Convnext V2

Code release for ConvNeXt V2 model

ConvNeXt V2
Official PyTorch Implementation

This repo contains the PyTorch version of 8 model definitions (Atto, Femto, Pico, Nano, Tiny, Base, Large, Huge), pre-training/fine-tuning code and pre-trained weights (converted from JAX weights trained on TPU) for our ConvNeXt V2 paper.

ConvNeXt V2: Co-designing and Scaling ConvNets with Masked Autoencoders
Sanghyun Woo, Shoubhik Debnath, Ronghang Hu, Xinlei Chen, Zhuang Liu, In So Kweon and Saining Xie

KAIST, Meta AI and New York University

We propose a fully convolutional masked autoencoder framework (FCMAE) and a new Global Response Normalization (GRN) layer that can be added to the ConvNeXt architecture to enhance inter-channel feature competition. This co-design of self-supervised learning techniques and architectural improvement results in a new model family called ConvNeXt V2, which significantly improves the performance of pure ConvNets on various recognition benchmarks. We also provide pre-trained ConvNeXt V2 models of various sizes.

Results and Pre-trained Models

ImageNet-1K FCMAE pre-trained weights (self-supervised)

name resolution #params model
ConvNeXt V2-A 224x224 3.7M model
ConvNeXt V2-F 224x224 5.2M model
ConvNeXt V2-P 224x224 9.1M model
ConvNeXt V2-N 224x224 15.6M model
ConvNeXt V2-T 224x224 28.6M model
ConvNeXt V2-B 224x224 89M model
ConvNeXt V2-L 224x224 198M model
ConvNeXt V2-H 224x224 660M model

ImageNet-1K fine-tuned models

name resolution [email protected] #params FLOPs model
ConvNeXt V2-A 224x224 76.7 3.7M 0.55G model
ConvNeXt V2-F 224x224 78.5 5.2M 0.78G model
ConvNeXt V2-P 224x224 80.3 9.1M 1.37G model
ConvNeXt V2-N 224x224 81.9 15.6M 2.45G model
ConvNeXt V2-T 224x224 83.0 28.6M 4.47G model
ConvNeXt V2-B 224x224 84.9 89M 15.4G model
ConvNeXt V2-L 224x224 85.8 198M 34.4G model
ConvNeXt V2-H 224x224 86.3 660M 115G model

ImageNet-22K fine-tuned models

name resolution [email protected] #params FLOPs model
ConvNeXt V2-N 224x224 82.1 15.6M 2.45G model
ConvNeXt V2-N 384x384 83.4 15.6M 7.21G model
ConvNeXt V2-T 224x224 83.9 28.6M 4.47G model
ConvNeXt V2-T 384x384 85.1 28.6M 13.1G model
ConvNeXt V2-B 224x224 86.8 89M 15.4G model
ConvNeXt V2-B 384x384 87.7 89M 45.2G model
ConvNeXt V2-L 224x224 87.3 198M 34.4G model
ConvNeXt V2-L 384x384 88.2 198M 101.1G model
ConvNeXt V2-H 384x384 88.7 660M 337.9G model
ConvNeXt V2-H 512x512 88.9 660M 600.8G model


Please check for installation instructions.


We provide example evaluation commands for ConvNeXt V2-Base:


python \
--model convnextv2_base \
--eval true \
--resume /path/to/checkpoint \
--input_size 224 \
--data_path /path/to/imagenet-1k \


python -m torch.distributed.launch --nproc_per_node=8 \
--model convnextv2_base \
--eval true \
--resume /path/to/checkpoint \
--input_size 224 \
--data_path /path/to/imagenet-1k \
  • For evaluating other model variants, change --model, --resume, --input_size accordingly. URLs for the pre-trained models can be found from the result tables.
  • Setting model-specific --drop_path is not strictly required in evaluation, as the DropPath module in timm behaves the same during evaluation; but it is required in training. See or our paper (appendix) for the values used for different models.


See for pre-training and fine-tuning instructions.


This repository borrows from timm, ConvNeXt and MAE.

We thank Ross Wightman for the initial design of the small-compute ConvNeXt model variants and the associated training recipe. We also appreciate the helpful discussions and feedback provided by Kaiming He.


This project is released under the MIT license except ImageNet pre-trained and fine-tuned models which are licensed under a CC-BY-NC. Please see the LICENSE file for more information.


If you find this repository helpful, please consider citing:

  title={ConvNeXt V2: Co-designing and Scaling ConvNets with Masked Autoencoders},
  author={Sanghyun Woo, Shoubhik Debnath, Ronghang Hu, Xinlei Chen, Zhuang Liu, In So Kweon and Saining Xie},
  journal={arXiv preprint arXiv:2301.00808},

Get A Weekly Email With Trending Projects For These Categories
No Spam. Unsubscribe easily at any time.