Skip to content

Instantly share code, notes, and snippets.

What would you like to do?
#!/usr/bin/env python
# -*- coding: utf-8 -*-
# this is a quick implementation of
# BUT! This is kind of dirty. Lots of hard coding. 
import numpy as np
import math
from chainer import cuda, Function, FunctionSet, gradient_check, Variable, optimizers
import chainer.functions as Fu
from chainer.functions import caffe
import chainer
import matplotlib.pyplot as plt
from scipy.misc import imread, imresize, imsave
use_gpu = False
def readimage(filename):
img = imread(filename)
img = imresize(img,[224, 224])
img = np.transpose(img,(2,0,1))
img = img.reshape((1,3,224,224))
p_data = np.ascontiguousarray(img,dtype=np.float32)
return Variable(cuda.to_gpu(p_data) if use_gpu else p_data)
def reshape2(conv1_1):[1][2]*[3]
return chainer.functions.reshape(conv1_1,(k,pixels))
# save the image x
def save_x(img,filename='output.png'):
img = img.reshape((3,224,224))
img = np.transpose(img,(1,2,0))
def forward(x, p, a):
conv1_1, conv2_1, conv3_1, conv4_1,conv5_1, = func(inputs={'data': x}, outputs=['conv1_1', 'conv2_1', 'conv3_1', 'conv4_1', 'conv5_1'])
conv1_1F,conv2_1F, conv3_1F, conv4_1F,conv5_1F, = [ reshape2(x) for x in [conv1_1,conv2_1, conv3_1, conv4_1,conv5_1]]
conv1_1G,conv2_1G, conv3_1G, conv4_1G,conv5_1G, = [ Fu.matmul(x, x, transa=False, transb=True) for x in [conv1_1F,conv2_1F, conv3_1F, conv4_1F,conv5_1F]]
conv1_1,conv2_1, conv3_1, conv4_1,conv5_1, = func(inputs={'data': p}, outputs=['conv1_1', 'conv2_1', 'conv3_1', 'conv4_1', 'conv5_1'])
conv1_1P,conv2_1P, conv3_1P, conv4_1P,conv5_1P, = [ reshape2(x) for x in [conv1_1,conv2_1, conv3_1, conv4_1,conv5_1]]
L_content = Fu.mean_squared_error(conv4_1F,conv4_1P)/2
conv1_1,conv2_1, conv3_1, conv4_1,conv5_1, = func(inputs={'data': a}, outputs=['conv1_1', 'conv2_1', 'conv3_1', 'conv4_1', 'conv5_1'])
conv1_1A0,conv2_1A0, conv3_1A0, conv4_1A0,conv5_1A0, = [ reshape2(x) for x in [conv1_1,conv2_1, conv3_1, conv4_1,conv5_1]]
conv1_1A,conv2_1A, conv3_1A, conv4_1A,conv5_1A, = [ Fu.matmul(x, x, transa=False, transb=True) for x in [conv1_1A0,conv2_1A0, conv3_1A0, conv4_1A0,conv5_1A0]]
#caution! the deviding number is hard coding!
#this part is correspnding to equation (4) in the original paper
#to check the current N and M, run the following
#[ for x in [conv1_1F,conv2_1F, conv3_1F, conv4_1F,conv5_1F]]
L_style = (Fu.mean_squared_error(conv1_1G,conv1_1A)/(4*64*64*50176*50176)
+ Fu.mean_squared_error(conv2_1G,conv2_1A)/(4*128**128*12544*12544)
+ Fu.mean_squared_error(conv3_1G,conv3_1A)/(4*256*256*3136*3136)
+ Fu.mean_squared_error(conv4_1G,conv4_1A)/(4*512*512*784*784)\
)/4 # this is equal weighting of E_l
ratio = 0.0001 #alpha/beta
loss = ratio*L_content + L_style
return loss
if use_gpu:
print('initializing gpu')
cuda.init(0)# is GPU ID!!
print('loading images')
p=readimage('input/tuebingen.jpg')#read a content image
a=readimage('input/starry_night.jpg')#read a style image
x = Variable(cuda.to_gpu(x_data) if use_gpu else x_data)
x = readimage('input/tuebingen.jpg') # if you want to start from a exsiting image
print('loading network')
#download a pretraind caffe model from here:
func = caffe.CaffeFunction('VGG_ILSVRC_19_layers.caffemodel')#it takes some time.
if use_gpu:
print('moving net to gpu')
#optimize x(=image) with adam
#note we use numpy for optimization
v=np.zeros_like(cuda.to_cpu( if use_gpu else np.zeros_like(
print('starting gradient descent')
for epoch in xrange(10000):
grad = cuda.to_cpu(x.grad.copy()) if use_gpu else x.grad.copy()
t +=1
m = beta1*m + (1-beta1)*grad
v = beta2*v + (1-beta2)*(grad*grad)
v_hat=v/(1-np.power(beta2, t))
if use_gpu: -= cuda.to_gpu( alpha * m_hat / (np.sqrt(v_hat) + eps) )#back it to cuda
else: -= ( alpha * m_hat / (np.sqrt(v_hat) + eps) )
with open(savedir+'/log.txt', 'a') as f:
savename = savedir+'/frames'+str(epoch)+'.png'
if use_gpu:
save_x(, savename)
# #optimize x(=image) with momment
# momentum= 0.9
# lr=100
# v=np.zeros_like(
# for epoch in xrange(10000):
# loss=forward(x,p,a)
# loss.backward()
# grad=x.grad.copy()
# v *= momentum
# v -= lr * grad
# += v
# with open(savedir+'/log.txt', 'a') as f:
# f.write(str(epoch+315)+','+str(','+str(np.linalg.norm(x.grad))+'\n')
# savename = savedir+'/imge'+str(epoch+315)+'.png'
# save_x(,savename)

This comment has been minimized.

Copy link

@apple2373 apple2373 commented Sep 6, 2015

Hi, I refined the code. According to your advise.Thanks!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment