lab.generic module¶

lab.generic.abs[source]¶

Absolute value.

Parameters: a (tensor) – Tensor.
Returns: Absolute value of a.
Return type: tensor

lab.generic.add[source]¶

Add two tensors.

Parameters

a (tensor) – First tensor.
b (tensor) – Second tensor.

Returns

Sum of a and b.

Return type

tensor

lab.generic.all[source]¶

Logical all of a tensor, possibly along an axis.

Parameters

a (tensor) – Tensor.
axis (int, optional) – Optional axis.

Returns

Reduced tensor.

Return type

tensor

lab.generic.any[source]¶

Logical any of a tensor, possibly along an axis.

Parameters

a (tensor) – Tensor.
axis (int, optional) – Optional axis.

Returns

Reduced tensor.

Return type

tensor

lab.generic.argsort[source]¶

Get the indices that would a tensor along an axis in ascending order.

Parameters

a (tensor) – Tensor to sort.
axis (int, optional) – Axis to sort along. Defaults to -1.
descending (bool, optional) – Sort in descending order. Defaults to False.

Returns

The indices that would sort a.

Return type

tensor

lab.generic.bvn_cdf[source]¶

Standard bivariate normal CDF. Computes the probability that X < a and Y < b if X ~ N(0, 1), Y ~ N(0, 1), and X and Y have correlation c.

Parameters

a (tensor) – First upper limit. Must be a rank-one tensor.
b (tensor) – Second upper limit. Must be a rank-one tensor.
c (tensor) – Correlation coefficient. Must be a rank-one tensor.

Returns

Probabilities of the same shape as the input.

Return type

tensor

lab.generic.cast[source]¶

Cast an object to another data type.

Parameters

dtype (dtype) – New data type.
a (tensor) – Tensor to cast.

Returns

a, but of data type dtype.

Return type

tensor

lab.generic.cos[source]¶

Cosine function.

Parameters: a (tensor) – Tensor.
Returns: Cosine function evaluated at a.
Return type: tensor

lab.generic.divide[source]¶

Divide two tensors.

Parameters

a (tensor) – First tensor.
b (tensor) – Second tensor.

Returns

a divided by b.

Return type

tensor

lab.generic.erf[source]¶

Error function.

Parameters: a (tensor) – Tensor.
Returns: Error function evaluated at a.
Return type: tensor

lab.generic.exp[source]¶

Exponential function.

Parameters: a (tensor) – Tensor.
Returns: Exponential function evaluated at a.
Return type: tensor

lab.generic.eye[source]¶

Create an identity matrix.

Can also give a reference tensor whose data type and shape will be used to construct an identity matrix.

Parameters

dtype (dtype, optional) – Data type. Defaults to the default data type.
*shape (shape) – Shape of the matrix.

Returns

Identity matrix of shape shape and data type dtype.

Return type

tensor

lab.generic.ge[source]¶

Check whether one tensor is greater than or equal to another.

Parameters

a (tensor) – First tensor.
b (tensor) – Second tensor.

Returns

a is greater than or equal to b.

Return type

tensor[bool]

lab.generic.gt[source]¶

Check whether one tensor is strictly greater than another.

Parameters

a (tensor) – First tensor.
b (tensor) – Second tensor.

Returns

a is strictly greater than b.

Return type

tensor[bool]

lab.generic.identity[source]¶

Identity function

Parameters: a (tensor) – Tensor.
Returns: a exactly.
Return type: tensor

lab.generic.isnan[source]¶

Check whether a tensor is NaN.

Parameters: a (tensor) – Tensor.
Returns: a is NaN.
Return type: tensor[bool]

lab.generic.le[source]¶

Check whether one tensor is less than or equal to another.

Parameters

a (tensor) – First tensor.
b (tensor) – Second tensor.

Returns

a is less than or equal to b.

Return type

tensor[bool]

lab.generic.leaky_relu[source]¶

Leaky rectified linear unit.

Parameters

a (tensor) – Input.
alpha (tensor) – Coefficient of leak.

Returns

Activation value.

Return type

tensor

lab.generic.linspace[source]¶

Create a vector of c numbers ranging from a to c, distributed linearly.

Parameters

dtype (dtype, optional) – Data type. Defaults to the default data type.
a (number) – Lower bound.
b (number) – Upper bound.
num (int) – Number of numbers.

Returns

c numbers ranging from a to c, distributed linearly.

Return type

vector

lab.generic.log[source]¶

Logarithmic function

Parameters: a (tensor) – Tensor.
Returns: Logarithmic function evaluated at a.
Return type: tensor

lab.generic.log_2_pi = 1.8378770664093453¶: Value of log(2 * pi).

lab.generic.logsumexp[source]¶

Exponentiate a tensor, sum it, and then take the logarithm, possibly along an axis.

Parameters

a (tensor) – Tensor.
axis (int, optional) – Optional axis.

Returns

Reduced tensor.

Return type

tensor

lab.generic.lt[source]¶

Check whether one tensor is strictly less than another.

Parameters

a (tensor) – First tensor.
b (tensor) – Second tensor.

Returns

a is strictly less than b.

Return type

tensor[bool]

lab.generic.max[source]¶

Take the maximum of a tensor, possibly along an axis.

Parameters

a (tensor) – Tensor.
axis (int, optional) – Optional axis.

Returns

Reduced tensor.

Return type

tensor

lab.generic.maximum[source]¶

Take the maximum of two tensors.

Parameters

a (tensor) – First tensor.
b (tensor) – Second tensor.

Returns

Maximum of a and b.

Return type

tensor

lab.generic.mean[source]¶

Take the mean of a tensor, possibly along an axis.

Parameters

a (tensor) – Tensor.
axis (int, optional) – Optional axis.

Returns

Reduced tensor.

Return type

tensor

lab.generic.min[source]¶

Take the minimum of a tensor, possibly along an axis.

Parameters

a (tensor) – Tensor.
axis (int, optional) – Optional axis.

Returns

Reduced tensor.

Return type

tensor

lab.generic.minimum[source]¶

Take the minimum of two tensors.

Parameters

a (tensor) – First tensor.
b (tensor) – Second tensor.

Returns

Minimum of a and b.

Return type

tensor

lab.generic.multiply[source]¶

Multiply two tensors.

Parameters

a (tensor) – First tensor.
b (tensor) – Second tensor.

Returns

Product of a and b.

Return type

tensor

lab.generic.nan = nan¶: NaN.

lab.generic.negative[source]¶

Negate a tensor.

Parameters: a (tensor) – Tensor.
Returns: Negative of a.
Return type: tensor

lab.generic.one[source]¶

Create a 1 with a particular data type.

Parameters: dtype (dtype) – Data type.
Returns: 1 with data type dtype.
Return type: scalar

lab.generic.ones[source]¶

Create a tensor of ones.

Can also give a reference tensor whose data type and shape will be used to construct a tensor of ones.

Parameters

dtype (dtype, optional) – Data type. Defaults to the default data type.
*shape (shape) – Shape of the tensor.

Returns

Tensor of ones of shape shape and data type dtype.

Return type

tensor

lab.generic.pi = 3.141592653589793¶: Value of pi.

lab.generic.power[source]¶

Raise a tensor to a power.

Parameters

a (tensor) – Tensor.
power (tensor) – Power.

Returns

a to the power of power.

Return type

tensor

lab.generic.range[source]¶

Create a vector of numbers ranging from start to stop with step size step.

Parameters

dtype (dtype, optional) – Data type. Defaults to int.
start (number, optional) – Start of range. Defaults to 0.
stop (number) – End of range.
step (number, optional) – Step size. Defaults to 1.

Returns

Numbers ranging from start to stop with step size step.

Return type

vector

lab.generic.relu[source]¶

Rectified linear unit.

Parameters: a (tensor) – Tensor.
Returns: Rectified linear unit evaluated at a.
Return type: tensor

lab.generic.scan[source]¶

Perform a TensorFlow-style scanning operation.

Parameters

f (function) – Scanning function.
xs (tensor) – Tensor to scan over.
*init_state (tensor) – Initial state.

lab.generic.sigmoid[source]¶

Sigmoid function.

Parameters: a (tensor) – Tensor.
Returns: Sigmoid function evaluated at a.
Return type: tensor

lab.generic.sign[source]¶

Sign function.

Parameters: a (tensor) – Tensor.
Returns: Sign of a.
Return type: tensor

lab.generic.sin[source]¶

Sine function.

Parameters: a (tensor) – Tensor.
Returns: Sine function evaluated at a.
Return type: tensor

lab.generic.softplus[source]¶

Softplus function.

Parameters: a (tensor) – Tensor.
Returns: Softplus function evaluated at a.
Return type: tensor

lab.generic.sort[source]¶

Sort a tensor along an axis in ascending order.

Parameters

a (tensor) – Tensor to sort.
axis (int, optional) – Axis to sort along. Defaults to -1.
descending (bool, optional) – Sort in descending order. Defaults to False.

Returns

a, but sorted.

Return type

tensor

lab.generic.sqrt[source]¶

Square root.

Parameters: a (tensor) – Tensor.
Returns: Square root of a.
Return type: tensor

lab.generic.std[source]¶

Compute the standard deviation of a tensor, possibly along an axis.

Parameters

a (tensor) – Tensor.
axis (int, optional) – Optional axis.

Returns

Reduced tensor.

Return type

tensor

lab.generic.subtract[source]¶

Subtract two tensors.

Parameters

a (tensor) – First tensor.
b (tensor) – Second tensor.

Returns

a minus b.

Return type

tensor

lab.generic.sum[source]¶

Sum a tensor, possibly along an axis.

Parameters

a (tensor) – Tensor.
axis (int, optional) – Optional axis.

Returns

Reduced tensor.

Return type

tensor

lab.generic.tan[source]¶

Tangent function.

Parameters: a (tensor) – Tensor.
Returns: Tangent function evaluated at a.
Return type: tensor

lab.generic.tanh[source]¶

Tangent hyperbolic function.

Parameters: a (tensor) – Tensor.
Returns: Tangent hyperbolic function evaluated at a.
Return type: tensor

lab.generic.to_numpy[source]¶

Convert an object to NumPy.

Parameters: a (object) – Object to convert.
Returns: a as NumPy.
Return type: np.ndarray

lab.generic.zero[source]¶

Create a 0 with a particular data type.

Parameters: dtype (dtype) – Data type.
Returns: 0 with data type dtype.
Return type: scalar

lab.generic.zeros[source]¶

Create a tensor of zeros.

Can also give a reference tensor whose data type and shape will be used to construct a tensor of zeros.

Parameters

dtype (dtype, optional) – Data type. Defaults to the default data type.
*shape (shape) – Shape of the tensor.

Returns

Tensor of zeros of shape shape and data type dtype.

Return type

tensor