CheXNet is a 121-layer convolutional neural network designed to detect pneumonia from chest X-rays, outperforming practicing radiologists in terms of F1 score. Trained on the ChestX-ray14 dataset, which contains over 100,000 frontal-view X-ray images labeled with 14 different thoracic diseases, CheXNet achieves an F1 score of 0.435, significantly higher than the average radiologist's score of 0.387. The model not only excels in pneumonia detection but also performs state-of-the-art on all 14 diseases in the dataset. CheXNet's performance is validated through comparisons with radiologists and through the use of Class Activation Maps (CAMs) to visualize the most indicative areas of the X-ray for each pathology. The study highlights the potential of automated chest X-ray analysis to improve healthcare delivery, particularly in regions with limited access to expert radiologists.CheXNet is a 121-layer convolutional neural network designed to detect pneumonia from chest X-rays, outperforming practicing radiologists in terms of F1 score. Trained on the ChestX-ray14 dataset, which contains over 100,000 frontal-view X-ray images labeled with 14 different thoracic diseases, CheXNet achieves an F1 score of 0.435, significantly higher than the average radiologist's score of 0.387. The model not only excels in pneumonia detection but also performs state-of-the-art on all 14 diseases in the dataset. CheXNet's performance is validated through comparisons with radiologists and through the use of Class Activation Maps (CAMs) to visualize the most indicative areas of the X-ray for each pathology. The study highlights the potential of automated chest X-ray analysis to improve healthcare delivery, particularly in regions with limited access to expert radiologists.