CondenseNetV2

This repository is the official Pytorch implementation for CVPR2021 paper "CondenseNet V2: Sparse Feature Reactivation for Deep Networks" by Le Yang*, Haojun Jiang*, Ruojin Cai, Yulin Wang, Shiji Song, Gao Huang and Qi Tian (* Authors contributed equally).

Update on 2021/04/14: Release the training code on ImageNet.

Reference

If you find our project useful in your research, please consider citing:

@inproceedings{yang2021condensenetv2,
  title={CondenseNet V2: Sparse Feature Reactivation for Deep Networks},
  author={Yang, Le and Jiang, Haojun and Cai, Ruojin and Wang, Yulin and Song, Shiji and Huang, Gao and Tian, Qi},
  booktitle={Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition},
  year={2021}
}

Introduction

Reusing features in deep networks through dense connectivity is an effective way to achieve high computational efficiency. The recent proposed CondenseNet has shown that this mechanism can be further improved if redundant features are removed. In this paper, we propose an alternative approach named sparse feature reactivation (SFR), aiming at actively increasing the utility of features for reusing. In the proposed network, named CondenseNetV2, each layer can simultaneously learn to 1) selectively reuse a set of most important features from preceding layers; and 2) actively update a set of preceding features to increase their utility for later layers. Our experiments show that the proposed models achieve promising performance on image classification (ImageNet and CIFAR) and object detection (MS COCO) in terms of both theoretical efficiency and practical speed.

DenseNet, CondenseNet and CondenseNetV2.
Comparison

Sparse feature reactivation.
SFR

Usage

Dependencies

Training

As an example, use the following command to train a CondenseNetV2-A/B/C on ImageNet

python -m torch.distributed.launch --nproc_per_node=8 train.py --model cdnv2_a/b/c 
  --batch-size 1024 --lr 0.4 --warmup-lr 0.1 --warmup-epochs 5 --opt sgd --sched cosine \
  --epochs 350 --weight-decay 4e-5 --aa rand-m9-mstd0.5 --remode pixel --reprob 0.2 \
  --data_url /PATH/TO/IMAGENET --train_url /PATH/TO/LOG_DIR

Evaluation

We take the ImageNet model trained above as an example.

To evaluate the non-converted trained model, use test.py to evaluate from a given checkpoint path:

python test.py --model cdnv2_a/b/c \
  --data_url /PATH/TO/IMAGENET -b 32 -j 8 \
  --train_url /PATH/TO/LOG_DIR \
  --evaluate_from /PATH/TO/MODEL_WEIGHT

To evaluate the converted trained model, use --model converted_cdnv2_a/b/c:

python test.py --model converted_cdnv2_a/b/c \
  --data_url /PATH/TO/IMAGENET -b 32 -j 8 \
  --train_url /PATH/TO/LOG_DIR \
  --evaluate_from /PATH/TO/MODEL_WEIGHT

Note that these models are still the large models after training. To convert the model to standard group-convolution version as described in the paper, use the convert_and_eval.py:

python convert_and_eval.py --model cdnv2_a/b/c \
  --data_url /PATH/TO/IMAGENET -b 64 -j 8 \
  --train_url /PATH/TO/LOG_DIR \
  --convert_from /PATH/TO/MODEL_WEIGHT

Results

Results on ImageNet

Model FLOPs Params Top-1 Error Tsinghua Cloud Google Drive
CondenseNetV2-A 46M 2.0M 35.6 Download Download
CondenseNetV2-B 146M 3.6M 28.1 Download Download
CondenseNetV2-C 309M 6.1M 24.1 Download Download

Results on COCO2017 Detection

The detection experiments are conducted based on the mmdetection repository. We simply replace the backbones of FasterRCNN and RetinaNet with our CondenseNetV2s.

Detection Framework Backbone Backbone FLOPs mAP
FasterRCNN ShuffleNetV2 0.5x 41M 22.1
FasterRCNN CondenseNetV2-A 46M 23.5
FasterRCNN ShuffleNetV2 1.0x 146M 27.4
FasterRCNN CondenseNetV2-B 146M 27.9
FasterRCNN MobileNet 1.0x 300M 30.6
FasterRCNN ShuffleNetV2 1.5x 299M 30.2
FasterRCNN CondenseNetV2-C 309M 31.4
RetinaNet MobileNet 1.0x 300M 29.7
RetinaNet ShuffleNetV2 1.5x 299M 29.1
RetinaNet CondenseNetV2-C 309M 31.7

Results on CIFAR

Model FLOPs Params CIFAR-10 CIFAR-100
CondenseNet-50 28.6M 0.22M 6.22 -
CondenseNet-74 51.9M 0.41M 5.28 -
CondenseNet-86 65.8M 0.52M 5.06 23.64
CondenseNet-98 81.3M 0.65M 4.83 -
CondenseNet-110 98.2M 0.79M 4.63 -
CondenseNet-122 116.7M 0.95M 4.48 -
CondenseNetV2-110 41M 0.48M 4.65 23.94
CondenseNetV2-146 62M 0.78M 4.35 22.52

Contacts

[email protected]

[email protected]

Any discussions or concerns are welcomed!

GitHub

https://github.com/jianghaojun/CondenseNetV2