def __init__(self,
in_channels,
out_channels,
kernel_size,
stride=1,
pad_mode='same',
padding=0,
dilation=1,
group=1,
has_bias=False,
weight_init='normal',
bias_init='zeros'):
super(DepthwiseConv2d, self).__init__()
self.kernel_size = twice(kernel_size)
self.stride = twice(stride)
self.dilation = twice(dilation)
self.in_channels = check_int_positive(in_channels)
self.out_channels = check_int_positive(out_channels)
validator.check_integer('group', group, in_channels, Rel.EQ)
validator.check_integer('group', group, out_channels, Rel.EQ)
validator.check_integer('group', group, 1, Rel.GE)
self.pad_mode = pad_mode
self.padding = padding
self.dilation = dilation
self.group = group
self.has_bias = has_bias
self.conv = P.DepthwiseConv2dNative(channel_multiplier=1,
kernel_size=self.kernel_size,
pad_mode=self.pad_mode,
pad=self.padding,
stride=self.stride,
dilation=self.dilation)
self.bias_add = P.BiasAdd()
weight_shape = [1, in_channels, *self.kernel_size]
self.weight = Parameter(initializer(weight_init, weight_shape),
name='weight')
if check_bool(has_bias):
self.bias = Parameter(initializer(bias_init, [out_channels]),
name='bias')
else:
if bias_init != 'zeros':
logger.warning(
"value of `has_bias` is False, value of `bias_init` will be ignore."
)
self.bias = None
def __init__(self,
in_channel,
out_channel,
in_drop_ratio=0.0,
coef_drop_ratio=0.0,
residual=False,
coef_activation=nn.LeakyReLU(),
activation=nn.ELU()):
super(AttentionHead, self).__init__()
self.in_channel = check_int_positive(in_channel)
self.out_channel = check_int_positive(out_channel)
self.in_drop_ratio = in_drop_ratio
self.in_drop = nn.Dropout(keep_prob=1 - in_drop_ratio)
self.in_drop_2 = nn.Dropout(keep_prob=1 - in_drop_ratio)
self.feature_transform = GNNFeatureTransform(
in_channels=self.in_channel,
out_channels=self.out_channel,
has_bias=False,
weight_init='XavierUniform')
self.f_1_transform = GNNFeatureTransform(
in_channels=self.out_channel,
out_channels=1,
weight_init='XavierUniform')
self.f_2_transform = GNNFeatureTransform(
in_channels=self.out_channel,
out_channels=1,
weight_init='XavierUniform')
self.softmax = nn.Softmax()
self.coef_drop = nn.Dropout(keep_prob=1 - coef_drop_ratio)
self.matmul = P.MatMul()
self.bias_add = P.BiasAdd()
self.bias = Parameter(initializer('zeros', self.out_channel), name='bias')
self.residual = check_bool(residual)
if self.residual:
if in_channel != out_channel:
self.residual_transform_flag = True
self.residual_transform = GNNFeatureTransform(
in_channels=self.in_channel,
out_channels=self.out_channel)
else:
self.residual_transform = None
self.coef_activation = coef_activation
self.activation = activation
def __init__(self,
in_channels,
out_channels,
kernel_size,
stride,
pad_mode,
padding,
dilation,
group,
has_bias,
weight_init,
bias_init):
super(_Conv, self).__init__()
self.in_channels = check_int_positive(in_channels)
self.out_channels = check_int_positive(out_channels)
self.kernel_size = kernel_size
self.stride = check_int_positive(stride)
self.pad_mode = pad_mode
self.padding = check_int_non_negative(padding)
self.dilation = check_int(dilation)
self.group = check_int_positive(group)
self.has_bias = has_bias
if (not isinstance(kernel_size, tuple)) or len(kernel_size) != 2 or \
(not isinstance(kernel_size[0], int)) or (not isinstance(kernel_size[1], int)) or \
kernel_size[0] < 1 or kernel_size[1] < 1:
raise ValueError("Attr 'kernel_size' of 'Conv2D' Op passed "
+ str(self.kernel_size) + ", should be a int or tuple and equal to or greater than 1.")
if in_channels % group != 0:
raise ValueError("Attr 'in_channels' of 'Conv2D' Op must be divisible by "
"attr 'group' of 'Conv2D' Op.")
if out_channels % group != 0:
raise ValueError("Attr 'out_channels' of 'Conv2D' Op must be divisible by "
"attr 'group' of 'Conv2D' Op.")
self.weight = Parameter(initializer(weight_init, [out_channels, in_channels // group, *kernel_size]),
name='weight')
if check_bool(has_bias):
self.bias = Parameter(initializer(bias_init, [out_channels]), name='bias')
else:
if bias_init != 'zeros':
logger.warning("Value of 'has_bias' is False, value of 'bias_init' will be ignored.")
self.bias = None
def __init__(self,
save_checkpoint_steps=1,
save_checkpoint_seconds=0,
keep_checkpoint_max=5,
keep_checkpoint_per_n_minutes=0,
integrated_save=True,
model_type="normal"):
if not save_checkpoint_steps and not save_checkpoint_seconds and \
not keep_checkpoint_max and not keep_checkpoint_per_n_minutes:
raise ValueError("The input_param can't be all None or 0")
if save_checkpoint_steps:
save_checkpoint_steps = check_int_non_negative(
save_checkpoint_steps)
if save_checkpoint_seconds:
save_checkpoint_seconds = check_int_non_negative(
save_checkpoint_seconds)
if keep_checkpoint_max:
keep_checkpoint_max = check_int_non_negative(keep_checkpoint_max)
if keep_checkpoint_per_n_minutes:
keep_checkpoint_per_n_minutes = check_int_non_negative(
keep_checkpoint_per_n_minutes)
if model_type:
model_type = check_string(model_type,
["normal", "fusion", "quant"])
self._save_checkpoint_steps = save_checkpoint_steps
self._save_checkpoint_seconds = save_checkpoint_seconds
if self._save_checkpoint_steps and self._save_checkpoint_steps > 0:
self._save_checkpoint_seconds = None
self._keep_checkpoint_max = keep_checkpoint_max
self._keep_checkpoint_per_n_minutes = keep_checkpoint_per_n_minutes
if self._keep_checkpoint_max and self._keep_checkpoint_max > 0:
self._keep_checkpoint_per_n_minutes = None
else:
if not self._keep_checkpoint_per_n_minutes or self._keep_checkpoint_per_n_minutes == 0:
self._keep_checkpoint_max = 1
self._model_type = model_type
self._integrated_save = check_bool(integrated_save)
def __init__(self,
params,
learning_rate,
momentum,
weight_decay=0.0,
loss_scale=1.0,
use_nesterov=False):
super(Momentum, self).__init__(learning_rate, params, weight_decay,
loss_scale)
if isinstance(momentum, float) and momentum < 0.0:
raise ValueError(
"momentum should be at least 0.0, but got momentum {}".format(
momentum))
self.momentum = Parameter(Tensor(momentum, mstype.float32),
name="momentum")
self.params = self.parameters
self.use_nesterov = check_bool(use_nesterov)
self.moments = self.params.clone(prefix="moments", init='zeros')
self.hyper_map = C.HyperMap()
self.opt = P.ApplyMomentum(use_nesterov=self.use_nesterov)
def __init__(
self,
in_channels,
out_channels,
activation=None,
has_bias=True,
weight_prior_fn=NormalPrior,
weight_posterior_fn=lambda name, shape: NormalPosterior(name=name, shape=shape),
bias_prior_fn=NormalPrior,
bias_posterior_fn=lambda name, shape: NormalPosterior(name=name, shape=shape)):
super(_DenseVariational, self).__init__()
self.in_channels = check_int_positive(in_channels)
self.out_channels = check_int_positive(out_channels)
self.has_bias = check_bool(has_bias)
if isinstance(weight_prior_fn, Cell):
self.weight_prior = weight_prior_fn
else:
self.weight_prior = weight_prior_fn()
self.weight_posterior = weight_posterior_fn(shape=[self.out_channels, self.in_channels], name='bnn_weight')
if self.has_bias:
if isinstance(bias_prior_fn, Cell):
self.bias_prior = bias_prior_fn
else:
self.bias_prior = bias_prior_fn()
self.bias_posterior = bias_posterior_fn(shape=[self.out_channels], name='bnn_bias')
self.activation = activation
if isinstance(self.activation, str):
self.activation = get_activation(activation)
self.activation_flag = self.activation is not None
self.matmul = P.MatMul(transpose_b=True)
self.bias_add = P.BiasAdd()
self.sum = P.ReduceSum()
def __init__(self, num_groups, num_channels, eps=1e-05, affine=True, gamma_init='ones', beta_init='zeros'):
super(GroupNorm, self).__init__()
self.num_groups = check_int_positive(num_groups)
self.num_channels = check_int_positive(num_channels)
if num_channels % num_groups != 0:
raise ValueError("num_channels should be divided by num_groups")
self.eps = check_typename('eps', eps, (float,))
self.affine = check_bool(affine)
gamma = initializer(gamma_init, [num_channels, 1, 1])
beta = initializer(beta_init, [num_channels, 1, 1])
if self.affine:
self.gamma = Parameter(gamma, name='gamma')
self.beta = Parameter(beta, name='beta')
else:
self.gamma = gamma
self.beta = beta
self.shape = F.shape
self.reshape = F.reshape
self.reduce_mean = P.ReduceMean(keep_dims=True)
self.square = F.square
self.reduce_sum = P.ReduceSum(keep_dims=True)
self.sqrt = P.Sqrt()
def __init__(self,
model,
train_dataset,
task_type,
num_classes=None,
epochs=1,
epi_uncer_model_path=None,
ale_uncer_model_path=None,
save_model=False):
self.epi_model = model
self.ale_model = deepcopy(model)
self.epi_train_dataset = train_dataset
self.ale_train_dataset = deepcopy(train_dataset)
self.task_type = task_type
self.epochs = check_int_positive(epochs)
self.epi_uncer_model_path = epi_uncer_model_path
self.ale_uncer_model_path = ale_uncer_model_path
self.save_model = check_bool(save_model)
self.epi_uncer_model = None
self.ale_uncer_model = None
self.concat = P.Concat(axis=0)
self.sum = P.ReduceSum()
self.pow = P.Pow()
if not isinstance(model, Cell):
raise TypeError('The model should be Cell type.')
if task_type not in ('regression', 'classification'):
raise ValueError(
'The task should be regression or classification.')
if task_type == 'classification':
self.num_classes = check_int_positive(num_classes)
else:
self.num_classes = num_classes
if save_model:
if epi_uncer_model_path is None or ale_uncer_model_path is None:
raise ValueError(
"If save_model is True, the epi_uncer_model_path and "
"ale_uncer_model_path should not be None.")
def __init__(self,
save_checkpoint_steps=1,
save_checkpoint_seconds=0,
keep_checkpoint_max=5,
keep_checkpoint_per_n_minutes=0,
integrated_save=True):
if not save_checkpoint_steps and not save_checkpoint_seconds and \
not keep_checkpoint_max and not keep_checkpoint_per_n_minutes:
raise ValueError("The input_param can't be all None or 0")
if save_checkpoint_steps:
save_checkpoint_steps = check_int_non_negative(
save_checkpoint_steps)
if save_checkpoint_seconds:
save_checkpoint_seconds = check_int_non_negative(
save_checkpoint_seconds)
if keep_checkpoint_max:
keep_checkpoint_max = check_int_non_negative(keep_checkpoint_max)
if keep_checkpoint_per_n_minutes:
keep_checkpoint_per_n_minutes = check_int_non_negative(
keep_checkpoint_per_n_minutes)
self._save_checkpoint_steps = save_checkpoint_steps
self._save_checkpoint_seconds = save_checkpoint_seconds
if self._save_checkpoint_steps and self._save_checkpoint_steps > 0:
self._save_checkpoint_seconds = None
self._keep_checkpoint_max = keep_checkpoint_max
self._keep_checkpoint_per_n_minutes = keep_checkpoint_per_n_minutes
if self._keep_checkpoint_max and self._keep_checkpoint_max > 0:
self._keep_checkpoint_per_n_minutes = None
else:
if not self._keep_checkpoint_per_n_minutes or self._keep_checkpoint_per_n_minutes == 0:
self._keep_checkpoint_max = 1
self._integrated_save = check_bool(integrated_save)
def __init__(self,
in_channels,
out_channels,
kernel_size,
stride,
pad_mode,
padding,
dilation,
group,
has_bias,
weight_init,
bias_init,
transposed=False):
super(_Conv, self).__init__()
self.in_channels = check_int_positive(in_channels)
self.out_channels = check_int_positive(out_channels)
self.kernel_size = kernel_size
self.stride = stride
self.pad_mode = pad_mode
self.weight_init = weight_init
self.bias_init = bias_init
if isinstance(padding, int):
Validator.check_integer('padding', padding, 0, Rel.GE, self.cls_name)
self.padding = padding
elif isinstance(padding, tuple):
for pad in padding:
Validator.check_integer('padding item', pad, 0, Rel.GE, self.cls_name)
self.padding = padding
else:
raise TypeError("padding type must be int/tuple(int) cannot be {}!".format(type(padding)))
self.dilation = dilation
self.group = check_int_positive(group)
self.has_bias = has_bias
if (not isinstance(kernel_size[0], int)) or (not isinstance(kernel_size[1], int)) or \
kernel_size[0] < 1 or kernel_size[1] < 1:
raise ValueError("Attr 'kernel_size' of 'Conv2D' Op passed "
+ str(self.kernel_size) + ", should be a int or tuple and equal to or greater than 1.")
if (not isinstance(stride[0], int)) or (not isinstance(stride[1], int)) or stride[0] < 1 or stride[1] < 1:
raise ValueError("Attr 'stride' of 'Conv2D' Op passed "
+ str(self.stride) + ", should be a int or tuple and equal to or greater than 1.")
if (not isinstance(dilation[0], int)) or (not isinstance(dilation[1], int)) or \
dilation[0] < 1 or dilation[1] < 1:
raise ValueError("Attr 'dilation' of 'Conv2D' Op passed "
+ str(self.dilation) + ", should equal to or greater than 1.")
if in_channels % group != 0:
raise ValueError("Attr 'in_channels' of 'Conv2D' Op must be divisible by "
"attr 'group' of 'Conv2D' Op.")
if out_channels % group != 0:
raise ValueError("Attr 'out_channels' of 'Conv2D' Op must be divisible by "
"attr 'group' of 'Conv2D' Op.")
if transposed:
shape = [in_channels, out_channels // group, *kernel_size]
else:
shape = [out_channels, in_channels // group, *kernel_size]
self.weight = Parameter(initializer(self.weight_init, shape), name='weight')
if check_bool(has_bias):
self.bias = Parameter(initializer(self.bias_init, [out_channels]), name='bias')
else:
if self.bias_init != 'zeros':
logger.warning("Value of 'has_bias' is False, value of 'bias_init' will be ignored.")
self.bias = None
def __init__(self, dataset, dataset_sink_mode=True, iter_first_order=0):
check_bool(dataset_sink_mode)
self.iter = _DatasetIterMSLoopSink(dataset, iter_first_order)
def __init__(self,
in_channels,
out_channels,
weight_init='normal',
bias_init='zeros',
damping=0.03,
loss_scale=1,
frequency=278,
has_bias=True,
activation=None):
super(Dense_Thor, self).__init__()
self.in_channels = check_int_positive(in_channels)
self.out_channels = check_int_positive(out_channels)
self.has_bias = check_bool(has_bias)
self.thor = True
if isinstance(weight_init, Tensor):
if weight_init.dim() != 2 or weight_init.shape[0] != out_channels or \
weight_init.shape[1] != in_channels:
raise ValueError("weight_init shape error")
self.weight = Parameter(initializer(weight_init, [out_channels, in_channels]), name="weight")
if self.has_bias:
if isinstance(bias_init, Tensor):
if bias_init.dim() != 1 or bias_init.shape[0] != out_channels:
raise ValueError("bias_init shape error")
self.bias = Parameter(initializer(bias_init, [out_channels]), name="bias")
self.matmul = P.MatMul(transpose_b=True)
self.bias_add = P.BiasAdd()
self.activation = get_activation(activation)
self.activation_flag = self.activation is not None
self.matrix_A_inv = Parameter(Tensor(np.zeros([128, 128, 16, 16]).astype(np.float16)), name='matrix_A_inv',
requires_grad=False)
self.matrix_G_inv = Parameter(Tensor(np.zeros([63, 63, 16, 16]).astype(np.float16)), name="matrix_G_inv",
requires_grad=False)
self.fake_G = Tensor(np.zeros([63, 63, 16, 16]).astype(np.float16))
self.matmul = P.MatMul(transpose_b=True)
self.cube_matmul = P.CusMatMulCube(transpose_a=True)
self.matrix_combine = P.CusMatrixCombine()
self.cholesky = P.CusCholeskyTrsm()
self.shape = P.Shape()
self.reshape = P.Reshape()
self.transpose = P.Transpose()
self.cov_step = Parameter(initializer(0, [1], mstype.int32), name="cov_step", requires_grad=False)
self.mul = P.Mul()
self.cast = P.Cast()
self.damping = Tensor(damping)
self.loss_scale = Tensor(1 / loss_scale, mstype.float16)
self.vector_matmul = P.CusBatchMatMul()
self.pad = P.Pad(((0, 24), (0, 24)))
self.pad1 = P.Pad(((0, 8), (0, 8)))
self.slice = P.Slice()
self.gather = P.GatherV2()
self.assignadd = P.AssignAdd()
self.freq = Tensor(frequency, mstype.int32)
self.axis = 0
self.A_inv_max = Parameter(initializer(0, [1], mstype.float32), name="A_inv_max", requires_grad=False)
self.G_inv_max = Parameter(initializer(0, [1], mstype.float32), name="G_inv_max", requires_grad=False)
self.fused_abs_max1 = P.CusFusedAbsMax1([1000, 1000])
self.fused_abs_max2 = P.CusFusedAbsMax1()
self.log = P.Log()
self.exp = P.Exp()
self.dampingA = Tensor(np.identity(2048), mstype.float32)
self.dampingG = Tensor(np.identity(1024), mstype.float32)
self.add = P.TensorAdd()
self.sqrt = P.Sqrt()
self.getG = P.InsertGradientOf(self.save_gradient)
def test_check_bool_1():
assert check_bool(True)
def test_check_bool_3():
with pytest.raises(TypeError):
check_bool("str")
def __init__(self,
params,
learning_rate,
momentum,
matrix_A,
matrix_G,
A_inv_max,
G_inv_max,
weight_decay=0.0,
loss_scale=1.0,
use_nesterov=False,
decay_filter=lambda x: x.name not in []):
super(THOR_GPU, self).__init__(learning_rate, params, weight_decay,
loss_scale)
validator.check_value_type("momentum", momentum, [float],
self.cls_name)
if isinstance(momentum, float) and momentum < 0.0:
raise ValueError(
"momentum should be at least 0.0, but got momentum {}".format(
momentum))
self.momentum = Parameter(Tensor(momentum, mstype.float32),
name="momentum")
self.params = self.parameters
self.use_nesterov = check_bool(use_nesterov)
self.moments = self.params.clone(prefix="moments", init='zeros')
self.hyper_map = C.HyperMap()
self.opt = P.ApplyMomentum(use_nesterov=self.use_nesterov)
self.feature_map = [
1.0 / 12544, 1.0 / 3136, 1.0 / 3136, 1.0 / 3136, 1.0 / 3136,
1.0 / 3136, 1.0 / 3136, 1.0 / 3136, 1.0 / 3136, 1.0 / 3136,
1.0 / 3136, 1.0 / 3136, 1.0 / 784, 1.0 / 784, 1.0 / 784, 1.0 / 784,
1.0 / 784, 1.0 / 784, 1.0 / 784, 1.0 / 784, 1.0 / 784, 1.0 / 784,
1.0 / 784, 1.0 / 784, 1.0 / 784, 1.0 / 196, 1.0 / 196, 1.0 / 196,
1.0 / 196, 1.0 / 196, 1.0 / 196, 1.0 / 196, 1.0 / 196, 1.0 / 196,
1.0 / 196, 1.0 / 196, 1.0 / 196, 1.0 / 196, 1.0 / 196, 1.0 / 196,
1.0 / 196, 1.0 / 196, 1.0 / 196, 1.0 / 196, 1.0 / 49, 1.0 / 49,
1.0 / 49, 1.0 / 49, 1.0 / 49, 1.0 / 49, 1.0 / 49, 1.0 / 49,
1.0 / 49, 1.0
]
self.feature_map_new = [x**0.5 for x in self.feature_map]
self.transpose = P.Transpose()
self.shape = P.Shape()
self.reshape = P.Reshape()
self.matmul = P.MatMul()
self.matrix_A = ParameterTuple(matrix_A)
self.matrix_G = ParameterTuple(matrix_G)
self.A_inv_max = ParameterTuple(A_inv_max)
self.G_inv_max = ParameterTuple(G_inv_max)
self.assign = P.Assign()
self.mul = P.Mul()
mean = _get_gradients_mean()
degree = _get_device_num()
parameter_length = len(self.feature_map)
self.grad_reducer_thorA = DistributedGradReducerThor(
parameter_length, ((parameter_length, ), 0), mean, degree)
self.grad_reducer_thorG = DistributedGradReducerThor(
parameter_length, ((parameter_length, ), 0), mean, degree)
self.weight_decay = weight_decay
self.decay_flags = tuple(decay_filter(x) for x in self.parameters)
self.update_gradient = P.UpdateThorGradient(split_dim=128)
def __init__(self,
in_channels,
out_channels,
weight_init='normal',
bias_init='zeros',
damping=0.03,
loss_scale=1,
frequency=278,
batch_size=32,
has_bias=True,
activation=None):
super(Dense_Thor_GPU, self).__init__()
self.in_channels = check_int_positive(in_channels)
self.out_channels = check_int_positive(out_channels)
self.has_bias = check_bool(has_bias)
self.thor = True
if isinstance(weight_init, Tensor):
if weight_init.dim() != 2 or weight_init.shape[0] != out_channels or \
weight_init.shape[1] != in_channels:
raise ValueError("weight_init shape error")
self.weight = Parameter(initializer(weight_init,
[out_channels, in_channels]),
name="weight")
if self.has_bias:
if isinstance(bias_init, Tensor):
if bias_init.dim() != 1 or bias_init.shape[0] != out_channels:
raise ValueError("bias_init shape error")
self.bias = Parameter(initializer(bias_init, [out_channels]),
name="bias")
self.matmul = P.MatMul(transpose_b=True)
self.bias_add = P.BiasAdd()
self.activation = get_activation(activation)
self.activation_flag = self.activation is not None
split_dim = 128
matrix_A_shape, matrix_G_shape = caculate_matmul_shape(
self.in_channels, self.out_channels, split_dim)
self.matrix_A_inv = Parameter(Tensor(
np.zeros(matrix_A_shape).astype(np.float32)),
name='matrix_A_inv',
requires_grad=False)
self.matrix_G_inv = Parameter(Tensor(
np.zeros(matrix_G_shape).astype(np.float32)),
name="matrix_G_inv",
requires_grad=False)
self.broadcast_to = P.BroadcastTo(matrix_A_shape)
self.cov_step = Parameter(initializer(0, [1], mstype.int32),
name="cov_step",
requires_grad=False)
self.shape = P.Shape()
self.reshape = P.Reshape()
self.transpose = P.Transpose()
self.mul = P.Mul()
self.cube_matmul = P.MatMul(transpose_a=True)
self.loss_scale = Tensor(1 / loss_scale, mstype.float16)
self.batch_size = Tensor(batch_size, mstype.float16)
self.getG = P.InsertGradientOf(self.save_gradient)
self.damping = Parameter(Tensor(damping),
name="damping_value",
requires_grad=False)
self.dampingA = Tensor(np.identity(in_channels), mstype.float32)
self.dampingG = Tensor(np.identity(out_channels), mstype.float32)
self.cast = P.Cast()
self.gather = P.GatherV2()
self.freq = Tensor(frequency, mstype.int32)
self.axis = 0
self.add = P.TensorAdd()
self.sqrt = P.Sqrt()
self.cholesky = P.Cholesky(split_dim=split_dim)
self.vector_matmul = P.BatchMatMul(transpose_a=True)
def __init__(self,
in_channels,
out_channels,
activation=None,
has_bias=True,
weight_prior_fn=NormalPrior,
weight_posterior_fn=lambda name, shape: NormalPosterior(
name=name, shape=shape),
bias_prior_fn=NormalPrior,
bias_posterior_fn=lambda name, shape: NormalPosterior(
name=name, shape=shape)):
super(_DenseVariational, self).__init__()
self.in_channels = check_int_positive(in_channels)
self.out_channels = check_int_positive(out_channels)
self.has_bias = check_bool(has_bias)
if isinstance(weight_prior_fn, Cell):
self.weight_prior = weight_prior_fn
else:
self.weight_prior = weight_prior_fn()
for prior_name, prior_dist in self.weight_prior.name_cells().items():
if prior_name != 'normal':
raise TypeError(
"The type of distribution of `weight_prior_fn` should be `normal`"
)
if not (isinstance(getattr(prior_dist, '_mean_value'), Tensor)
and isinstance(getattr(prior_dist, '_sd_value'), Tensor)):
raise TypeError(
"The input form of `weight_prior_fn` is incorrect")
try:
self.weight_posterior = weight_posterior_fn(
shape=[self.out_channels, self.in_channels], name='bnn_weight')
except TypeError:
raise TypeError(
'The type of `weight_posterior_fn` should be `NormalPosterior`'
)
for posterior_name, _ in self.weight_posterior.name_cells().items():
if posterior_name != 'normal':
raise TypeError(
"The type of distribution of `weight_posterior_fn` should be `normal`"
)
if self.has_bias:
if isinstance(bias_prior_fn, Cell):
self.bias_prior = bias_prior_fn
else:
self.bias_prior = bias_prior_fn()
for prior_name, prior_dist in self.bias_prior.name_cells().items():
if prior_name != 'normal':
raise TypeError(
"The type of distribution of `bias_prior_fn` should be `normal`"
)
if not (isinstance(getattr(prior_dist, '_mean_value'), Tensor)
and isinstance(getattr(prior_dist, '_sd_value'),
Tensor)):
raise TypeError(
"The input form of `bias_prior_fn` is incorrect")
try:
self.bias_posterior = bias_posterior_fn(
shape=[self.out_channels], name='bnn_bias')
except TypeError:
raise TypeError(
'The type of `bias_posterior_fn` should be `NormalPosterior`'
)
for posterior_name, _ in self.bias_posterior.name_cells().items():
if posterior_name != 'normal':
raise TypeError(
"The type of distribution of `bias_posterior_fn` should be `normal`"
)
self.activation = activation
if not self.activation:
self.activation_flag = False
else:
self.activation_flag = True
if isinstance(self.activation, str):
self.activation = get_activation(activation)
elif isinstance(self.activation, Cell):
self.activation = activation
else:
raise ValueError('The type of `activation` is wrong.')
self.matmul = P.MatMul(transpose_b=True)
self.bias_add = P.BiasAdd()
self.sum = P.ReduceSum()
def train(self,
epoch,
train_dataset,
callbacks=None,
dataset_sink_mode=True):
"""
Training API where the iteration is controlled by python front-end.
When setting pynative mode, the training process will be performed with dataset not sink.
Note:
CPU is not supported when dataset_sink_mode is true.
If dataset_sink_mode is True, epoch of training should be equal to the count of repeat
operation in dataset processing. Otherwise, errors could occur since the amount of data
is not the amount training requires.
If dataset_sink_mode is True, data will be sent to device. If device is Ascend, features
of data will be transferred one by one. The limitation of data transmission per time is 256M.
Args:
epoch (int): Total number of iterations on the data.
train_dataset (Dataset): A training dataset iterator. If there is no
loss_fn, a tuple with multiply data (data1, data2, data3, ...) should be
returned and passed to the network. Otherwise, a tuple (data, label) should
be returned, and the data and label are passed to the network and loss
function respectively.
callbacks (list): List of callback object. Callbacks which should be excuted while training. Default: None.
dataset_sink_mode (bool): Determines whether to pass the data through dataset channel. Default: True.
Configure pynative mode, the training process will be performed with
dataset not sink.
Examples:
>>> dataset = get_dataset()
>>> net = Net()
>>> loss = nn.SoftmaxCrossEntropyWithLogits()
>>> loss_scale_manager = FixedLossScaleManager()
>>> optim = Momentum(params=net.trainable_params(), learning_rate=0.1, momentum=0.9)
>>> model = Model(net, loss_fn=loss, optimizer=optim, metrics=None, loss_scale_manager=loss_scale_manager)
>>> model.train(2, dataset)
"""
repeat_count = train_dataset.get_repeat_count()
if epoch != repeat_count and dataset_sink_mode is True:
logger.warning(
f"The epoch_size {epoch} is not the same with dataset repeat_count {repeat_count}"
)
check_bool(dataset_sink_mode)
_device_number_check(self._parallel_mode, self._device_number)
_parameter_broadcast_check(self._parallel_mode,
self._parameter_broadcast)
if context.get_context("device_target") in [
"CPU", "GPU"
] and context.get_context("enable_loop_sink"):
raise ValueError(
"CPU and GPU can't support loop sink, please set enable_loop_sink=False."
)
self._train(epoch,
train_dataset,
callbacks=callbacks,
dataset_sink_mode=dataset_sink_mode)
def test_check_bool_5():
with pytest.raises(TypeError):
check_bool(3.5)
def __init__(self,
in_channels,
out_channels,
kernel_size,
stride=1,
pad_mode='same',
padding=0,
dilation=1,
group=1,
has_bias=False,
weight_init='normal',
bias_init='zeros'):
kernel_size = twice(kernel_size)
stride = twice(stride)
dilation = twice(dilation)
Validator.check_value_type('padding', padding, (int, tuple), self.cls_name)
if isinstance(padding, tuple):
Validator.check_integer('padding size', len(padding), 4, Rel.EQ, self.cls_name)
# out_channels and in_channels swap.
# cause Conv2DBackpropInput's out_channel refers to Conv2D's out_channel,
# then Conv2dTranspose's out_channel refers to Conv2DBackpropInput's in_channel.
super(Conv2dTranspose, self).__init__(
in_channels,
out_channels,
kernel_size,
stride,
pad_mode,
padding,
dilation,
group,
has_bias,
weight_init,
bias_init,
transposed=True)
self.in_channels = in_channels
self.out_channels = out_channels
self.shape = P.Shape()
if pad_mode not in ('valid', 'same', 'pad'):
raise ValueError('Attr \'pad_mode\' of \'Conv2dTranspose\' Op passed '
+ str(pad_mode) + ', should be one of values in \'valid\', \'same\', \'pad\'.')
self.is_valid = self.pad_mode == 'valid'
self.is_same = self.pad_mode == 'same'
self.is_pad = self.pad_mode == 'pad'
if check_bool(has_bias):
self.bias = Parameter(initializer(bias_init, [out_channels]), name='bias')
# cause Conv2DBackpropInput's out_channel refers to Conv2D's out_channel.
self.conv2d_transpose = P.Conv2DBackpropInput(out_channel=in_channels,
kernel_size=kernel_size,
mode=1,
pad_mode=pad_mode,
pad=padding,
stride=stride,
dilation=dilation,
group=group)
self.bias_add = P.BiasAdd()
if isinstance(self.padding, int):
self.padding_top, self.padding_bottom, self.padding_left, self.padding_right = (self.padding,) * 4
else:
self.padding_top, self.padding_bottom, self.padding_left, self.padding_right = self.padding
def __init__(self,
in_channels,
out_channels,
kernel_size,
stride=1,
pad_mode='same',
padding=0,
dilation=1,
group=1,
has_bias=False,
weight_init='normal',
bias_init='zeros'):
Validator.check_value_type("kernel_size", kernel_size, [int], self.cls_name)
Validator.check_value_type("stride", stride, [int], self.cls_name)
Validator.check_value_type("padding", padding, [int], self.cls_name)
Validator.check_value_type("dilation", dilation, [int], self.cls_name)
Validator.check_integer('kernel_size', kernel_size, 1, Rel.GE, self.cls_name)
Validator.check_integer('stride', stride, 1, Rel.GE, self.cls_name)
Validator.check_integer('padding', padding, 0, Rel.GE, self.cls_name)
Validator.check_integer('dilation', dilation, 1, Rel.GE, self.cls_name)
kernel_size = (1, kernel_size)
stride = (1, stride)
dilation = (1, dilation)
get_shape = P.Shape()
get_dtype = P.DType()
if isinstance(weight_init, Tensor):
weight_init_shape = get_shape(weight_init)
Validator.check_integer('weight_init_shape', len(weight_init_shape), 3, Rel.EQ, self.cls_name)
weight_init_dtype = get_dtype(weight_init)
weight_init_value = weight_init.asnumpy()
weight_init_value = np.expand_dims(weight_init_value, 2)
weight_init = Tensor(weight_init_value, weight_init_dtype)
# out_channels and in_channels swap.
# cause Conv2DBackpropInput's out_channel refers to Conv2D's out_channel,
# then Conv1dTranspose's out_channel refers to Conv2DBackpropInput's in_channel.
super(Conv1dTranspose, self).__init__(
in_channels,
out_channels,
kernel_size,
stride,
pad_mode,
padding,
dilation,
group,
has_bias,
weight_init,
bias_init,
transposed=True)
self.padding = (0, 0, padding, padding)
self.in_channels = in_channels
self.out_channels = out_channels
self.shape = P.Shape()
if pad_mode not in ('valid', 'same', 'pad'):
raise ValueError('Attr \'pad_mode\' of \'Conv1dTranspose\' Op passed '
+ str(pad_mode) + ', should be one of values in \'valid\', \'same\', \'pad\'.')
self.is_valid = self.pad_mode == 'valid'
self.is_same = self.pad_mode == 'same'
self.is_pad = self.pad_mode == 'pad'
if check_bool(has_bias):
self.bias = Parameter(initializer(bias_init, [out_channels]), name='bias')
# cause Conv2DBackpropInput's out_channel refers to Conv2D's out_channel.
self.conv2d_transpose = P.Conv2DBackpropInput(out_channel=in_channels,
kernel_size=kernel_size,
mode=1,
pad_mode=pad_mode,
pad=self.padding,
stride=stride,
dilation=dilation,
group=group)
self.bias_add = P.BiasAdd()
self.expand_dims = P.ExpandDims()
self.squeeze = P.Squeeze(2)
def test_check_bool_2():
assert check_bool(False) is not True
