Uformer is a Transformer-based architecture for image restoration, combining a hierarchical encoder-decoder structure with a novel Locally-enhanced Window (LeWin) Transformer block and a learnable multi-scale restoration modulator. The LeWin block uses non-overlapping window-based self-attention to reduce computational complexity while capturing local context, and the modulator adjusts features in multiple layers to enhance restoration quality. Uformer achieves state-of-the-art performance on tasks like image denoising, motion deblurring, defocus deblurring, and deraining, outperforming existing methods in terms of PSNR and SSIM metrics. It is efficient, with a lightweight design that introduces minimal additional parameters and computational cost. The model is evaluated on multiple datasets and shows strong generalization across different image restoration tasks. The paper also discusses the effectiveness of the proposed components through extensive experiments and ablation studies, demonstrating the superiority of Uformer in capturing both local and global dependencies for image restoration.Uformer is a Transformer-based architecture for image restoration, combining a hierarchical encoder-decoder structure with a novel Locally-enhanced Window (LeWin) Transformer block and a learnable multi-scale restoration modulator. The LeWin block uses non-overlapping window-based self-attention to reduce computational complexity while capturing local context, and the modulator adjusts features in multiple layers to enhance restoration quality. Uformer achieves state-of-the-art performance on tasks like image denoising, motion deblurring, defocus deblurring, and deraining, outperforming existing methods in terms of PSNR and SSIM metrics. It is efficient, with a lightweight design that introduces minimal additional parameters and computational cost. The model is evaluated on multiple datasets and shows strong generalization across different image restoration tasks. The paper also discusses the effectiveness of the proposed components through extensive experiments and ablation studies, demonstrating the superiority of Uformer in capturing both local and global dependencies for image restoration.