Vision Transformers (ViTs)
PulseAugur coverage of Vision Transformers (ViTs) — every cluster mentioning Vision Transformers (ViTs) across labs, papers, and developer communities, ranked by signal.
-
New FACTS method improves Vision Transformer compression efficiency
Researchers have developed a new method called FACTS (Fisher Approximation tailored to Compressing ViTs) to improve the efficiency of Vision Transformers (ViTs) through model compression. This technique utilizes Fisher-…
-
New OptiModNet model achieves state-of-the-art optic disc segmentation with high efficiency
Researchers have developed OptiModNet, a novel hybrid deep learning architecture designed for efficient and accurate segmentation of the optic disc and cup. This model combines U-Net and Transformer elements, incorporat…
-
New framework reveals geometric consistency is key for stable diffusion models
Researchers have developed a unified framework to analyze the impact of geometric transformations on diffusion model architectures like UNets, ViTs, and DiTs. By applying dihedral group elements to intermediate hidden s…
-
New FiPS framework compresses transformer models with minimal accuracy loss
Researchers have developed a new framework called Fine-grained Parameter Sharing (FiPS) to compress large transformer models. FiPS combines cross-block parameter sharing, low-rank factorization, and sparsity within a si…
-
New MDSE attack fools Spiking Neural Networks and traditional models
Researchers have developed a new adversarial attack method called Mixed Dynamic Spiking Estimation (MDSE) specifically for Spiking Neural Networks (SNNs). This attack demonstrates that the effectiveness of white-box adv…