def logistic_loss(observed_ratings, predicted_ratings, sample_weights=None):
"""
Logistic loss for explicit data.
Parameters
----------
observed_ratings: tensor
Tensor containing observed ratings which
should be +1 or -1 for this loss function.
predicted_ratings: tensor
Tensor containing rating predictions.
sample_weights: tensor, optional
Tensor containing weights to scale the loss by.
Returns
-------
loss, float
The mean value of the loss function.
"""
assert_no_grad(observed_ratings)
# Convert target classes from (-1, 1) to (0, 1)
observed_ratings = torch.clamp(observed_ratings, 0, 1)
if sample_weights is not None:
loss = F.binary_cross_entropy_with_logits(predicted_ratings,
observed_ratings,
size_average=False)
return _weighted_loss(loss, sample_weights)
return F.binary_cross_entropy_with_logits(predicted_ratings,
observed_ratings,
size_average=True)
def logistic_loss(observed_ratings, predicted_ratings):
"""
Logistic loss for explicit data.
Parameters
----------
observed_ratings: tensor
Tensor containing observed ratings which
should be +1 or -1 for this loss function.
predicted_ratings: tensor
Tensor containing rating predictions.
Returns
-------
loss, float
The mean value of the loss function.
"""
assert_no_grad(observed_ratings)
# Convert target classes from (-1, 1) to (0, 1)
observed_ratings = torch.clamp(observed_ratings, 0, 1)
return F.binary_cross_entropy_with_logits(predicted_ratings,
observed_ratings,
size_average=True)
def poisson_loss(observed_ratings, predicted_ratings, sample_weights=None):
"""
Poisson loss.
Parameters
----------
observed_ratings: tensor
Tensor containing observed ratings.
predicted_ratings: tensor
Tensor containing rating predictions.
sample_weights: tensor, optional
Tensor containing weights to scale the loss by.
Returns
-------
loss, float
The mean value of the loss function.
"""
assert_no_grad(observed_ratings)
loss = predicted_ratings - observed_ratings * torch.log(predicted_ratings)
return _weighted_loss(loss, sample_weights)
def regression_loss(observed_ratings, predicted_ratings):
"""
Regression loss.
Parameters
----------
observed_ratings: tensor
Tensor containing observed ratings.
predicted_ratings: tensor
Tensor containing rating predictions.
Returns
-------
loss, float
The mean value of the loss function.
"""
assert_no_grad(observed_ratings)
return ((observed_ratings - predicted_ratings)**2).mean()
def poisson_loss(observed_ratings, predicted_ratings):
"""
Poisson loss.
Parameters
----------
observed_ratings: tensor
Tensor containing observed ratings.
negative_predictions: tensor
Tensor containing rating predictions.
Returns
-------
loss, float
The mean value of the loss function.
"""
assert_no_grad(observed_ratings)
return (predicted_ratings - observed_ratings * torch.log(predicted_ratings)).mean()
def regression_loss(observed_ratings, predicted_ratings):
"""
Regression loss.
Parameters
----------
observed_ratings: tensor
Tensor containing observed ratings.
negative_predictions: tensor
Tensor containing rating predictions.
Returns
-------
loss, float
The mean value of the loss function.
"""
assert_no_grad(observed_ratings)
return ((observed_ratings - predicted_ratings) ** 2).mean()
def poisson_loss(observed_ratings, predicted_ratings):
"""
Poisson loss.
Parameters
----------
observed_ratings: tensor
Tensor containing observed ratings.
negative_predictions: tensor
Tensor containing rating predictions.
Returns
-------
loss, float
The mean value of the loss function.
"""
assert_no_grad(observed_ratings)
return (predicted_ratings - observed_ratings * torch.log(predicted_ratings)).mean()
