# Pure Convolutional Architecture

ConvNeXT is a pure convolutional model inspired by vision Transformer designs, pre-trained on ImageNet-22k and fine-tuned on ImageNet-1k, outperforming traditional Transformers.

ConvNeXT is a pure convolutional model inspired by vision Transformers, trained on the ImageNet-1k dataset for image classification tasks.

ConvNeXT is a pure convolutional model inspired by the design of vision Transformers, trained on the ImageNet-22k dataset, with performance surpassing Transformers.

ConvNeXT is a pure convolutional model inspired by vision transformers, trained on the ImageNet-1k dataset, outperforming traditional vision transformers.

ConvNeXT is a pure convolutional model inspired by vision Transformer designs, pretrained on ImageNet-22k and fine-tuned on ImageNet-1k, outperforming Transformers.