The FAO estimates that 60 percent of the world’s population makes their liv- ing from agriculture. The rapid increase in the global populations demand for food is also quite fast. In this case, plant diseases pose a substantial threat to the agricultural industry. Therefore, deep learning algorithms are applied to spot them at an early stage as a move towards protecting farmers against such losses while increasing crop yield. We applied CNNs in developing a technique for identifying different diseases of cassava leaf which lead to low yields. We created a cost-effective model that will help farmers to save costs and special- ize in farming operations. Early diagnosis of these diseases is proposed by EfficientNet-B0, which may serve well since they provide a remedy for minor cases of cassava leaf illnesses. This may lead to better cassava crop health, and therefore more food security especially in some particularly vulnerable places.