Handwritten digit recognizer AI.
- Python 3.6+
- Anaconda 4+ (optional)
- Python modules from the
requirements.txt
Dowload and install the dependencies with the command:
$ python -m pip install -r requirements.txt
To train the model you should run:
$ python trainer.py
If you want to use the model in HTML, you should run porter model.pkl --js --pipe > "public/js/model.js"
.
After this, run python port.py
and the estimator is ready to be used.
To try the model.pkl
estimator, go to this project's github page,
or open the index.html
in the main directory and remove the /DigitRecognizer/
root from the public links, like
/DigitRecognizer/public/css/main.css
will be
public/css/main.css
after this you will be able to work with the index.html
file localy on your computer.
This project is open for any kind of contribution from anyone.
- Fork this repository
- Create a new branch (optional)
- Clone it
- Make your changes
- Upload them
- Make a pull request here
- Profit.
See the LICENSE file for details.