This dataset provides images of handwritten words for OCR models to be trained on. The dataset contains 25,132 images with the labels embedded in the image names. For easy access to the labels a label.csv file is provided with the corresponding image name and the label.