def __init__(self,**kargs):
self.patch_sz = kargs.get('patch_sz',4)
self.tstring = kargs.get('tstring', tstring()) # used for saving
self.model_name = kargs.get('model_name','BaseModel')
if self.patch_sz:
self.D = kargs.get('D',self.patch_sz*self.patch_sz)
else:
self.D = kargs.get('D',None)
def display(self,save_string=None,save=True,normalize_A=True,max_factors=256,zerophasewhiten=True):
output = {}
if hasattr(self.A,'get_value'):
A = self.A.get_value()
else:
A = self.A
if self.whiten:
A = np.dot(self.dewhitenmatrix,A)
if zerophasewhiten:
A = np.dot(self.zerophasewhitenmatrix,A)
if self.patch_sz:
psz = self.patch_sz
else:
psz = int(np.ceil(np.sqrt(A.shape[0])))
if not psz**2 == A.shape[0]:
A = np.vstack((A,np.zeros((psz**2 - A.shape[0],A.shape[1]))))
# plot the vectors in A
NN = min(self.NN,max_factors)
buf = 1
sz = int(np.sqrt(NN))
hval = np.max(np.abs(A))
array = -np.ones(((psz+buf)*sz+buf,(psz+buf)*sz+buf))
Aind = 0
for r in range(sz):
for c in range(sz):
if normalize_A:
hval = np.max(np.abs(A[:,Aind]))
Avalues = A[:,Aind].reshape(psz,psz)/hval
array[buf+(psz+buf)*c:buf+(psz+buf)*c+psz,buf+(psz+buf)*r:buf+(psz+buf)*r+psz] = Avalues
Aind += 1
hval = 1.
plt.figure(1)
plt.clf()
plt.imshow(array,vmin=-hval,vmax=hval,interpolation='nearest',cmap=plt.cm.gray)
plt.colorbar()
output['A'] = array
plt.draw()
if save:
from config import state_dir, tstring
savepath = os.path.join(state_dir,self.model_name + '_' + self.tstring)
if not os.path.isdir(savepath): os.makedirs(savepath)
if save_string is None:
save_string = tstring()
plt.figure(1)
fname = os.path.join(savepath, 'A_' + save_string + '.png')
plt.savefig(fname)
return output
def display(self,save_string=None,save=True,normalize_A=True,max_factors=64,zerophasewhiten=True):
if self.patch_sz is None:
raise NotImplemented
from display import display_color_patches
output = {}
if hasattr(self.A,'get_value'):
A = self.A.get_value()
else:
A = self.A
if self.whiten:
A = np.dot(self.dewhitenmatrix,A)
if zerophasewhiten:
A = np.dot(self.zerophasewhitenmatrix,A)
# plot the vectors in A
NN = min(self.NN,2*max_factors)
patches = np.zeros((self.patch_sz,self.patch_sz,3,2*NN))
count = 0
for nn in range(NN):
if normalize_A:
Ann = A[...,nn]
Ann /= np.abs(Ann).max() / 127.5
Ann += 127.5
else:
Ann = A[...,nn]
patches[...,count ] = Ann.reshape(6,self.patch_sz,self.patch_sz)[:3,...].T
patches[...,count+1] = Ann.reshape(6,self.patch_sz,self.patch_sz)[3:,...].T
count += 2
patches = patches.reshape((self.patch_sz*self.patch_sz*3,2*NN))
array = display_color_patches(patches,self.patch_sz,fig_num=1,normalize=False)
output['A'] = array
plt.draw()
if save:
from config import state_dir, tstring
savepath = os.path.join(state_dir,self.model_name + '_' + self.tstring)
if not os.path.isdir(savepath): os.makedirs(savepath)
if save_string is None:
save_string = tstring()
plt.figure(1)
fname = os.path.join(savepath, 'A_' + save_string + '.png')
plt.savefig(fname)
return output
def display_whitening(self,save_string=None,save=True,normalize_A=True,max_factors=64,zerophasewhiten=True):
if self.patch_sz is None:
raise NotImplemented
from display import display_color_patches
output = {}
if zerophasewhiten:
A = self.zerophasewhitenmatrix
else:
A = self.whitenmatrix
if hasattr(A,'get_value'):
A = A.get_value()
A = A.T # make the columns the filters
# plot the vectors in A
NN = min(A.shape[1],2*max_factors)
patches = np.zeros((self.patch_sz,self.patch_sz,3,2*NN))
count = 0
for nn in range(NN):
patches[...,count ] = A[...,nn].reshape(6,self.patch_sz,self.patch_sz)[:3,...].T
patches[...,count+1] = A[...,nn].reshape(6,self.patch_sz,self.patch_sz)[3:,...].T
count += 2
patches = patches.reshape((self.patch_sz*self.patch_sz*3,2*NN))
array = display_color_patches(patches,self.patch_sz,fig_num=1,normalize=normalize_A)
output['whitenmatrix'] = array
plt.draw()
if save:
from config import state_dir, tstring
savepath = os.path.join(state_dir,self.model_name + '_' + self.tstring)
if not os.path.isdir(savepath): os.makedirs(savepath)
if save_string is None:
save_string = tstring()
plt.figure(1)
fname = os.path.join(savepath, 'whitenmatrix_' + save_string + '.png')
plt.savefig(fname)
return output