def copyto(dst, src):
"""Copies the elements of an ndarray to those of another one.
This function can copy the CPU/GPU arrays to the destination arrays on
another device.
Args:
dst (`numpy.ndarray`, `cupy.ndarray` or `ideep4py.mdarray`):
Destination array.
src (`numpy.ndarray`, `cupy.ndarray` or `ideep4py.mdarray`):
Source array.
"""
if isinstance(dst, numpy.ndarray):
numpy.copyto(dst, numpy.asarray(cuda.to_cpu(src)))
elif isinstance(dst, intel64.mdarray):
intel64.ideep.basic_copyto(dst, cuda.to_cpu(src))
elif isinstance(dst, cuda.ndarray):
if isinstance(src, chainer.get_cpu_array_types()):
src = numpy.asarray(src)
if dst.flags.c_contiguous or dst.flags.f_contiguous:
dst.set(src)
else:
cuda.cupy.copyto(dst, cuda.to_gpu(src, device=dst.device))
elif isinstance(src, cuda.ndarray):
cuda.cupy.copyto(dst, src)
else:
raise TypeError(
'cannot copy from non-array object of type {}'.format(
type(src)))
else:
raise TypeError('cannot copy to non-array object of type {}'.format(
type(dst)))
def copyto(dst, src):
"""Copies the elements of an ndarray to those of another one.
This function can copy the CPU/GPU arrays to the destination arrays on
another device.
Args:
dst (`numpy.ndarray`, `cupy.ndarray` or `ideep4py.mdarray`):
Destination array.
src (`numpy.ndarray`, `cupy.ndarray` or `ideep4py.mdarray`):
Source array.
"""
if isinstance(dst, numpy.ndarray):
numpy.copyto(dst, _cpu._to_cpu(src))
elif isinstance(dst, intel64.mdarray):
intel64.ideep.basic_copyto(
dst, _cpu._to_cpu(src))
elif isinstance(dst, cuda.ndarray):
if isinstance(src, chainer.get_cpu_array_types()):
src = numpy.asarray(src)
if dst.flags.c_contiguous or dst.flags.f_contiguous:
dst.set(src)
else:
cuda.cupy.copyto(dst, cuda.to_gpu(src, device=dst.device))
elif isinstance(src, cuda.ndarray):
cuda.cupy.copyto(dst, src)
else:
raise TypeError('cannot copy from non-array object of type {}'
.format(type(src)))
else:
raise TypeError('cannot copy to non-array object of type {}'.format(
type(dst)))
def forward(self, inputs):
self.retain_inputs((0, 1))
c_prev, x = inputs
a, i, f, o = _extract_gates(x)
batch = len(x)
if isinstance(x, chainer.get_cpu_array_types()):
if intel64.should_use_ideep('>=auto'):
xp = intel64.ideep.get_array_module(x)
else:
xp = numpy
a = xp.tanh(a)
i = _sigmoid(i, xp)
f = _sigmoid(f, xp)
o = _sigmoid(o, xp)
c_next = numpy.empty_like(c_prev)
c_next[:batch] = a * i + f * c_prev[:batch]
h = o * xp.tanh(c_next[:batch])
else:
c_next = cuda.cupy.empty_like(c_prev)
h = cuda.cupy.empty_like(c_next[:batch])
cuda.elementwise(
'T c_prev, T a, T i_, T f, T o', 'T c, T h',
'''
COMMON_ROUTINE;
c = aa * ai + af * c_prev;
h = ao * tanh(c);
''',
'lstm_fwd', preamble=_preamble)(
c_prev[:batch], a, i, f, o, c_next[:batch], h)
c_next[batch:] = c_prev[batch:]
self.retain_outputs((0,))
return c_next, h
def forward(self, inputs):
self.retain_inputs((0, 1))
c_prev, x = inputs
a, i, f, o = _extract_gates(x)
batch = len(x)
if isinstance(x, chainer.get_cpu_array_types()):
if intel64.should_use_ideep('>=auto'):
xp = intel64.ideep.get_array_module(x)
else:
xp = numpy
a = xp.tanh(a)
i = _sigmoid(i, xp)
f = _sigmoid(f, xp)
o = _sigmoid(o, xp)
c_next = numpy.empty_like(c_prev)
c_next[:batch] = a * i + f * c_prev[:batch]
h = o * xp.tanh(c_next[:batch])
else:
c_next = cuda.cupy.empty_like(c_prev)
h = cuda.cupy.empty_like(c_next[:batch])
cuda.elementwise(
'T c_prev, T a, T i_, T f, T o', 'T c, T h',
'''
COMMON_ROUTINE;
c = aa * ai + af * c_prev;
h = ao * tanh(c);
''',
'lstm_fwd', preamble=_preamble)(
c_prev[:batch], a, i, f, o, c_next[:batch], h)
c_next[batch:] = c_prev[batch:]
self.retain_outputs((0,))
return c_next, h
def _array_to_cpu(array, stream):
if array is None:
return None
if isinstance(array, ndarray):
check_cuda_available()
with get_device_from_array(array):
return array.get(stream)
elif isinstance(array, (numpy.number, numpy.bool_)):
return numpy.asarray(array)
elif isinstance(array, chainer.get_cpu_array_types()):
return array
else:
raise TypeError(
'The array sent to cpu must be numpy.ndarray or cupy.ndarray, '
'or a NumPy scalar.'
'\nActual type: {0}.'.format(type(array)))
def _array_to_cpu(array, stream):
if array is None:
return None
if isinstance(array, ndarray):
check_cuda_available()
with get_device_from_array(array):
return array.get(stream)
elif isinstance(array, (numpy.number, numpy.bool_)):
return numpy.asarray(array)
elif isinstance(array, chainer.get_cpu_array_types()):
return array
else:
raise TypeError(
'The array sent to cpu must be numpy.ndarray or cupy.ndarray, '
'or a NumPy scalar.'
'\nActual type: {0}.'.format(type(array)))
def forward(self, inputs):
cs, x = inputs[:-1], inputs[-1]
n_ary = len(cs)
gates = list(_extract_gates(x, 3 + n_ary))
a, i, o = gates[:3]
fs = gates[3:]
if isinstance(x, chainer.get_cpu_array_types()):
self.a = numpy.tanh(a)
self.i = _sigmoid(i)
self.o = _sigmoid(o)
self.fs = [_sigmoid(f) for f in fs]
self.c = self.a * self.i + sum(f * c for f, c in zip(self.fs, cs))
h = self.o * numpy.tanh(self.c)
else:
preamble = _preamble + \
' '.join('T af{} = sigmoid(f{});'.format(j, j)
for j in six.moves.range(1, n_ary + 1))
cells_str = ', '.join('T c{}'.format(j)
for j in six.moves.range(1, n_ary + 1))
fgates_str = ', '.join('T f{}'.format(j)
for j in six.moves.range(1, n_ary + 1))
fc_calc_str = ' + '.join('af{} * c{}'.format(j, j)
for j in six.moves.range(1, n_ary + 1))
self.c, h = cuda.elementwise('T a, T i_, T o, {}, {}'.format(
cells_str, fgates_str),
'T c, T h',
'''
COMMON_ROUTINE;
c = aa * ai + {};
h = ao * tanh(c);
'''.format(fc_calc_str),
'treelstm_fwd',
preamble=preamble)(a, i, o,
*(list(cs) + fs))
return self.c, h
def forward(self, inputs):
cs, x = inputs[:-1], inputs[-1]
n_ary = len(cs)
gates = list(_extract_gates(x, 3 + n_ary))
a, i, o = gates[:3]
fs = gates[3:]
if isinstance(x, chainer.get_cpu_array_types()):
self.a = numpy.tanh(a)
self.i = _sigmoid(i)
self.o = _sigmoid(o)
self.fs = [_sigmoid(f) for f in fs]
self.c = self.a * self.i + sum(f * c for f, c in zip(self.fs, cs))
h = self.o * numpy.tanh(self.c)
else:
preamble = _preamble + \
' '.join('T af{} = sigmoid(f{});'.format(j, j)
for j in six.moves.range(1, n_ary + 1))
cells_str = ', '.join('T c{}'.format(j)
for j in six.moves.range(1, n_ary + 1))
fgates_str = ', '.join('T f{}'.format(j)
for j in six.moves.range(1, n_ary + 1))
fc_calc_str = ' + '.join('af{} * c{}'.format(j, j)
for j in six.moves.range(1, n_ary + 1))
self.c, h = cuda.elementwise(
'T a, T i_, T o, {}, {}'.format(cells_str, fgates_str),
'T c, T h',
'''
COMMON_ROUTINE;
c = aa * ai + {};
h = ao * tanh(c);
'''.format(fc_calc_str),
'treelstm_fwd', preamble=preamble)(
a, i, o, *(list(cs) + fs))
return self.c, h
