from numbers import Real
from typing import Optional
import numpy as np
import mygrad._utils.graph_tracking as _tracking
from mygrad.operation_base import Operation
from mygrad.tensor_base import Tensor, asarray
from mygrad.typing import ArrayLike
class MarginRanking(Operation):
def __call__(self, x1, x2, y, margin):
"""Computes the margin ranking loss between ``x1``
and ``x2``.
Parameters
----------
x1 : mygrad.Tensor, shape=(N,) or (N, D)
x2 : mygrad.Tensor, shape=(N,) or (N, D)
y : numpy.ndarray
margin : float
Returns
-------
numpy.ndarray, shape=()
"""
self.variables = (x1, x2)
x1 = x1.data
x2 = x2.data
self.y = y
M = margin - self.y * (x1 - x2)
not_thresh = M <= 0
loss = M
loss[not_thresh] = 0.0
if _tracking.TRACK_GRAPH:
self._grad = np.ones_like(M)
self._grad[not_thresh] = 0.0
self._grad /= M.size
return np.mean(loss)
def backward_var(self, grad, index, **kwargs):
sign = -self.y if index == 0 else self.y
return grad * (sign * self._grad)
[docs]def margin_ranking_loss(
x1: ArrayLike,
x2: ArrayLike,
y: ArrayLike,
margin: float,
*,
constant: Optional[bool] = None
) -> Tensor:
r"""Computes the margin average margin ranking loss.
Equivalent to::
>>> import mygrad as mg
>>> mg.mean(mg.maximum(0, margin - y * (x1 - x2)))
Parameters
----------
x1 : ArrayLike, shape=(N,) or (N, D)
A batch of scores or descriptors to compare against those in `x2`
x2 : ArrayLike, shape=(N,) or (N, D)
A batch of scores or descriptors to compare against those in `x1`
y : Union[int, ArrayLike], scalar or shape=(N,)
1 or -1. Specifies whether the margin is compared against `(x1 - x2)`
or `(x2 - x1)`, for each of the N comparisons.
margin : float
A non-negative value to be used as the margin for the loss.
constant : bool, optional(default=False)
If ``True``, the returned tensor is a constant (it
does not back-propagate a gradient)
Returns
-------
mygrad.Tensor, shape=()
The mean margin ranking loss.
"""
if not 0 < x1.ndim < 3:
raise ValueError("`x1` must have shape (N,) or (N, D)")
if not x1.shape == x2.shape:
raise ValueError("`x1` and `x2` must have the same shape")
if not np.issubdtype(x1.dtype, np.floating):
raise TypeError("`x1` must contain floats")
if not np.issubdtype(x2.dtype, np.floating):
raise TypeError("`x2` must contain floats")
if not isinstance(margin, Real) or margin < 0:
raise ValueError("`margin` must be a non-negative scalar")
y = asarray(y)
if y.size == 1:
y = np.array(y.item())
if not y.ndim == 0 and not (y.ndim == 1 and len(y) == len(x1)):
raise ValueError("`y` must be a scalar or shape-(N,) array of ones")
if y.ndim:
if x1.ndim == 2:
y = y.reshape(-1, 1)
return Tensor._op(MarginRanking, x1, x2, op_args=(y, margin), constant=constant)