Residual Attention Networks (RANs) are a class of Convolutional Neural Networks (CNNs) that integrate attention mechanisms into deep architectures. RANs employ stacked attention modules to generate attention-aware features from low-resolution images and map them back to the original feature maps. This synergistic approach leverages the strengths of CNNs and attention mechanisms, leading to enhanced image classification performance. In this study, we propose an improved RAN architecture by replacing the single-type attention mechanism with a hybrid attention mechanism. The use of two distinct hybrid attention mechanisms is anticipated to further improve the CNN-based model's performance.