/
deepdream.py
156 lines (132 loc) · 5.67 KB
/
deepdream.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
# Source: Google Deepdream code @ https://github.com/google/deepdream/
# Slightly modified in order to be run inside the container as a script instead of an IPython Notebook
from cStringIO import StringIO
import numpy as np
import scipy.ndimage as nd
import PIL.Image
import os
from IPython.display import clear_output, Image, display
from google.protobuf import text_format
import time
import caffe
def showarray(a):
a = np.uint8(np.clip(a, 0, 255))
f = StringIO()
millis = int(round(time.time() * 1000))
filename = "/data/output/tmp/steps-%i.jpg" % millis
PIL.Image.fromarray(np.uint8(a)).save(filename)
input_file = os.getenv('INPUT', 'input.png')
iterations = os.getenv('ITER', 50)
try:
iterations = int(iterations)
except ValueError:
iterations = 50
scale = os.getenv('SCALE', 0.05)
try:
scale = float(scale)
except ValueError:
scale = 0.05
model_name = os.getenv('MODEL', 'inception_4c/output')
print "Processing file: " + input_file
img = np.float32(PIL.Image.open('/data/%s' % input_file))
model_path = '/caffe/models/bvlc_googlenet/' # substitute your path here
net_fn = model_path + 'deploy.prototxt'
param_fn = model_path + 'bvlc_googlenet.caffemodel'
# Patching model to be able to compute gradients.
# Note that you can also manually add "force_backward: true" line to "deploy.prototxt".
model = caffe.io.caffe_pb2.NetParameter()
text_format.Merge(open(net_fn).read(), model)
model.force_backward = True
open('tmp.prototxt', 'w').write(str(model))
net = caffe.Classifier('tmp.prototxt', param_fn,
mean = np.float32([104.0, 116.0, 122.0]), # ImageNet mean, training set dependent
channel_swap = (2,1,0)) # the reference model has channels in BGR order instead of RGB
# a couple of utility functions for converting to and from Caffe's input image layout
def preprocess(net, img):
return np.float32(np.rollaxis(img, 2)[::-1]) - net.transformer.mean['data']
def deprocess(net, img):
return np.dstack((img + net.transformer.mean['data'])[::-1])
def make_step(net, step_size=1.5, end='inception_4c/output', jitter=32, clip=True):
'''Basic gradient ascent step.'''
src = net.blobs['data'] # input image is stored in Net's 'data' blob
dst = net.blobs[end]
ox, oy = np.random.randint(-jitter, jitter+1, 2)
src.data[0] = np.roll(np.roll(src.data[0], ox, -1), oy, -2) # apply jitter shift
net.forward(end=end)
dst.diff[:] = dst.data # specify the optimization objective
net.backward(start=end)
g = src.diff[0]
# apply normalized ascent step to the input image
src.data[:] += step_size/np.abs(g).mean() * g
src.data[0] = np.roll(np.roll(src.data[0], -ox, -1), -oy, -2) # unshift image
if clip:
bias = net.transformer.mean['data']
src.data[:] = np.clip(src.data, -bias, 255-bias)
def deepdream(net, base_img, iter_n=10, octave_n=4, octave_scale=1.4, end='inception_4c/output', clip=True, **step_params):
# prepare base images for all octaves
octaves = [preprocess(net, base_img)]
for i in xrange(octave_n-1):
octaves.append(nd.zoom(octaves[-1], (1, 1.0/octave_scale,1.0/octave_scale), order=1))
src = net.blobs['data']
detail = np.zeros_like(octaves[-1]) # allocate image for network-produced details
for octave, octave_base in enumerate(octaves[::-1]):
h, w = octave_base.shape[-2:]
if octave > 0:
# upscale details from the previous octave
h1, w1 = detail.shape[-2:]
detail = nd.zoom(detail, (1, 1.0*h/h1,1.0*w/w1), order=1)
src.reshape(1,3,h,w) # resize the network's input image size
src.data[0] = octave_base+detail
for i in xrange(iter_n):
make_step(net, end=end, clip=clip, **step_params)
# visualization
vis = deprocess(net, src.data[0])
if not clip: # adjust image contrast if clipping is disabled
vis = vis*(255.0/np.percentile(vis, 99.98))
showarray(vis)
print octave, i, end, vis.shape
clear_output(wait=True)
# extract details produced on the current octave
detail = src.data[0]-octave_base
# returning the resulting image
return deprocess(net, src.data[0])
def verifyModel(net, model):
print "Verifying model: %s" % model
keys = net.blobs.keys()
if model in keys:
print "Model %s is valid." %model
return True
else:
print "Invalid model: %s. Valid models are:" % model
for k in keys:
print k
return False
if not verifyModel(net, model_name):
os._exit(1)
if not os.path.exists("/data/output"):
os.mkdir("/data/output")
if not os.path.exists("/data/output/tmp"):
os.mkdir("/data/output/tmp")
print "This might take a little while..."
print "Generating first sample..."
step_one = deepdream(net, img)
PIL.Image.fromarray(np.uint8(step_one)).save("/data/output/step_one.jpg")
print "Generating second sample..."
step_two = deepdream(net, img, end='inception_3b/5x5_reduce')
PIL.Image.fromarray(np.uint8(step_two)).save("/data/output/step_two.jpg")
frame = img
frame_i = 0
h, w = frame.shape[:2]
s = float(scale) # scale coefficient
print "Entering dream mode..."
print "Iterations = %s" % iterations
print "Scale = %s" % scale
print "Model = %s" % model_name
for i in xrange(int(iterations)):
print "Step %d of %d is starting..." % (i, int(iterations))
frame = deepdream(net, frame, end=model_name)
PIL.Image.fromarray(np.uint8(frame)).save("/data/output/%04d.jpg"%frame_i)
frame = nd.affine_transform(frame, [1-s,1-s,1], [h*s/2,w*s/2,0], order=1)
frame_i += 1
print "Step %d of %d is complete." % (i, int(iterations))
print "All done! Check the /output folder for results"