search

Hvass-Labs / TensorFlow-Tutorials

TensorFlow Tutorials with YouTube Videos
MIT License
9.28k stars 4.19k forks source link

Tutorial 14: Parameter "tile_size" is not passed down to proper function. #94

Closed d4ichi closed 5 years ago

d4ichi commented 5 years ago

14_DeepDream.ipynb

Function "optimize_image" is calling "tiled_gradient":

tiled_gradient(gradient=gradient, image=img)

But I believe this should be something like:

tiled_gradient(gradient=gradient, image=img, tile_size=tile_size)

No matter we set tile_size in "recursive_optimize", that parameter is not passed down to proper function, i think.

Could you check this out please. Thanks.

Hvass-Labs commented 5 years ago

It has been over 2 years since I did this so I honestly can't remember. But I looked at it briefly and it appears you might be right, that the tile_size parameter is not passed from the optimize_image() function when it calls tiled_gradient(). Could you try and change it and report back here what happens? Do you get an error if you change the tile_size?

Hvass-Labs commented 5 years ago

Thanks for reporting this. It has now been fixed. There does seem to be a difference in the images that are produced depending on the tile-size.

d4ichi commented 5 years ago

I'm sorry, it seems I missed notification, didn't realize your comment. Thank you so much for fixing this, also thank you again for this awesome repo.