There are 3 notebooks:
sampleDataset
is how the dataset was sampledtrain
where it is shown how the model was fine-tuned and trainedevaluate
provides code for evaluating the model
All of them are self-explanatory.
Inference is available via REST API, made using a simple Flask app. Run with:
python app.py
python3.8 -m venv venv
source venv/bin/activate
pip install git+https://github.com/tensorflow/examples.git
pip install -r requirements.txt
Mappilary Vistas v1.2, which contains 25 000 fully-annotated high-resolution images with 66 object categories.
pix2pix : Image-to-image translation with a conditional GAN.