Skip to content
Switch branches/tags

Convolutional Recurrent Neural Network

This software implements the Convolutional Recurrent Neural Network (CRNN) in pytorch. Origin software could be found in crnn

Run demo

A demo program can be found in Before running the demo, download a pretrained model from Baidu Netdisk or Dropbox. This pretrained model is converted from auther offered one by tool. Put the downloaded model file crnn.pth into directory data/. Then launch the demo by:


The demo reads an example image and recognizes its text content.

Example image: Example Image

Expected output: loading pretrained model from ./data/crnn.pth a-----v--a-i-l-a-bb-l-ee-- => available


Train a new model

  1. Construct dataset following origin guide. If you want to train with variable length images (keep the origin ratio for example), please modify the tool/ and sort the image according to the text length.
  2. Execute python --adadelta --trainRoot {train_path} --valRoot {val_path} --cuda. Explore for details.