| Model | Parameters | FLOPs | Top 1 Acc. | Download |
| :------------------- | :--------- | :------- | :--------- | :------- |
| CycleMLP-B1 | 15M | 2.1G | 78.9% |[model](https://github.com/ShoufaChen/CycleMLP/releases/download/v0.1/CycleMLP_B1.pth)|
| CycleMLP-B2 | 27M | 3.9G | 81.6% |[model](https://github.com/ShoufaChen/CycleMLP/releases/download/v0.1/CycleMLP_B2.pth)|
| CycleMLP-B3 | 38M | 6.9G | 82.4% |[model](https://github.com/ShoufaChen/CycleMLP/releases/download/v0.1/CycleMLP_B3.pth)|
| CycleMLP-B4 | 52M | 10.1G | 83.0% |[model](https://github.com/ShoufaChen/CycleMLP/releases/download/v0.1/CycleMLP_B4.pth)|
| CycleMLP-B5 | 76M | 12.3G | 83.2% |[model](https://github.com/ShoufaChen/CycleMLP/releases/download/v0.1/CycleMLP_B5.pth)|
| CycleMLP-T | 28M | 4.4G | 81.3% |[model](https://github.com/ShoufaChen/CycleMLP/releases/download/v0.1/CycleMLP_tiny.pth)|
| CycleMLP-S | 50M | 8.5G | 82.9% |[model](https://github.com/ShoufaChen/CycleMLP/releases/download/v0.1/CycleMLP_small.pth)|
| CycleMLP-B | 88M | 15.2G | 83.4% |[model](https://github.com/ShoufaChen/CycleMLP/releases/download/v0.1/CycleMLP_base.pth)|
## Usage
### Install
- PyTorch 1.7.0+ and torchvision 0.8.1+
- [timm](https://github.com/rwightman/pytorch-image-models/tree/c2ba229d995c33aaaf20e00a5686b4dc857044be):
```
pip install 'git+https://github.com/rwightman/pytorch-image-models@c2ba229d995c33aaaf20e00a5686b4dc857044be'
or
git clone https://github.com/rwightman/pytorch-image-models
cd pytorch-image-models
git checkout c2ba229d995c33aaaf20e00a5686b4dc857044be
pip install -e .
```
- fvcore (optional, for FLOPs calculation)
- mmcv, mmdetection, mmsegmentation (optional)
### Data preparation
Download and extract ImageNet train and val images from http://image-net.org/.
The directory structure is:
```
│path/to/imagenet/
├──train/
│ ├── n01440764
│ │ ├── n01440764_10026.JPEG
│ │ ├── n01440764_10027.JPEG
│ │ ├── ......
│ ├── ......
├──val/
│ ├── n01440764
│ │ ├── ILSVRC2012_val_00000293.JPEG
│ │ ├── ILSVRC2012_val_00002138.JPEG
│ │ ├── ......
│ ├── ......
```
### Evaluation
To evaluate a pre-trained CycleMLP-B5 on ImageNet val with a single GPU run:
```
python main.py --eval --model CycleMLP_B5 --resume path/to/CycleMLP_B5.pth --data-path /path/to/imagenet
```
### Training
To train CycleMLP-B5 on ImageNet on a single node with 8 gpus for 300 epochs run:
```
python -m torch.distributed.launch --nproc_per_node=8 --use_env main.py --model CycleMLP_B5 --batch-size 128 --data-path /path/to/imagenet --output_dir /path/to/save
```
### Acknowledgement
This code is based on [DeiT](https://github.com/facebookresearch/deit) and [pytorch-image-models](https://github.com/rwightman/pytorch-image-models). Thanks for their wonderful works
## Citing
```bibtex
@inproceedings{
chen2022cyclemlp,
title={Cycle{MLP}: A {MLP}-like Architecture for Dense Prediction},
author={Shoufa Chen and Enze Xie and Chongjian GE and Runjian Chen and Ding Liang and Ping Luo},
booktitle={International Conference on Learning Representations},
year={2022},
url={https://openreview.net/forum?id=NMEceG4v69Y}
}
```
## License
CycleMLP is released under MIT License.