Requirements

Code is written in python3, so you should have it

• Install cuda
• Install cudnn
• Install opencv for python
• Install mxnet

pip3 install opencv-python
pip3 install mxnet-cu80 (CUDA 8.0)
git clone https://github.com/denemmy/captcha-recognizer.git

How to run

After you have cloned the repository, you can test model running the script below.

cd deploy
python3 deploy.py --gpu 0 --input <directory with images>

This script create results.txt file in the input directory with predicted labels for each image.