Improvements to computing losses (3b2c17d9) · Commits · 钟慕尧 / deepchem

deepchem/models/losses.py

+4 −4

Original line number	Diff line number	Diff line
		@@ -138,10 +138,10 @@ class SigmoidCrossEntropy(Loss):

		def _create_pytorch_loss(self):
		import torch
		bce = torch.nn.BCELoss(reduction='none')
		bce = torch.nn.BCEWithLogitsLoss(reduction='none')

		def loss(output, labels):
		return bce(torch.sigmoid(output), labels)
		return bce(output, labels)

		return loss

		@@ -163,10 +163,10 @@ class SoftmaxCrossEntropy(Loss):

		def _create_pytorch_loss(self):
		import torch
		ls = torch.nn.LogSoftmax(dim=1)

		def loss(output, labels):
		return -torch.sum(
		labels * torch.log(torch.nn.functional.softmax(output, 1)), dim=-1)
		return -torch.sum(labels * ls(output), dim=-1)

		return loss