Awesome

DN4 in PyTorch (2023 Version)

We provide a PyTorch implementation of DN4 for few-shot learning. If you use this code, please cite:

Revisiting Local Descriptor based Image-to-Class Measure for Few-shot Learning.<br> Wenbin Li, Lei Wang, Jinglin Xu, Jing Huo, Yang Gao and Jiebo Luo. In CVPR 2019.<br> <img src='DN4_2019_Version/imgs/Flowchart.bmp' width=600/>

Prerequisites

Linux
Python 3.8
Pytorch 1.7.0
GPU + CUDA CuDNN
pillow, torchvision, scipy, numpy

Getting Started

Installation

Clone this repo:

git clone https://github.com/WenbinLee/DN4.git
cd DN4

Install PyTorch 1.7.0 and other dependencies.

Datasets

Caltech-UCSD Birds-200-2011, Standford Cars, Standford Dogs, miniImageNet and tieredImageNet are available at Google Drive and 百度网盘(提取码：yr1w).

miniImageNet Few-shot Classification

Train a 5-way 1-shot model based on Conv64:

python Train_DN4.py --dataset_dir ./path/to/miniImageNet --data_name miniImageNet --encoder_model Conv64F_Local --way_num 5 --shot_num 1

Train a 5-way 1-shot model based on ResNet12:

python Train_DN4.py --dataset_dir ./path/to/miniImageNet --data_name miniImageNet --encoder_model ResNet12 --way_num 5 --shot_num 1

Test the model (specify the dataset_dir, encoder_model, and data_name first):

python Test_DN4.py --resume ./results/SGD_Cosine_Lr0.05_DN4_Conv64F_Local_Epoch_30_miniImageNet_84_84_5Way_1Shot/ --encoder_model Conv64F_Local

Latest results on miniImageNet (2023)

(Compared to the originally reported results in the paper. * denotes that ResNet256F is used.)

<table> <tr> <td rowspan="2">Method</td> <td rowspan="2">Backbone</td> <td colspan="2">5-way 1-shot</td> <td colspan="2">5-way 5-shot</td> </tr> <tr> <td>2019 Version</td> <td>2023 Version</td> <td>2019 Version</td> <td>2023 Version</td> </tr> <tr> <td rowspan="2">DN4</td> <td> Conv64F_Local </td> <td> 51.24 </td> <td> 51.97 </td> <td> 71.02 </td> <td> 73.19 </td> </tr> <tr> <td> ResNet12 </td> <td> 54.37* </td> <td> 61.23 </td> <td> 74.44* </td> <td> 75.66 </td> </tr> </table>

The results on the miniImageNet dataset reported in the orinigal paper: <img src='DN4_2019_Version/imgs/Results_miniImageNet2.bmp' align="center" width=710>

Citation

If you use this code for your research, please cite our paper.

@inproceedings{DN4_CVPR_2019,
  author       = {Wenbin Li and
                  Lei Wang and
                  Jinglin Xu and
                  Jing Huo and
                  Yang Gao and
                  Jiebo Luo},
  title        = {Revisiting Local Descriptor Based Image-To-Class Measure for Few-Shot Learning},
  booktitle    = {{IEEE} Conference on Computer Vision and Pattern Recognition (CVPR)},
  pages        = {7260--7268},
  year         = {2019}
}