An api for developing cartoon-style images with CartoonGAN (CVPR 2018), powered by TensorFlow 2.0. Deployed using Flask and Flask-RESTful.
Clone the repository.
$ git clone https://github.com/Lancasterg/cartoon-gan-flask-api.git
Install the requirements (cpu).
$ python -m virtualenv cartoon-gan-venv // create a venv
$ source activate cartoon-gan-venv // init venv
$ pip install -r requirements-server.txt // install requirements for flask server
$ pip install -r requirements-cpu.txt // install requirements for ml
Start the API in development mode.
$ python flask_api/app.py
Or run using Docker
$ bash docker-build-dev.sh
$ docker run -d -p 5000:5000 cartoon-gan-api:latest
Test the API.
$ bash shell_scripts/hit_api.sh
Run the cartoonize.py
command to test the GAN.
python cartoonize.py \
--input_dir input_images \
--output_dir output_images \
--styles shinkai hayao \
--comparison_view horizontal
- Thanks to the author
[Chen et al., CVPR18]
who published this great work - Massive thank you to the authors of the tensorflow implementation of CartoonGan-tensorflow
- CartoonGAN-Test-Pytorch-Torch where we extracted pretrained Pytorch model weights for TensorFlow usage