SustAInML

A comprehensive benchmark and analysis of more than 45 transformer models for image classification to evaluate their efficiency, considering various performance metrics. We find the optimal architectures to use and uncover that model-scaling is more efficient than image scaling.

This paper introduces TaylorShift, a novel reformulation of the attention mechanism using Taylor softmax that enables computing full token-to-token interactions in linear time. We analytically and empirically determine the crossover points where employing TaylorShift becomes more efficient than traditional attention. TaylorShift outperforms the traditional transformer architecture in 4 out of 5 tasks.

We improve dataset distillation by distilling only a representative coreset.

We speed up diffusion classifiers by utilizing a label hierarchy and pruning unrelated paths.

We extend pretrained super-resolution models to larger images by using local-aware prompts.

We utilize the TaylorShift attention mechanism for global pixel-wise-attention in image super-resolution.