Disease identification in paddy crops is a critical aspect in food security and high yields in the world. The previous approaches towards identification of plant diseases are more dependent on the observation, which is time consuming and quite subjective in nature. In this paper, a new deep learning model, namely Swin Transformers, is introduced for the identification of paddy leaf disease with the help of data augmentation techniques. The model utilizes the Swin Transformer’s capability to extract features at different levels to categorize the images into 12 diseases and healthy classes such as bacterial leaf blight, blast, brown spot, and others. In addition, Auto Augment based on reinforcement learning is used to reduce overfitting to some extent by creating new training data with different and balanced distributions. For the training and validation of the model, the dataset consisted of 16,225 high-resolution images of paddy leaves in the fields of Tamil Nadu, India. The accuracy of the proposed model was 98.6% which is higher than the traditional CNN-based models in terms of accuracy and robustness. This work, therefore, presents an efficient way of automating the process of identifying diseases within paddy leaves using the latest deep learning techniques and data augmentation. The authors’ procedure does not only reduce the dangers associated with manual inspection but also provides a viable method of monitoring the state of crops in real-time and their spread. The highlights also reveal that AI has the potential of transforming the agricultural sector to improve food production and yields.