Test-Agnostic Long-Tailed Recognition

August 14, 2025 · View on GitHub

This repository is the official Pytorch implementation of Breaking Long-Tailed Learning Bottlenecks: A Controllable Paradigm with Hypernetwork-Generated Diverse Experts (NeurIPS 2024).

1. Requirements

To install requirements:

pip install -r requirements.txt

2. Datasets

(1) Four bechmark datasets

Please download these datasets and put them to the /data file.
ImageNet-LT and Places-LT can be found at here.
iNaturalist data should be the 2018 version from here.
CIFAR-100 will be downloaded automatically with the dataloader.

data
├── ImageNet_LT
│   ├── test
│   ├── train
│   └── val
├── CIFAR100
│   └── cifar-100-python
├── Place365
│   ├── data_256
│   ├── test_256
│   └── val_256
└── iNaturalist 
    ├── test2018
    └── train_val2018

(2) Txt files

We provide txt files for test-agnostic long-tailed recognition for ImageNet-LT, Places-LT and iNaturalist 2018. CIFAR-100 will be generated automatically with the code.
For iNaturalist 2018, please unzip the iNaturalist_train.zip.

data_txt
├── ImageNet_LT
│   ├── ImageNet_LT_backward2.txt
│   ├── ImageNet_LT_backward5.txt
│   ├── ImageNet_LT_backward10.txt
│   ├── ImageNet_LT_backward25.txt
│   ├── ImageNet_LT_backward50.txt
│   ├── ImageNet_LT_forward2.txt
│   ├── ImageNet_LT_forward5.txt
│   ├── ImageNet_LT_forward10.txt
│   ├── ImageNet_LT_forward25.txt
│   ├── ImageNet_LT_forward50.txt
│   ├── ImageNet_LT_test.txt
│   ├── ImageNet_LT_train.txt
│   ├── ImageNet_LT_uniform.txt
│   └── ImageNet_LT_val.txt
├── Places_LT_v2
│   ├── Places_LT_backward2.txt
│   ├── Places_LT_backward5.txt
│   ├── Places_LT_backward10.txt
│   ├── Places_LT_backward25.txt
│   ├── Places_LT_backward50.txt
│   ├── Places_LT_forward2.txt
│   ├── Places_LT_forward5.txt
│   ├── Places_LT_forward10.txt
│   ├── Places_LT_forward25.txt
│   ├── Places_LT_forward50.txt
│   ├── Places_LT_test.txt
│   ├── Places_LT_train.txt
│   ├── Places_LT_uniform.txt
│   └── Places_LT_val.txt
└── iNaturalist18
    ├── iNaturalist18_backward2.txt
    ├── iNaturalist18_backward3.txt
    ├── iNaturalist18_forward2.txt
    ├── iNaturalist18_forward3.txt
    ├── iNaturalist18_train.txt
    ├── iNaturalist18_uniform.txt
    └── iNaturalist18_val.txt

3. Script

(1) CIFAR100-LT

Training

To train the expertise-diverse model, run this command:

python train.py -configs/mixup/standard_training/config_cifar100_ir100_bs-experts.json

One can change the imbalance ratio from 100 to 10/50 by changing the config file.

Evaluate

To evaluate expertise-diverse model on the uniform test class distribution, run:

python test.py -r checkpoint_path

To evaluate expertise-diverse model on agnostic test class distributions, run:

python test_all_cifar.py -r checkpoint_path

Test-time training

To test-time train the expertise-diverse model for agnostic test class distributions, run:

python test_training_cifar.py -c configs/mixup/standard_training/config_cifar100_ir100_bs-experts.json -r checkpoint_path

One can change the imbalance ratio from 100 to 10/50 by changing the config file.

(2) ImageNet-LT

Training

To train the expertise-diverse model, run this command:

python train.py -c configs/mixup/standard_training/config_imagenet_lt_resnet50_bs-experts.json

Evaluate

To evaluate expertise-diverse model on the uniform test class distribution, run:

python test.py -r checkpoint_path

To evaluate expertise-diverse model on agnostic test class distributions, run:

python test_all_imagenet.py -r checkpoint_path

Test-time training

To test-time train the expertise-diverse model for agnostic test class distributions, run:

python test_training_imagenet.py -c configs/mixup/standard_training/config_imagenet_lt_resnet50_bs-experts.json -r checkpoint_path

(3) Places-LT

Training

To train the expertise-diverse model, run this command:

python train.py -c configs/config_places_lt_resnet152_sade.json

Evaluate

To evaluate expertise-diverse model on the uniform test class distribution, run:

python test_places.py -r checkpoint_path

To evaluate expertise-diverse model on agnostic test class distributions, run:

python test_all_places.py -r checkpoint_path

Test-time training

To test-time train the expertise-diverse model for agnostic test class distributions, run:

python test_training_places.py -c configs/test_time_places_lt_resnet152_sade.json -r checkpoint_path

(4) iNaturalist 2018

Training

To train the expertise-diverse model, run this command:

python train.py -c configs/config_iNaturalist_resnet50_sade.json

Evaluate

To evaluate expertise-diverse model on the uniform test class distribution, run:

python test.py -r checkpoint_path

To evaluate expertise-diverse model on agnostic test class distributions, run:

python test_all_inat.py -r checkpoint_path

Test-time training

To test-time train the expertise-diverse model for agnostic test class distributions, run:

python test_training_inat.py -c configs/test_time_iNaturalist_resnet50_sade.json -r checkpoint_path

4. Citation

If you find our work inspiring or use our codebase in your research, please cite our work.

@article{zhao2024breaking,
  title={Breaking Long-Tailed Learning Bottlenecks: A Controllable Paradigm with Hypernetwork-Generated Diverse Experts},
  author={Zhao, Zhe and Wen, HaiBin and Wang, Zikang and Wang, Pengkun and Wang, Fanfu and Lai, Song and Zhang, Qingfu and Wang, Yang},
  journal={Advances in Neural Information Processing Systems},
  volume={37},
  pages={7493--7520},
  year={2024}
}

5. Acknowledgements

This is a project based on this pytorch template.

The mutli-expert framework are based on RIDE. The data generation of agnostic test class distributions takes references from LADE.