# ImageNet Optimization

Hiera is a hierarchical vision Transformer model that is fast, powerful, and concise. It surpasses state-of-the-art performance in a wide range of image and video tasks while significantly improving runtime speed.

A vision-language model initialized from OpenAI CLIP, adversarially fine-tuned on ImageNet with enhanced robustness features

Vision-language model initialized with OpenAI CLIP, enhanced robustness through unsupervised adversarial fine-tuning

The Hybrid Vision Transformer (ViT) model combines convolutional networks and Transformer architectures for image classification tasks, excelling on ImageNet.

ConvNeXT is a pure convolutional model inspired by vision Transformers, trained on the ImageNet-1k dataset at 224x224 resolution.

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 vision transformers, trained on the ImageNet-1k dataset, outperforming traditional vision transformers.