README.md

SADA

This is the pytorch version code of Few-Shot Learning with Visual Distribution Calibration and Cross-Modal Distribution Alignment in CVPR2023.

Content

• SADA
  • Content
  • Introduce of SADA
  • Datasets
  • Pretrained CLIP
  • Insturction

Introduce of SADA

SADA is introduced from《Few-Shot Learning with Visual Distribution Calibration and Cross-Modal Distribution Alignment》

Paper ：Runqi Wang, Hao Zheng, Xiaoyue Duan, Jianzhuang Liu, Yuning Lu, Tian Wang, Songcen Xu, Baochang Zhang. "Few-Shot Learning with Visual Distribution Calibration and Cross-Modal Distribution Alignment". In CVPR, 2023.

Datasets

The test Datasets：CIFAR10, Download link
Datasets size：10 classes and 32*32 pixels for each image.
Training set：50,000 images.
Testing set：10,000 images.

Pretrained CLIP

We use the pretrained CLIP model from here

Insturction

pip install -r requirements.txt
sh run.sh