def plot_relu_layer(input):
num = 1
for channel in range(input.shape[-1]):
result = input[:, :, channel]
result = result.reshape(input.shape[0], input.shape[1])
ax = subplot(5, 2, num, frame_on=False)
ax.xaxis.set_major_locator(NullLocator()) # remove ticks
ax.yaxis.set_major_locator(NullLocator())
num += 1
plt.imshow(result, origin='upper')
plt.title('channel #' + str(channel))
plt.tight_layout()
plt.savefig('relu_activations.png')
]),
color="green",
lw=1,
linestyle=lss[i],
label=None if i else "env. friendly individuals")
ax12.plot(t[3:],
100 * mean([
array(traj["SocialSystem.has_emissions_tax"][s][3:] * 1)
for s in social_systems
],
axis=0),
color="cyan",
lw=2,
label="regions w/ emissions tax")
ax12.set_ylim(-5, 105)
ax12.yaxis.set_major_locator(NullLocator())
ax12.legend(loc=7)
# metabolic
for i, s in enumerate(social_systems):
Es = array(traj["SocialSystem.secondary_energy_flow"][s][3:])
ax22.plot(t[3:],
100 * array(traj["SocialSystem.biomass_input_flow"][s][3:]) *
40e9 / Es,
color="green",
lw=lws,
alpha=al,
linestyle=lss[i],
label=None if i else "biomass")
ax22.plot(t[3:],
100 * array(traj["SocialSystem.fossil_fuel_input_flow"][s][3:]) *
color=(0, 0.25, 0))
else:
text(1.8 * n - 2,
0.14 * t_numpy - 6e-6,
label_numpy,
horizontalalignment='right',
size=11,
color=(0, 0.25, 0))
from matplotlib.text import FontProperties
for n, t_python, t_numpy, label_python, label_numpy in \
zip(sizes, time_python, time_numpy, labels_python, labels_numpy)[:-1]:
make_mark(n, t_python, t_numpy, label_python, label_numpy)
ax.xaxis.minor.locator = NullLocator()
yticks((1e-5, 1e-4, 1e-3, 1e-2, 1e-1, 1),
(u'10µs', u'100µs', u'1ms', u'10ms', u'100ms', u'1s'))
ax.yaxis.tick_right()
ax.yaxis.grid(alpha=0.4)
for t in ax.xaxis.majorTicks:
t.tick2line.set_visible(False)
legend(loc='upper left', numpoints=1, prop=FontProperties(size=12))
text(2, 1e-6, 'n')
ylabel(u"Temps d'éxécution")
savefig('timings.eps')
A = imread('../data/shakira.png') # load an image
A = mean(A, 2) # to get a 2-D array
full_pc = size(A, axis=1) # numbers of all the principal components
i = 1
dist = []
for numpc in range(0, full_pc + 10, 10): # 0 10 20 ... full_pc
coeff, score, latent = princomp(A, numpc)
Ar = dot(coeff, score).T + mean(A, axis=0) # image reconstruction
# difference in Frobenius norm
dist.append(linalg.norm(A - Ar, 'fro'))
# showing the pics reconstructed with less than 50 PCs
print "trying %s principal components with a distance of: %f" % (numpc,
dist[-1])
if numpc <= 50:
if numpc == 50:
imsave(fname='../data/' + str(numpc) + '.jpg', arr=Ar)
ax = subplot(2, 3, i, frame_on=False)
ax.xaxis.set_major_locator(NullLocator()) # remove ticks
ax.yaxis.set_major_locator(NullLocator())
i += 1
imshow(Ar)
title('PCs # ' + str(numpc))
gray()
figure()
imshow(A)
title('numpc FULL')
gray()
show()
10): # 0 10 20 ... full_pc + 10 sayısı içinde döngü
coeff, score, latent = princomp(A, numpc)
Ar = dot(coeff, score).T + mean(A, axis=0) # imajı tekrar yapılandırma
# Frobenius normdaki farklılık
dist.append(linalg.norm(A - Ar, 'fro'))
# temel bileşenlerle yeniden yapılandırılan imajları göster
# 50'den az olan temel bileşenler kullanılmıştır
if numpc <= 50:
ax = subplot(2, 3, i, frame_on=False)
ax.xaxis.set_major_locator(NullLocator())
ax.yaxis.set_major_locator(NullLocator())
i += 1
plt.imshow(Ar)
title('PCs # ' + str(numpc))
gray()
figure()
imshow(A)
title('numpc FULL')
gray()
show()
from pylab import plot, axis