India is the world's second-largest producer of potatoes, demonstrating the country's importance in the agricultural sector. In order to develop a sustainable agricultural system, it is essential to perform relevant research, especially in light of the recent advances in farming technology and the application of artificial intelligence in the detection of plant diseases. Early blight and Late blight have a substantial effect on potato yield and quality, and detecting these leaf diseases by hand can be laborious and time consuming. Due to the complexity involved, computerized and precise identification of these problems during the germination phase can help increase potato crop yield. Various models have been put up in the past to identify various plant diseases. The model offered in this study makes use of pre-trained models, such as Painter's embedding method, to generate results that are more accurate and to extract the required properties from the dataset. After running the results through various classifiers, it became clear that Support Vector Machine (SVM) - Polynomial Kernel was the clear winner, with an accuracy of 99.48% throughout the whole test dataset.