from torch import optim, nn
from ..models.model import Model
from ..utils import tolist
from ..losses import NLL
[docs]class ML(Model):
"""
Maximum Likelihood (log-likelihood)
"""
def __init__(self, p,
other_distributions=[],
optimizer=optim.Adam,
optimizer_params={}):
# set distributions (for training)
distributions = [p] + tolist(other_distributions)
# set losses
self.nll = NLL(p)
loss = self.nll.mean()
super().__init__(loss, test_loss=loss,
distributions=distributions,
optimizer=optimizer, optimizer_params=optimizer_params)
[docs] def train(self, train_x={}, **kwargs):
return super().train(train_x, **kwargs)
[docs] def test(self, test_x={}, **kwargs):
return super().test(test_x, **kwargs)