In typical pathology procedures, there often exists a divergence of opinions among pathologists leading to diagnostic uncertainties. To address this issue, the integration of deep learning methodologies is recommended to enhance decision-making consistency and operational efficiency. Presently, diseases such as Colon, Stomach, and Kidney cancers pose significant mortality risks. This research introduces a novel approach aimed at training a deep learning model capable of categorizing gastric, colon, and renal cancers using a unified model. The utilization of Whole Slide Images as input data for the Efficient-Net model, which has been pre-trained on the ImageNet dataset, forms the basis of this study. The model is fine-tuned through a transfer learning technique involving partial transfer learning. Various strategies have been proposed to accurately classify these pathology images individually through the application of partial transfer learning. This study aims to showcase the generalization capabilities of partial transfer learning in the classification of pathology images.