-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathvisualizations.py
183 lines (150 loc) · 6.46 KB
/
visualizations.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
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
import numpy as np
import matplotlib as mpl
mpl.use('Agg')
import matplotlib.pyplot as plt
from matplotlib.collections import LineCollection
from matplotlib import colors
from matplotlib.ticker import MaxNLocator
from collections import namedtuple
from mpl_toolkits.axes_grid1 import ImageGrid
import pandas as pd
import cv2
import os
import glob
from datetime import datetime
def memory_usage_vis(csv_file, names=None):
if names is None:
names=['Dimensions', 'CPU-Mem', 'GPU-Mem']
dataset = pd.read_csv(csv_file, header=None, names=names, index_col=0)
plt.figure()
dataset.plot.bar(figsize=(11, 4), align='edge', width=0.5, rot=0)
plt.title('Memory Consumption')
plt.xlabel('Dimensions')
plt.ylabel('Memory (MB)')
plt.legend()
plt.savefig(csv_file[:-3]+'jpg')
def loss_vis(compliance_loss_array, title, save=True, path=None, **kwargs):
if path is None:
path = 'tmp/'
title_ = title
if os.path.isfile(path + title + '.png'):
title_ += str(int(datetime.timestamp(datetime.now())))
if save:
ylim = kwargs['ylim'] if 'ylim' in kwargs.keys() else 5000.0
plt.rcParams.update({'font.size': 18})
plt.figure(figsize=(14, 10))
plt.plot(np.arange(0, len(compliance_loss_array)), compliance_loss_array, label='compliance loss')
plt.title('Compliance')
plt.xlabel('Iteration')
plt.ylabel('Compliance Loss')
plt.ylim(0, ylim)
plt.suptitle(title, fontsize=18)
plt.savefig(path + title_ + '.png')
plt.close()
return title_
def density_vis(density, loss, gridDimensions, title, prediction=True, save=True, same_size=True, binary_loss=None, path=None):
plt.rcParams.update({'font.size': 18})
ratio = gridDimensions[0] / gridDimensions[1]
# for including titles properly (not related to solution)
if ratio == 1.0:
ratio = ratio * 3.5
if path is None:
path = 'tmp/'
if prediction:
if same_size:
plt.figure(figsize=(ratio * 5, ratio//ratio * 5 + 1))
else:
plt.figure(figsize=(gridDimensions[0]/10, gridDimensions[1]/10+1))
pred_density = -density.view(gridDimensions).detach().cpu().numpy()[:, :].T
plt.imshow(pred_density, cmap='gray')
if binary_loss is not None:
plt.title('Prediction (loss={:.3f}, b-loss={:.3f}, vol={:.3f})'.format(loss, binary_loss, -pred_density.mean()))
else:
plt.title('Prediction (loss={:.3f}, vol={:.3f})'.format(loss, -pred_density.mean()))
plt.suptitle('Prediction ('+title+')', fontsize=13)
if save:
plt.suptitle(title, fontsize=18)
title_ = title
if os.path.isfile(path + title + '.png'):
title_ += str(int(datetime.timestamp(datetime.now())))
plt.savefig(path + title_ + '.png')
else:
if same_size:
plt.figure(figsize=(ratio * 5, ratio//ratio * 5 + 1))
else:
plt.figure(figsize=(gridDimensions[0]/10, gridDimensions[1]/10+1))
plt.imshow(-density.reshape(gridDimensions[0], gridDimensions[1]).T, cmap='gray')
if save:
if binary_loss is not None:
plt.title('{} (loss={:.3f}, b-loss={:.3f})'.format(title, loss, binary_loss))
else:
plt.title('{} (loss={:.3f})'.format(title, loss))
title_ = title
if os.path.isfile(path + title + '_gt.png'):
title_ += str(int(datetime.timestamp(datetime.now())))
plt.savefig(path + title_ + '_gt.png')
plt.close()
def pred_gt_density_vis(pred, gt, gridDimensions, pred_loss, gt_loss, title, save=True, path=None):
density = pred
gt_densities = gt
fig, axes = plt.subplots(1, 2, figsize=(14, 5))
pred_density = -density.view(gridDimensions).detach().cpu().numpy()[:, :].T
axes[0].imshow(pred_density, cmap='gray')
axes[0].set_title('Prediction (loss={:.3f}, vol={:.3f})'.format(pred_loss, -pred_density.mean()))
axes[1].imshow(-gt_densities.reshape(gridDimensions[0], gridDimensions[1]).T, cmap='gray')
axes[1].set_title('Ground Truth (loss={:.3f})'.format(gt_loss))
fig.suptitle('VoxelFEM vs CNN VoxelFEM ('+title+')', fontsize=13)
if path is None:
path = 'logs/images/cnn/vanilla'
if save:
fig.suptitle(title, fontsize=18)
if os.path.isfile(path + title + '.png'):
title += str(int(datetime.timestamp(datetime.now())))
plt.savefig(path + title + '.png')
def n_column_image_grid(title, image_list=None, path=None, patterns=None):
if (image_list is None) and (path is None):
raise Exception('"image_list" and "path" both cannot be "None"')
if patterns is None:
raise Exception('Provide patterns for file names (relative path)')
if image_list is None:
path = ''
def key_for_sort(string: str):
idx = string.find('_s') + 2 # scale value
scale_str = string[idx: idx + 4]
try:
scale = float(scale_str) # from 10.0 to 99.9
except ValueError:
scale = float(scale_str[:-1]) # from 0.1 to 9.9
return scale
images = []
for p in patterns:
images_path = glob.glob(p)
images_path.sort(key=key_for_sort)
images_col = []
for j in images_path:
images_col.append(cv2.imread(j, 0))
images.append(images_col)
cols = len(images[0])
rows = len(patterns)
factor = 1
rows = rows * factor
cols = cols // factor
fig = plt.figure(figsize=(rows*5, cols*5))
grid = ImageGrid(fig, 111, # similar to subplot(111)
nrows_ncols=(cols, rows),
axes_pad=0.1,
)
images = np.array(images)
images = list(images.transpose(1, 0, 2, 3).reshape(-1, images.shape[-2], images.shape[-1]))
for ax, im in zip(grid, images):
ax.tick_params(left=False, labelleft=False) #remove ticks
ax.tick_params(bottom=False, labelbottom=False) #remove ticks
ax.set_frame_on(True)
ax.imshow(im[5:-240, 120:-120], cmap='gray')
plt.savefig(path+title+'_.png')
plt.close()
def plot_n_large_scale_fourfeat_outputs(title, patterns=None):
patterns = [r'tmp/fourfeat_vscale/*150x50*MbbBeamSeed88*.png',
r'tmp/fourfeat_vscale/*300x100*MbbBeamSeed88*.png']
n_column_image_grid(title='fourfeat_150x50_VS_300x100_S[0.5-30]_MbbBeamSeed88',
image_list=None, path='', patterns=patterns)