Quantization Aware Training for Grape Leaf Disease Detection

A major obstacle to agricultural output is grape leaf stress conditions.Conventional diagnosis techniques are time-consuming and necessitate domain knowledge.This study investigates the use of Quantization-Aware Training (QAT) to create an effective deep learning model for the classification of grape leaf diseases.The model is refined using a dataset supplied by the Indian Council of Agricultural Research (ICAR) using the ResNet50 architecture.Real-time illness diagnosis is made possible by the quantized TFLite model, which guarantees interoperability with edge devices such as IoT modules.What makes it perfect for deployment in settings with limited resources is high accuracy and a smaller model size.