研究人员开发了 ViT-FREE 方法，可在无需重新训练的情况下提高 Vision Transformers (ViTs) 在人脸识别方面的效率。该方法允许从预训练 ViT 的中间层提前退出，从而降低计算成本，同时保持高精度。一种额外的微调策略 ViT-FREE_FT，通过仅用合成数据调整投影层，进一步优化了浅层退出的性能。 AI
影响 使得强大的 Vision Transformer 模型能够在资源受限的设备上更高效地部署，用于人脸识别任务。
排序理由 该集群包含一篇研究论文，详细介绍了一种提高现有模型效率的新方法。
AI 生成摘要 · Google Gemini · 来自 2 个来源。 我们如何撰写摘要 →
arXiv:2606.12023v1 Announce Type: new Abstract: Vision Transformers (ViTs) have gained significant attention in computer vision and shown strong potential for face recognition (FR). However, their high computational cost makes deployment on resource-constrained devices challengin…
Vision Transformers (ViTs) have gained significant attention in computer vision and shown strong potential for face recognition (FR). However, their high computational cost makes deployment on resource-constrained devices challenging, motivating the need for methods that balance …