def kaiming_uniform_(tensor, a=0, mode='fan_in', nonlinearity='leaky_relu'):
fan = _calculate_correct_fan(tensor, mode)
gain = calculate_gain(nonlinearity, a)
std = gain / math.sqrt(fan)
bound = math.sqrt(3.0) * std # Calculate uniform bounds from standard deviation
with no_grad():
return tensor.uniform_(-bound, bound)
def _type_to(input, dtype='float32', inplace=False):
if dtype == input._dtype: return input
ctx = MakeContext(inputs=[input])
key = 'torch/ops/astype/{}:{}/dtype:{}/inplace:{}'.format(
ctx[0].lower(), ctx[1], dtype, 'true' if inplace else 'false')
module = get_module(AsType, key, ctx, dtype=dtype, inplace=inplace)
with no_grad():
return module.forward(input)
def dirac_(tensor):
dimensions = tensor.ndimension()
if dimensions not in [3, 4, 5]:
raise ValueError("Only tensors with 3, 4, or 5 dimensions are supported")
sizes = tensor.size()
min_dim = min(sizes[0], sizes[1])
with no_grad():
tensor.zero_()
for d in range(min_dim):
if dimensions == 3: # Temporal convolution
tensor[d, d, tensor.size(2) // 2] = 1
elif dimensions == 4: # Spatial convolution
tensor[d, d, tensor.size(2) // 2, tensor.size(3) // 2] = 1
else: # Volumetric convolution
tensor[d, d, tensor.size(2) // 2, tensor.size(3) // 2, tensor.size(4) // 2] = 1
return tensor
def constant_(tensor, val):
with no_grad():
return tensor.fill_(val)
def normal_(tensor, mean=0, std=1):
with no_grad():
return tensor.normal_(mean, std)
def uniform_(tensor, a=0, b=1):
with no_grad():
return tensor.uniform_(a, b)
def kaiming_normal_(tensor, a=0, mode='fan_in', nonlinearity='leaky_relu'):
fan = _calculate_correct_fan(tensor, mode)
gain = calculate_gain(nonlinearity, a)
std = gain / math.sqrt(fan)
with no_grad():
return tensor.normal_(0, std)
def xavier_normal_(tensor, gain=1):
fan_in, fan_out = _calculate_fan_in_and_fan_out(tensor)
std = gain * math.sqrt(2.0 / (fan_in + fan_out))
with no_grad():
return tensor.normal_(0, std)
def xavier_uniform_(tensor, gain=1):
fan_in, fan_out = _calculate_fan_in_and_fan_out(tensor)
std = gain * math.sqrt(2.0 / (fan_in + fan_out))
a = math.sqrt(3.0) * std # Calculate uniform bounds from standard deviation
with no_grad():
return tensor.uniform_(-a, a)