Add different kernels on gram matrix in style transfer algorithm

folder structure will be created on the run.

then after download this mat file - vgg197.mat which contains the weights of the vgg19 network. place it in the same location where neural_style_loop.py located (root directory)

you can change the sigma of the exponential kernel weights by changing in the neural_style_loop.py line 118.

range_sigma = [1e18, 1e19, 1e21, 1e22, 1e25, 1e30]

you can change the style weights by changing in the neural_style_loop.py line 119.

range_sw = [1e10, 1e15, 1e20, 1e25, 1e30, 1e40, 1e50]

then select the which you want to apply form neural_style_loop.py line 179

• 0 - dotproduct
• 1 - exponential
• 2 - mattern
• 3 - polynomial

in line 22 of neural_style_loop.py file you can change the number of Iternation to print the statics. `PRINT_ITERATIONS = 50'

after that run the code python neural_style_loop.py --content examples/content/1-content.jpg examples/content/2-content.jpg