Breast cancer is one of the leading causes of cancer death in women all over the world. The emergence of digital imaging and computational aids in medicine has improved the diagnostic accuracy of breast cancer and reduced the workload of pathologists. Convolutional Neural Networks (CNNs) have recently emerged as a favored deep learning technique for breast cancer detection and classification. This paper presents a comparison of various deep convolutional neural networks (CNN), EfficientNet architectures (B1-B7), VGG19, ResNet50, DenseNet169, and InceptionV3 architectures for the classification of histopathology images of breast cancer. All architectures are tested on a publicly accessible histopathology image dataset. To minimize overfitting, data augmentation techniques are also used during training CNN models. According to the findings of the investigation, the EfficientNet-B6 model had a validation accuracy of 96.9% and a validation loss of 0.0898 in comparison to other tested models.