This repo is the duplicated version of the official implementation of our NeurIPS2022 paper "Scaling & Shifting Your Features: A New Baseline for Efficient Model Tuning" (arXiv).

• Clone this repo:

git clone https://github.com/dercaft/SSF.git
cd SSF

• Create a conda virtual environment and activate it:

conda create -n ssf -f ssfn.yaml -y
conda activate ssf

在项目路径下，新建DATA_PATH.sh:

touch DATA_PATH.sh

DATA_PATH.sh内应是数据集路径以及要输出的log,checkpoint路径

export VTAB_PATH="<dataset_abspath>"
export OUTPUT_PATH="<output_abspath>"

每个新建的训练脚本sh文件，开头都要 source DATA_PATH.sh!!!

MAC203服务器上有数据集，可以直接使用，不需要下载。

数据集路径: /data/VTAB/vtab-1k

输出路径为: ${OUTPUT_PATH}

MAC235服务器上有数据集，可以直接使用，不需要下载。

数据集路径: /media/disk1/VTAB/vtab-1k

输出路径为: /media/disk1/wyh/SSF

如果要自己下载，按以下步骤：

• FGVC & vtab-1k

You can follow VPT to download them.

Since the original vtab dataset is processed with tensorflow scripts and the processing of some datasets is tricky, we also upload the extracted vtab-1k dataset in onedrive for your convenience. You can download from here and then use them with our vtab.py directly. (Note that the license is in vtab dataset).

• CIFAR-100

wget https://www.cs.toronto.edu/~kriz/cifar-100-python.tar.gz

• For ImageNet-1K, download it from http://image-net.org/, and move validation images to labeled sub-folders. The file structure should look like:

  $ tree data
  imagenet
  ├── train
  │   ├── class1
  │   │   ├── img1.jpeg
  │   │   ├── img2.jpeg
  │   │   └── ...
  │   ├── class2
  │   │   ├── img3.jpeg
  │   │   └── ...
  │   └── ...
  └── val
      ├── class1
      │   ├── img4.jpeg
      │   ├── img5.jpeg
      │   └── ...
      ├── class2
      │   ├── img6.jpeg
      │   └── ...
      └── ...
  

• Robustness & OOD datasets

Prepare ImageNet-A, ImageNet-R and ImageNet-C for evaluation.

  _logger.info(
      '{0}: [{1:>4d}/{2}]  '
      'Time: {batch_time.val:.3f} ({batch_time.avg:.3f})  '
      'Loss: {loss.val:>7.4f} ({loss.avg:>6.4f})  '
      'Acc@1: {top1.val:>7.4f} ({top1.avg:>7.4f})  '
      'Acc@5: {top5.val:>7.4f} ({top5.avg:>7.4f})'.format(
          log_name, batch_idx, last_idx, batch_time=batch_time_m,
          loss=losses_m, top1=top1_m, top5=top5_m))

这里每个 Acc@1/5: <log输出时，最后一个batch的准确率> (<整体平均的准确率>)

所以会出现前面acc都高，但是括号里的acc比前面都低的情况。

• For pre-trained ViT-B/16, Swin-B, and ConvNext-B models on ImageNet-21K, the model weights will be automatically downloaded when you fine-tune a pre-trained model via SSF. You can also manually download them from ViT,Swin Transformer, and ConvNext.

• For pre-trained AS-MLP-B model on ImageNet-1K, you can manually download them from AS-MLP.

To fine-tune a pre-trained ViT model via SSF on CIFAR-100 or ImageNet-1K, run:

bash train_scripts/vit/cifar_100/train_ssf.sh

or

bash train_scripts/vit/imagenet_1k/train_ssf.sh

You can also find the similar scripts for Swin, ConvNext, and AS-MLP models. You can easily reproduce our results. Enjoy!

To evaluate the performance of fine-tuned model via SSF on Robustness & OOD, run:

bash train_scripts/vit/imagenet_a(r, c)/eval_ssf.sh

If this project is helpful for you, you can cite our paper:

@InProceedings{Lian_2022_SSF,
  title={Scaling \& Shifting Your Features: A New Baseline for Efficient Model Tuning},
  author={Lian, Dongze and Zhou, Daquan and Feng, Jiashi and Wang, Xinchao},
  booktitle={Advances in Neural Information Processing Systems (NeurIPS)},
  year={2022}
}

The code is built upon timm. The processing of the vtab-1k dataset refers to vpt, vtab github repo, and NOAH.