Here are the examples of the python api numpy.add taken from open source projects. By voting up you can indicate which examples are most useful and appropriate.
83 Examples
5
Example 1
Project: tvb-library Source File: common_test.py
@unittest.skipIf(not hasattr(numpy.add, 'at'),
'Cannot test fallback numpy.add.at implementation without '
'a version of NumPy which provides this ufunc method (>=1.8).')
def test_add_at(self):
ri = numpy.random.randint
for nd in range(1, 5):
m, n, rest = ri(3, 50), ri(51, 100), tuple(ri(3, 10, nd - 1))
source = ri(-100, 100, (n,) + rest)
map = ri(0, m, n)
expected, actual = numpy.zeros((2, m) + rest)
numpy.add.at(expected, map, source)
common._add_at(actual, map, source)
self.assertTrue(numpy.allclose(expected, actual))
3
Example 2
def testshared():
for schedule in ['static', 'dynamic', 'guided']:
with Parallel(
Shared(a=[0, 0]),
Reduction(numpy.add, b=[0, 0])
) as p:
for i in p.forloop(range(100), schedule=schedule):
with p.critical:
p.var.a += numpy.array([i, i * 10])
p.var.b += numpy.array([i, i * 10])
assert (p.var.a == p.var.b).all()
3
Example 3
Project: NeuroM Source File: _neuritefunc.py
def segment_midpoints(neurites, neurite_type=NeuriteType.all):
'''Return a list of segment mid-points in a collection of neurites'''
def _seg_midpoint(sec):
'''Return the mid-points of segments in a section'''
pts = sec.points
return np.divide(np.add(pts[:-1], pts[1:])[:, :3], 2.0)
neurite_filter = is_type(neurite_type)
return [s for ss in iter_sections(neurites, neurite_filter=neurite_filter)
for s in _seg_midpoint(ss)]
3
Example 4
Project: tfdeploy Source File: tfdeploy.py
@Operation.factory(types=("SegmentSum", "UnsortedSegmentSum"))
def SegmentSum(a, ids, *args):
"""
Segmented sum op.
"""
func = lambda idxs: reduce(np.add, a[idxs])
return seg_map(func, a, ids),
3
Example 5
Project: biggus Source File: test__ufunc_wrapper.py
def test_nin2_docstring(self):
wrapped_fn = _ufunc_wrapper(np.add)
doc = inspect.getdoc(wrapped_fn)
expected = ('Return the elementwise evaluation of '
'np.add(a, b) as another Array.')
self.assertEqual(doc, expected)
3
Example 6
Project: iris Source File: test_Cube__operators.py
def test_lazy_biggus_add_scalar(self):
c1 = self.build_lazy_cube([1, 2])
cube = c1 + 5
self.assertEqual(c1 + 5, 5 + c1)
result = cube.lazy_data()
self.assertTrue(cube.has_lazy_data())
self.assert_elementwise(c1, None, result, np.add)
3
Example 7
Project: pyDive Source File: test_algorithm.py
def test_reduce(init_pyDive):
input_array = pyDive.h5.open(input_file, "fields").load()
energy_array = pyDive.empty(input_array.shape, dtype=input_array.dtype["fieldE"]["x"])
def energy(out, fields):
out[:] = fields["fieldE/x"]**2 + fields["fieldE/y"]**2 + fields["fieldB/z"]**2
pyDive.map(energy, energy_array, input_array)
test_total = pyDive.reduce(energy_array, np.add)
ref_total = np.add.reduce(energy_array.gather(), axis=None)
diff = abs(ref_total - test_total)
assert diff / ref_total < 1.0e-5
3
Example 8
@Operation.factory(types=("Add", "BiasAdd"))
def Add(a, b):
"""
Addition op.
"""
return np.add(a, b),
3
Example 9
Project: sharedmem Source File: parallel.py
def testraiseordered():
for schedule in ['static', 'dynamic', 'guided']:
try:
with Parallel(
Reduction(numpy.add, a=[0, 0])
) as p:
for i, ordered in p.forloop(range(20),
ordered=True, schedule=schedule):
with ordered:
p.var.a += numpy.array([i, i * 10])
if i == 19:
raise ValueError('raised at i == 19')
assert False
except ValueError as e:
pass
3
Example 10
Project: AWS-Lambda-ML-Microservice-Skeleton Source File: test_operators.py
def test_combine_generic(self):
df1 = self.frame
df2 = self.frame.ix[:-5, ['A', 'B', 'C']]
combined = df1.combine(df2, np.add)
combined2 = df2.combine(df1, np.add)
self.assertTrue(combined['D'].isnull().all())
self.assertTrue(combined2['D'].isnull().all())
chunk = combined.ix[:-5, ['A', 'B', 'C']]
chunk2 = combined2.ix[:-5, ['A', 'B', 'C']]
exp = self.frame.ix[:-5, ['A', 'B', 'C']].reindex_like(chunk) * 2
assert_frame_equal(chunk, exp)
assert_frame_equal(chunk2, exp)
3
Example 11
def K(self,X ,X2=None):
K = self.parts[0].K(X, X2) # compute 'base' kern everywhere
slices = [index_to_slices(X[:,i]) for i in self.extra_dims]
if X2 is None:
[[[np.add(K[s,s], k.K(X[s], None), K[s, s]) for s in slices_i] for slices_i in slices_k] for k, slices_k in zip(self.parts[1:], slices)]
else:
slices2 = [index_to_slices(X2[:,i]) for i in self.extra_dims]
[[[np.add(K[s,ss], k.K(X[s], X2[ss]), K[s, ss]) for s,ss in zip(slices_i, slices_j)] for slices_i, slices_j in zip(slices_k1, slices_k2)] for k, slices_k1, slices_k2 in zip(self.parts[1:], slices, slices2)]
return K
3
Example 12
def _compute_moments(self, *u_parents):
"""
Compute the moments of the sum
"""
u0 = functools.reduce(np.add,
(u_parent[0] for u_parent in u_parents))
u1 = functools.reduce(np.add,
(u_parent[1] for u_parent in u_parents))
for i in range(self.N):
for j in range(i+1, self.N):
xi_xj = linalg.outer(u_parents[i][0], u_parents[j][0], ndim=self.ndim)
xj_xi = linalg.transpose(xi_xj, ndim=self.ndim)
u1 = u1 + xi_xj + xj_xi
return [u0, u1]
3
Example 13
Project: NeuroM Source File: _neuritefunc.py
def segment_radial_distances(neurites, neurite_type=NeuriteType.all, origin=None):
'''Lengths of the segments in a collection of neurites'''
def _seg_rd(sec, pos):
'''list of radial distances of all segments of a section'''
mid_pts = np.divide(np.add(sec.points[:-1], sec.points[1:])[:, :3], 2.0)
return np.sqrt([mm.point_dist2(p, pos) for p in mid_pts])
dist = []
for n in iter_neurites(neurites, filt=is_type(neurite_type)):
pos = n.root_node.points[0] if origin is None else origin
dist.extend([s for ss in n.iter_sections() for s in _seg_rd(ss, pos)])
return dist
3
Example 14
Project: iris Source File: test_Cube__operators.py
def test_lazy_biggus_add_cubes(self):
c1 = self.build_lazy_cube([1, 2])
cube = c1 + c1
result = cube.lazy_data()
self.assertTrue(cube.has_lazy_data())
self.assert_elementwise(c1, c1.lazy_data(), result, np.add)
3
Example 15
def testreduction():
for schedule in ['static', 'dynamic', 'guided']:
with Parallel(
Reduction(numpy.add, a=[0, 0])
) as p:
for i in p.forloop(range(20), schedule=schedule) :
p.var.a += numpy.array([i, i * 10])
assert (p.var.a == [190, 1900]).all()
3
Example 16
Project: NeuroM Source File: morphmath.py
def segment_radial_dist(seg, pos):
'''Return the radial distance of a tree segment to a given point
The radial distance is the euclidian distance between the mid-point of
the segment and the point in question.
Parameters:
seg: tree segment
pos: origin to which disrances are measured. It must have at lease 3
components. The first 3 components are (x, y, z).
'''
return point_dist(pos, np.divide(np.add(seg[0], seg[1]), 2.0))
3
Example 17
@Cache_this(limit=3, force_kwargs=['which_parts'])
def K(self, X, X2=None, which_parts=None):
"""
Add all kernels together.
If a list of parts (of this kernel!) `which_parts` is given, only
the parts of the list are taken to compute the covariance.
"""
if which_parts is None:
which_parts = self.parts
elif not isinstance(which_parts, (list, tuple)):
# if only one part is given
which_parts = [which_parts]
return reduce(np.add, (p.K(X, X2) for p in which_parts))
3
Example 18
Project: landsat-util Source File: image.py
def _pansize(self, bands):
self.output('Calculating Pan Ratio', normal=True, arrow=True)
m = numpy.add(bands[0], bands[1])
m = numpy.add(m, bands[2])
pan = numpy.multiply(numpy.nan_to_num(numpy.true_divide(1, m)), bands[self.band8])
return pan
3
Example 19
Project: sharedmem Source File: parallel.py
def testraisecritical():
for schedule in ['static', 'dynamic', 'guided']:
try:
with Parallel(
Reduction(numpy.add, a=[0, 0])
) as p:
for i in p.forloop(range(20), schedule=schedule):
with p.critical:
p.var.a += numpy.array([i, i * 10])
if i == 19:
raise ValueError('raised at i == 19')
assert False
except ValueError as e:
pass
3
Example 20
def updateRasterInfo(self, **kwargs):
m = kwargs.get('op', 'Add').lower()
if m == 'add': self.op = np.add
elif m == 'subtract': self.op = np.subtract
elif m == 'multiply': self.op = np.multiply
elif m == 'divide': self.op = np.divide
kwargs['output_info']['statistics'] = ()
kwargs['output_info']['histogram'] = ()
return kwargs
3
Example 21
@Cache_this(limit=3, force_kwargs=['which_parts'])
def Kdiag(self, X, which_parts=None):
if which_parts is None:
which_parts = self.parts
elif not isinstance(which_parts, (list, tuple)):
# if only one part is given
which_parts = [which_parts]
return reduce(np.add, (p.Kdiag(X) for p in which_parts))
3
Example 22
def testbarrier():
now = time.time()
with Parallel(
Shared(a=[0, 0]),
Reduction(numpy.add, b=[0, 0])
) as p:
#time.sleep(p.rank * 0.01)
p.barrier()
p.barrier()
3
Example 23
def neighbours(self, idx):
"""
Return a list of indices to the neighbours of a given pixel.
This method can be overridden to handle custom layouts
(e.g., healpix maps, periodic boundaries, etc.)
Parameters
----------
idx : tuple
The N-dimensional location of a pixel in the data
Returns
-------
List of N-dimensional locations of each neighbour
"""
return [tuple(c) for c in np.add(_offsets[self.n_dim], idx)]
3
Example 24
Project: tfdeploy Source File: tfdeploy.py
@Operation.factory
def SparseSegmentSqrtN(a, idxs, ids):
"""
Sparse segmented sum / sqrt(n=len(idxs)) op.
"""
func = lambda _idxs: np.divide(reduce(np.add, a[idxs][_idxs]), np.math.sqrt(len(_idxs)))
return seg_map(func, a, ids),
3
Example 25
Project: AWS-Lambda-ML-Microservice-Skeleton Source File: test_ufunc.py
def test_kwarg_exact(self):
assert_raises(TypeError, np.add, 1, 2, castingx='safe')
assert_raises(TypeError, np.add, 1, 2, dtypex=np.int)
assert_raises(TypeError, np.add, 1, 2, extobjx=[4096])
assert_raises(TypeError, np.add, 1, 2, outx=None)
assert_raises(TypeError, np.add, 1, 2, sigx='ii->i')
assert_raises(TypeError, np.add, 1, 2, signaturex='ii->i')
assert_raises(TypeError, np.add, 1, 2, subokx=False)
assert_raises(TypeError, np.add, 1, 2, wherex=[True])
3
Example 26
def add_deviations(self, specresp):
# copy the old ARF (use new memory for deviations)
new_arf = np.add(specresp, self.bias)
# Include the perturbed effective area in each iteration.
rr = np.random.randint(self.ncomp)
return np.add(new_arf, self.simcomp[rr], new_arf)
3
Example 27
def add(A, b, offset=0):
"""
Add b to the view of A in place (!).
Returns modified A.
Broadcasting is allowed, thus b can be scalar.
if offset is not zero, make sure b is of right shape!
:param ndarray A: 2 dimensional array
:param ndarray-like b: either one dimensional or scalar
:param int offset: same as in view.
:rtype: view of A, which is adjusted inplace
"""
return _diag_ufunc(A, b, offset, np.add)
3
Example 28
@pytest.mark.skipif(platform.python_implementation() == 'PyPy',
reason="Skip numpy and scipy tests on PyPy")
def test_ufunc(self):
# test a numpy ufunc (universal function), which is a C-based function
# that is applied on a numpy array
if np:
# simple ufunc: np.add
self.assertEqual(pickle_depickle(np.add), np.add)
else: # skip if numpy is not available
pass
if spp:
# custom ufunc: scipy.special.iv
self.assertEqual(pickle_depickle(spp.iv), spp.iv)
else: # skip if scipy is not available
pass
3
Example 29
def execute(self, image):
self._images.append(image)
while len(self._images) >= self.nb_images.get():
del self._images[0]
try:
for img in self._images:
image = np.add(image, img)
except:
pass
return image
3
Example 30
@Operation.factory(unpack=False)
def AddN(inputs):
"""
Multi add op.
"""
return reduce(np.add, inputs),
3
Example 31
Project: hyperspy Source File: test_tools.py
def test_numpy_unfunc_two_arg_titled(self):
s1, s2 = self.signal.deepcopy(), self.signal.deepcopy()
s1.metadata.General.title = "A"
s2.metadata.General.title = "B"
result = np.add(s1, s2)
nt.assert_true(isinstance(result, signals.Signal1D))
np.testing.assert_array_equal(result.data, np.add(s1.data, s2.data))
nt.assert_equal(result.metadata.General.title, "add(A, B)")
3
Example 32
def testkill():
try:
with Parallel(
Shared(a=[0, 0]),
Reduction(numpy.add, b=[0, 0])
) as p:
# time.sleep(p.rank * 0.01)
p.barrier()
if p.rank == p.num_threads - 1:
os.kill(os.getpid(), signal.SIGKILL)
p.barrier()
assert False
except ParallelException as e:
return
assert False
3
Example 33
def add_deviations(self, specresp, rrin=None, rrsig=None):
# copy the old ARF (use new memory for deviations)
new_arf = np.add(specresp, self.bias)
rrout = np.random.standard_normal(self.ncomp)
if rrin is not None and rrsig is not None:
rrout = rrin + rrsig * rrout
self.rrout = rrout
tmp = self.eigenvec * self.eigenval[:,np.newaxis] * rrout[:,np.newaxis]
return np.add(new_arf, tmp.sum(axis=0), new_arf)
2
Example 34
Project: Reactor-3 Source File: maps.py
def render_lights(size=MAP_WINDOW_SIZE, show_weather=True):
if not SETTINGS['draw lights']:
return False
reset_lights(size=size)
_weather_light = weather.get_lighting()
#Not entirely my code. Made some changes to someone's code from libtcod's Python forum.
RGB_LIGHT_BUFFER[0] = numpy.add(RGB_LIGHT_BUFFER[0], _weather_light[0])
RGB_LIGHT_BUFFER[1] = numpy.add(RGB_LIGHT_BUFFER[1], _weather_light[1])
RGB_LIGHT_BUFFER[2] = numpy.add(RGB_LIGHT_BUFFER[2], _weather_light[2])
(x, y) = SETTINGS['light mesh grid']
if show_weather:
weather.generate_effects(size)
_remove_lights = []
for light in WORLD_INFO['lights']:
_x_range = light['pos'][0]-CAMERA_POS[0]
_y_range = light['pos'][1]-CAMERA_POS[1]
if _x_range <= -20 or _x_range>=size[0]+20:
continue
if _y_range <= -20 or _y_range>=size[1]+20:
continue
if not 'old_pos' in light:
light['old_pos'] = (0, 0, -2)
else:
light['old_pos'] = light['pos'][:]
if 'follow_item' in light:
if not light['follow_item'] in ITEMS:
_remove_lights.append(light)
continue
light['pos'] = items.get_pos(light['follow_item'])[:]
_render_x = light['pos'][0]-CAMERA_POS[0]
_render_y = light['pos'][1]-CAMERA_POS[1]
_x = numbers.clip(light['pos'][0]-(size[0]/2),0,MAP_SIZE[0])
_y = numbers.clip(light['pos'][1]-(size[1]/2),0,MAP_SIZE[1])
_top_left = (_x,_y,light['pos'][2])
#TODO: Render only on move
if not tuple(light['pos']) == tuple(light['old_pos']):
light['los'] = cython_render_los.render_los((light['pos'][0],light['pos'][1]), light['brightness']*2, view_size=size, top_left=_top_left)
los = light['los'].copy()
_x_scroll = _x-CAMERA_POS[0]
_x_scroll_over = 0
_y_scroll = _y-CAMERA_POS[1]
_y_scroll_over = 0
if _x_scroll<0:
_x_scroll_over = _x_scroll
_x_scroll = los.shape[1]+_x_scroll
if _y_scroll<0:
_y_scroll_over = _y_scroll
_y_scroll = los.shape[0]+_y_scroll
los = numpy.roll(los, _y_scroll, axis=0)
los = numpy.roll(los, _x_scroll, axis=1)
los[_y_scroll_over:_y_scroll,] = 1
los[:,_x_scroll_over:_x_scroll] = 1
if SETTINGS['diffuse light']:
_y, _x = diffuse_light((y, x))
(x, y) = numpy.meshgrid(_x, _y)
sqr_distance = (x - (_render_x))**2.0 + (y - (_render_y))**2.0
brightness = numbers.clip(random.uniform(light['brightness']*light['shake'], light['brightness']), 0.01, 50) / sqr_distance
brightness *= los
#brightness *= LOS_BUFFER[0]
#_mod = (abs((WORLD_INFO['length_of_day']/2)-WORLD_INFO['real_time_of_day'])/float(WORLD_INFO['length_of_day']))*5.0
#_mod = numbers.clip(_mod-1, 0, 1)
#(255*_mod, 165*_mod, 0*_mod)
#print brightness
#light['brightness'] = 25
#light['color'][0] = 255*(light['brightness']/255.0)
#light['color'][1] = (light['brightness']/255.0)
#light['color'][2] = 255*(light['brightness']/255.0)
RGB_LIGHT_BUFFER[0] -= (brightness.clip(0, 2)*(light['color'][0]))#numpy.subtract(RGB_LIGHT_BUFFER[0], light['color'][0]).clip(0, 255)
RGB_LIGHT_BUFFER[1] -= (brightness.clip(0, 2)*(light['color'][1]))#numpy.subtract(RGB_LIGHT_BUFFER[1], light['color'][1]).clip(0, 255)
RGB_LIGHT_BUFFER[2] -= (brightness.clip(0, 2)*(light['color'][2]))#numpy.subtract(RGB_LIGHT_BUFFER[2], light['color'][2]).clip(0, 255)
0
Example 35
def test_add(self):
self._test(np.add, ma.add)
0
Example 36
Project: ray Source File: core.py
@ray.remote
def add(x1, x2):
return np.add(x1, x2)
0
Example 37
def _compute_message_to_parent(self, index, m, *u_parents):
"""
Compute the message to a parent node.
.. math::
(\sum_i \mathbf{x}_i)^T \mathbf{M}_2 (\sum_j \mathbf{x}_j)
+ (\sum_i \mathbf{x}_i)^T \mathbf{m}_1
Moments of the parents are
.. math::
u_1^{(i)} = \langle \mathbf{x}_i \rangle
\\
u_2^{(i)} = \langle \mathbf{x}_i \mathbf{x}_i^T \rangle
Thus, the message for :math:`i`-th parent is
.. math::
\phi_{x_i}^{(1)} = \mathbf{m}_1 + 2 \mathbf{M}_2 \sum_{j\neq i} \mathbf{x}_j
\\
\phi_{x_i}^{(2)} = \mathbf{M}_2
"""
# Remove the moments of the parent that receives the message
u_parents = u_parents[:index] + u_parents[(index+1):]
m0 = (m[0] +
linalg.mvdot(
2*m[1],
functools.reduce(np.add,
(u_parent[0] for u_parent in u_parents)),
ndim=self.ndim))
m1 = m[1]
return [m0, m1]
0
Example 38
Project: peregrine Source File: satellite_glo.py
def getBatchSignals(self,
userTimeAll_s,
samples,
outputConfig,
noiseParams,
band,
debug):
'''
Generates signal samples.
Parameters
----------
userTimeAll_s : numpy.ndarray(n_samples, dtype=numpy.float64)
Vector of observer's timestamps in seconds for the interval start.
samples : numpy.ndarray((4, n_samples))
Array to which samples are added.
outputConfig : object
Output configuration object.
noiseParams : NoiseParameters
Noise parameters object
band : Band
Band description object.
debug : bool
Debug flag
Returns
-------
list
Debug information
'''
result = []
if (self.l1Enabled and band == outputConfig.GLONASS.L1):
intermediateFrequency_hz = band.INTERMEDIATE_FREQUENCIES_HZ[self.prn]
values = self.doppler.computeBatch(userTimeAll_s,
self.amplitude,
noiseParams,
signals.GLONASS.L1S[self.prn],
intermediateFrequency_hz,
self.l1Message,
self.caCode,
outputConfig,
debug)
numpy.add(samples[band.INDEX],
values[0],
out=samples[band.INDEX])
debugData = {'type': "GLOL1", 'doppler': values[1]}
result.append(debugData)
if (self.l2Enabled and band == outputConfig.GLONASS.L2):
intermediateFrequency_hz = band.INTERMEDIATE_FREQUENCIES_HZ[self.prn]
values = self.doppler.computeBatch(userTimeAll_s,
self.amplitude,
noiseParams,
signals.GLONASS.L2S[self.prn],
intermediateFrequency_hz,
self.l2Message,
self.caCode,
outputConfig,
debug)
numpy.add(samples[band.INDEX],
values[0],
out=samples[band.INDEX])
debugData = {'type': "GLOL2", 'doppler': values[1]}
result.append(debugData)
return result
0
Example 39
Project: peregrine Source File: satellite_gps.py
def getBatchSignals(self,
userTimeAll_s,
samples,
outputConfig,
noiseParams,
band,
debug):
'''
Generates signal samples.
Parameters
----------
userTimeAll_s : numpy.ndarray(n_samples, dtype=numpy.float64)
Vector of observer's timestamps in seconds for the interval start.
samples : numpy.ndarray((4, n_samples))
Array to which samples are added.
outputConfig : object
Output configuration object.
noiseParams : NoiseParameters
Noise parameters object
band : Band
Band description object.
debug : bool
Debug flag
Returns
-------
list
Debug information
'''
result = []
if (self.l1caEnabled and band == outputConfig.GPS.L1):
intermediateFrequency_hz = band.INTERMEDIATE_FREQUENCY_HZ
values = self.doppler.computeBatch(userTimeAll_s,
self.amplitude,
noiseParams,
signals.GPS.L1CA,
intermediateFrequency_hz,
self.l1caMessage,
self.l1caCode,
outputConfig,
debug)
numpy.add(samples[band.INDEX],
values[0],
out=samples[band.INDEX])
debugData = {'type': "GPSL1", 'doppler': values[1]}
result.append(debugData)
if (self.l2cEnabled and band == outputConfig.GPS.L2):
intermediateFrequency_hz = band.INTERMEDIATE_FREQUENCY_HZ
values = self.doppler.computeBatch(userTimeAll_s,
self.amplitude,
noiseParams,
signals.GPS.L2C,
intermediateFrequency_hz,
self.l2cMessage,
self.l2cCode,
outputConfig,
debug)
numpy.add(samples[band.INDEX],
values[0],
out=samples[band.INDEX])
debugData = {'type': "GPSL2", 'doppler': values[1]}
result.append(debugData)
return result
0
Example 40
def do_eval(hypes, eval_list, phase, sess):
"""
Run one evaluation against the full epoch of data.
Parameters
----------
hypes : dict
Hyperparameters
eval_list : list of tuples
Each tuple should contain a string (name if the metric) and a
tensor (storing the result of the metric).
phase : str
Describes the data the evaluation is run on.
sess : tf.Session
The session in which the model has been trained.
Returns
-------
tuple of lists
List of names and evaluation results
"""
# And run one epoch of eval.
# Checking for List for compability
if eval_list[phase] is None:
return [''], [0.0]
if type(eval_list[phase]) is list:
eval_names, eval_op = zip(*eval_list[phase])
else:
logging.warning("Passing eval_op directly is deprecated. "
"Pass a list of tuples instead.")
eval_names = ['Accuracy']
eval_op = [eval_list[phase]]
assert(len(eval_names) == len(eval_op))
if phase == 'train':
num_examples = hypes['data']['num_examples_per_epoch_for_train']
if phase == 'val':
num_examples = hypes['data']['num_examples_per_epoch_for_eval']
steps_per_epoch = num_examples // hypes['solver']['batch_size']
num_examples = steps_per_epoch * hypes['solver']['batch_size']
logging.info('Data: % s Num examples: % d ' % (phase, num_examples))
# run evaluation on num_examples many images
results = sess.run(eval_op)
logging.debug('Output of eval: %s', results)
for step in xrange(1, steps_per_epoch):
results = map(np.add, results, sess.run(eval_op))
avg_results = [result / steps_per_epoch for result in results]
for name, value in zip(eval_names, avg_results):
logging.info('%s : % 0.04f ' % (name, value))
return eval_names, avg_results
0
Example 41
Project: GPy Source File: add.py
@Cache_this(limit=3, force_kwargs=['which_parts'])
def psi0(self, Z, variational_posterior):
if not self._exact_psicomp: return Kern.psi0(self,Z,variational_posterior)
return reduce(np.add, (p.psi0(Z, variational_posterior) for p in self.parts))
0
Example 42
Project: GPy Source File: add.py
@Cache_this(limit=3, force_kwargs=['which_parts'])
def psi1(self, Z, variational_posterior):
if not self._exact_psicomp: return Kern.psi1(self,Z,variational_posterior)
return reduce(np.add, (p.psi1(Z, variational_posterior) for p in self.parts))
0
Example 43
@Cache_this(limit=3, force_kwargs=['which_parts'])
def psi2(self, Z, variational_posterior):
if not self._exact_psicomp: return Kern.psi2(self,Z,variational_posterior)
psi2 = reduce(np.add, (p.psi2(Z, variational_posterior) for p in self.parts))
#return psi2
# compute the "cross" terms
from .static import White, Bias
from .rbf import RBF
#from rbf_inv import RBFInv
from .linear import Linear
#ffrom fixed import Fixed
for p1, p2 in itertools.combinations(self.parts, 2):
# i1, i2 = p1._all_dims_active, p2._all_dims_active
# white doesn;t combine with anything
if isinstance(p1, White) or isinstance(p2, White):
pass
# rbf X bias
#elif isinstance(p1, (Bias, Fixed)) and isinstance(p2, (RBF, RBFInv)):
elif isinstance(p1, Bias) and isinstance(p2, (RBF, Linear)):
tmp = p2.psi1(Z, variational_posterior).sum(axis=0)
psi2 += p1.variance * (tmp[:,None]+tmp[None,:]) #(tmp[:, :, None] + tmp[:, None, :])
#elif isinstance(p2, (Bias, Fixed)) and isinstance(p1, (RBF, RBFInv)):
elif isinstance(p2, Bias) and isinstance(p1, (RBF, Linear)):
tmp = p1.psi1(Z, variational_posterior).sum(axis=0)
psi2 += p2.variance * (tmp[:,None]+tmp[None,:]) #(tmp[:, :, None] + tmp[:, None, :])
elif isinstance(p2, (RBF, Linear)) and isinstance(p1, (RBF, Linear)):
assert np.intersect1d(p1._all_dims_active, p2._all_dims_active).size == 0, "only non overlapping kernel dimensions allowed so far"
tmp1 = p1.psi1(Z, variational_posterior)
tmp2 = p2.psi1(Z, variational_posterior)
psi2 += np.einsum('nm,no->mo',tmp1,tmp2)+np.einsum('nm,no->mo',tmp2,tmp1)
#(tmp1[:, :, None] * tmp2[:, None, :]) + (tmp2[:, :, None] * tmp1[:, None, :])
else:
raise NotImplementedError("psi2 cannot be computed for this kernel")
return psi2
0
Example 44
Project: attention-lvcsr Source File: extending_theano_solution_1.py
@as_op(itypes=[theano.tensor.fmatrix, theano.tensor.fmatrix],
otypes=[theano.tensor.fmatrix], infer_shape=infer_shape_numpy_dot)
def numpy_add(a, b):
return numpy.add(a, b)
0
Example 45
Project: GPy Source File: add.py
@Cache_this(limit=3, force_kwargs=['which_parts'])
def psi2n(self, Z, variational_posterior):
if not self._exact_psicomp: return Kern.psi2n(self, Z, variational_posterior)
psi2 = reduce(np.add, (p.psi2n(Z, variational_posterior) for p in self.parts))
#return psi2
# compute the "cross" terms
from .static import White, Bias
from .rbf import RBF
#from rbf_inv import RBFInv
from .linear import Linear
#ffrom fixed import Fixed
for p1, p2 in itertools.combinations(self.parts, 2):
# i1, i2 = p1._all_dims_active, p2._all_dims_active
# white doesn;t combine with anything
if isinstance(p1, White) or isinstance(p2, White):
pass
# rbf X bias
#elif isinstance(p1, (Bias, Fixed)) and isinstance(p2, (RBF, RBFInv)):
elif isinstance(p1, Bias) and isinstance(p2, (RBF, Linear)):
tmp = p2.psi1(Z, variational_posterior)
psi2 += p1.variance * (tmp[:, :, None] + tmp[:, None, :])
#elif isinstance(p2, (Bias, Fixed)) and isinstance(p1, (RBF, RBFInv)):
elif isinstance(p2, Bias) and isinstance(p1, (RBF, Linear)):
tmp = p1.psi1(Z, variational_posterior)
psi2 += p2.variance * (tmp[:, :, None] + tmp[:, None, :])
elif isinstance(p2, (RBF, Linear)) and isinstance(p1, (RBF, Linear)):
assert np.intersect1d(p1._all_dims_active, p2._all_dims_active).size == 0, "only non overlapping kernel dimensions allowed so far"
tmp1 = p1.psi1(Z, variational_posterior)
tmp2 = p2.psi1(Z, variational_posterior)
psi2 += np.einsum('nm,no->nmo',tmp1,tmp2)+np.einsum('nm,no->nmo',tmp2,tmp1)
#(tmp1[:, :, None] * tmp2[:, None, :]) + (tmp2[:, :, None] * tmp1[:, None, :])
else:
raise NotImplementedError("psi2 cannot be computed for this kernel")
return psi2
0
Example 46
def update_weights(x, xbar, rho, w):
return numpy.add(w,rho*(x-xbar))
0
Example 47
Project: GPy Source File: finite_dimensional.py
def dKdiag_dtheta(self,X,target):
np.add(target[:,0],1.,target[:,0])
0
Example 48
Project: attention-lvcsr Source File: extending_theano_solution_1.py
@as_op(itypes=[theano.tensor.fmatrix, theano.tensor.fmatrix],
otypes=[theano.tensor.fmatrix], infer_shape=infer_shape_numpy_add_sub)
def numpy_add(a, b):
return numpy.add(a, b)
0
Example 49
def Kdiag(self, X, target):
"""Compute the diagonal of the covariance matrix for X."""
np.add(target, self.variance, target)
0
Example 50
def test_add(self):
self._test_elementwise(biggus.add, np.add)