Provides pre-trained language data files for the Tesseract OCR engine. Includes integerized LSTM neural network models that balance speed and accuracy for text extraction from images. Supports the modern neural net engine (–oem 1) as well as legacy models for backward compatibility.