Style images with convolutional neural nets with online parameters tuning. "Static" realization based on Stanford cs231n/assignment3/ StyleTransfer with Tensorflow notebook.
Run with python main.py from terminal.
Main command line arguments:
--default-style-decay - default value for style importance.
--default-content-weight - default value for content importance.
--content-max-side - max side of content image.
--style-max-side - max side of style image.
--save-each - step size for saving progress.
--update-each - step size for updating result.
| Content | Style | Progress |
|---|---|---|
 |
 |
|
 |
 |
 |
| |
|
 |