def random_positive_definite_matrix(shape,
dtype,
force_well_conditioned=False):
"""[batch] positive definite matrix.
Args:
shape: `TensorShape` or Python list. Shape of the returned matrix.
dtype: `TensorFlow` `dtype` or Python dtype.
force_well_conditioned: Python bool. If `True`, returned matrix has
eigenvalues with modulus in `(1, 4)`. Otherwise, eigenvalues are
chi-squared random variables.
Returns:
`Tensor` with desired shape and dtype.
"""
dtype = dtypes.as_dtype(dtype)
if not contrib_tensor_util.is_tensor(shape):
shape = tensor_shape.TensorShape(shape)
# Matrix must be square.
shape[-1].assert_is_compatible_with(shape[-2])
with ops.name_scope("random_positive_definite_matrix"):
tril = random_tril_matrix(
shape, dtype, force_well_conditioned=force_well_conditioned)
return math_ops.matmul(tril, tril, adjoint_b=True)
def random_positive_definite_matrix(shape, dtype, force_well_conditioned=False):
"""[batch] positive definite matrix.
Args:
shape: `TensorShape` or Python list. Shape of the returned matrix.
dtype: `TensorFlow` `dtype` or Python dtype.
force_well_conditioned: Python bool. If `True`, returned matrix has
eigenvalues with modulus in `(1, 4)`. Otherwise, eigenvalues are
chi-squared random variables.
Returns:
`Tensor` with desired shape and dtype.
"""
dtype = dtypes.as_dtype(dtype)
if not contrib_tensor_util.is_tensor(shape):
shape = tensor_shape.TensorShape(shape)
# Matrix must be square.
shape[-1].assert_is_compatible_with(shape[-2])
with ops.name_scope("random_positive_definite_matrix"):
tril = random_tril_matrix(shape, dtype, force_well_conditioned=force_well_conditioned)
return math_ops.matmul(tril, tril, adjoint_b=True)