paper-with-me

Papers

Pay Attention to Small Weights

2025-06-26 · Chao Zhou, Tom Jacobs, Advait Gadhikar, Rebekka Burkholz

Finetuning large pretrained neural networks is known to be resource-intensive, both in terms of memory and computational cost. To mitigate this, a common approach is to restrict training to a subset of the model parameters. By analyzing the relationship between gradients and weights during finetuning, we observe a notable pattern: large gradients are often associated with small-magnitude weights. This correlation is more pronounced in finetuning settings than in training from scratch. Motivated by this observation, we propose NANOADAM, which dynamically updates only the small-magnitude weights during finetuning and offers several practical advantages: first, this criterion is gradient-free -- the parameter subset can be determined without gradient computation; second, it preserves large-magnitude weights, which are likely to encode critical features learned during pretraining, thereby reducing the risk of catastrophic forgetting; thirdly, it permits the use of larger learning rates and consistently leads to better generalization performance in experiments. We demonstrate this for both NLP and vision tasks.

📄 PDF Abstract BibTeX arXiv:2506.21374

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

Accumulated Trivial Attention Matters in Vision Transformers on Small Datasets

2022-10-22 · Xiangyu Chen, Qinghao Hu, Kaidong Li, Cuncong Zhong 외

Vision Transformers has demonstrated competitive performance on computer vision tasks benefiting from their ability to capture long-range dependencies with multi-head self-attention modules and multi-layer perceptron. Ho…

Only Large Weights (And Not Skip Connections) Can Prevent the Perils of Rank Collapse

2025-05-22 · Josh Alman, Zhao Song

Attention mechanisms lie at the heart of modern large language models (LLMs). Straightforward algorithms for forward and backward (gradient) computation take quadratic time, and a line of work initiated by [Alman and Son…

SLA: Beyond Sparsity in Diffusion Transformers via Fine-Tunable Sparse-Linear Attention

2025-09-28 · Jintao Zhang, Haoxu Wang, Kai Jiang, Shuo Yang 외 arxiv

In Diffusion Transformer (DiT) models, particularly for video generation, attention latency is a major bottleneck due to the long sequence length and the quadratic complexity. We find that attention weights can be separa…

Video Generation

MultiResolution Attention Extractor for Small Object Detection

2020-06-10 · Fan Zhang, Licheng Jiao, Lingling Li, Fang Liu 외

Small objects are difficult to detect because of their low resolution and small size. The existing small object detection methods mainly focus on data preprocessing or narrowing the differences between large and small ob…

Hard AttentionObjectobject-detectionObject Detection+1

On the Connection between Local Attention and Dynamic Depth-wise Convolution

2021-06-08 · ICLR 2022 4 · Qi Han, Zejia Fan, Qi Dai, Lei Sun 외

Vision Transformer (ViT) attains state-of-the-art performance in visual recognition, and the variant, Local Vision Transformer, makes further improvements. The major component in Local Vision Transformer, local attention…

object-detectionObject DetectionPositionSemantic Segmentation