- 2 Convolutional Layers - 2 Fully Connected Layers
- Training Time - Approx 5min - 99% Accuracy
This repo contains 4753 (images) dataset
- Python 3
- Pip 3
pip3 install -r requirements.txt
python3 dataset.py
This command will generate two data files
x.pickle
y.pickle
python3 model.py
- Go to
dist
directory
cd dist
-
test
directory insidedist
directory contains some real captcha -
running
test.py
with an arg with directory will read the image and rename the file to actual captcha
python3 test.py test