colorizer
Overview
The python script color.py trains a Keras model that colorizes images of flowers.
You can run the model and log the results to Weights & Biases by running:
$ pip install -r requirements.txt
$ wandb login
$ python color.py
Submitting your results
You'll can submit your best runs to our benchmark. The user who submits the top result will receive a fully-paid scholarship to our Applied Deep Learning course in San Francisco (6 weeks, 2 evening hours per week, attendance in person required).
Evaluation
We use a perceptual distance metric on the validation set to rank results. This metric is generated by the color.py script.
You are welcome to use other ML frames besides Keras, but please make sure to log the same perceptual distance metric to wandb, so we can rank results.
Things to try:
- Fancier architectures
- Different loss functions
- Data augmentation
- More training data?