ConvNeXt V2: Co-designing and Scaling ConvNets with Masked Autoencoders
In this paper, the authors propose a fully convolutional masked autoencoder framework 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 (such as MAE) and architectural improvement results in a new model family called ConvNeXt V2, which significantly improves the performance of pure ConvNets on various recognition benchmarks, including ImageNet classification, COCO detection, and ADE20K segmentation.[1]
Figure 1. Architecture of ConvNeXt V2 [1]
Our reproduced model performance on ImageNet-1K is reported as follows.
| Model | Context | Top-1 (%) | Top-5 (%) | Params (M) | Recipe | Download |
|---|---|---|---|---|---|---|
| convnextv2_tiny | D910x8-G | 82.43 | 95.98 | 28.64 | yaml | weights |
- Context: Training context denoted as {device}x{pieces}-{MS mode}, where mindspore mode can be G - graph mode or F - pynative mode with ms function. For example, D910x8-G is for training on 8 pieces of Ascend 910 NPU using graph mode.
- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K.
Please refer to the installation instruction in MindCV.
Please download the ImageNet-1K dataset for model training and validation.
- Distributed Training
It is easy to reproduce the reported results with the pre-defined training recipe. For distributed training on multiple Ascend 910 devices, please run
# distributed training on multiple GPU/Ascend devices
mpirun -n 8 python train.py --config configs/convnextv2/convnextv2_tiny_ascend.yaml --data_dir /path/to/imagenetIf the script is executed by the root user, the
--allow-run-as-rootparameter must be added tompirun.
Similarly, you can train the model on multiple GPU devices with the above mpirun command.
For detailed illustration of all hyper-parameters, please refer to config.py.
Note: As the global batch size (batch_size x num_devices) is an important hyper-parameter, it is recommended to keep the global batch size unchanged for reproduction or adjust the learning rate linearly to a new global batch size.
- Standalone Training
If you want to train or finetune the model on a smaller dataset without distributed training, please run:
# standalone training on a CPU/GPU/Ascend device
python train.py --config configs/convnextv2/convnextv2_tiny_ascend.yaml --data_dir /path/to/dataset --distribute FalseTo validate the accuracy of the trained model, you can use validate.py and parse the checkpoint path with --ckpt_path.
python validate.py -c configs/convnextv2/convnextv2_tiny_ascend.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckptPlease refer to the deployment tutorial in MindCV.
[1] Woo S, Debnath S, Hu R, et al. ConvNeXt V2: Co-designing and Scaling ConvNets with Masked Autoencoders[J]. arXiv preprint arXiv:2301.00808, 2023.