Contains best (most accurate) trained LSTM models for Tesseract, an open source OCR engine. These models work with the LSTM OCR engine of Tesseract 4 and are distributed as trained model weights for optical character recognition tasks.