In this work, we investigate a new method for monkeypox detection with the EfficientNetB3 model through transfer learning. In order to stop the spread of the monkeypox outbreak and give prompt medical attention, it is imperative that quick and precise diagnosis be made. The monkeypox virus is the source of this zoonotic disease. We used a specific dataset of monkeypox photos to fine-tune the model by utilizing the pre-trained EfficientNetB3 architecture. This approach makes use of EfficientNetB3's strong performance and efficiency, which successfully strikes a balance between computational cost and model complexity. In order to optimize the training procedure, we added a customized Keras callback. In order to avoid overfitting and guarantee ideal model convergence, this callback is made to dynamically modify the learning rate or stop training in response to validation loss. If training performance degrades, the custom callback helps to keep the optimal model weights by restoring them from the epoch with the lowest validation loss. This adaptive technique not only increases the model's robustness, but it also speeds up the training process by eliminating unneeded epochs. Extensive experiments were carried out to assess the model's performance. Our findings show that the EfficientNetB3 model, when paired with the custom Keras callback, produced a high accuracy rate and an amazing F1 score. These metrics demonstrate the model's ability to accurately detect monkeypox from image data. The suggested approach has substantial clinical application potential, since it provides a dependable tool for early identification of monkeypox. This can help to speed up medical reactions and limit the virus's spread. Future study will increase the dataset and investigate additional deep learning techniques to improve detection accuracy and model generalizability. This strategy demonstrates the synergy of advanced transfer learning models and adaptive training techniques, resulting in a potent solution for infectious disease identification