Skip to content

Add comments for zero valued normalization factor in SoftmaxCrossEntropyLossGrad CUDA kernel. #3972

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
M. Zeeshan Siddiqui (codemzs) merged 1 commit into from
May 18, 2020
Merged

Add comments for zero valued normalization factor in SoftmaxCrossEntropyLossGrad CUDA kernel. #3972

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 6 additions & 0 deletions orttraining/orttraining/training_ops/cuda/loss/softmax_cross_entropy_loss_impl.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -118,6 +118,9 @@ __global__ void _WeightedSoftmaxCrossEntropyLossGrad(
int d = i % C;
CUDA_KERNEL_ASSERT(weight[row] == 0 || (label[row] >= 0 && label[row] < C));
if(0 == *normalize_factor){
// normalize_factor is sum of labels' weights. Because zero
// sum implies all weights are 0, the loss function should
// be constant 0 and its corresponding gradient should be 0 as well.
output_data[i] = 0;
} else {
output_data[i] = (*dY) * weight[row] * (_Exp(log_prob[i]) - 1.0 * (d == label[row])) / (*normalize_factor);
Expand All @@ -140,6 +143,9 @@ __global__ void _WeightedReductionNoneSoftmaxCrossEntropyLossGrad(
int d = i % C;
CUDA_KERNEL_ASSERT(weight[row] == 0 || (label[row] >= 0 && label[row] < C));
if(0 == *normalize_factor){
// normalize_factor is sum of labels' weights. Because zero
// sum implies all weights are 0, the loss function should
// be constant 0 and its corresponding gradient should be 0 as well.
output_data[i] = 0;
} else {
output_data[i] = dY[row] * weight[row] * (_Exp(log_prob[i]) - 1.0 * (d == label[row])) / (*normalize_factor);
Expand Down