Skip to content
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

Jump to bottom

r2.10 cherry-pick: af4a6a3c8b9 "Fix empty input crash for SparseFillEmptyRowsGrad." #58565

Merged
merged 1 commit into from Nov 14, 2022
Merged

r2.10 cherry-pick: af4a6a3c8b9 "Fix empty input crash for SparseFillEmptyRowsGrad." #58565

Show file tree
Hide file tree
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.
Jump to
Jump to file
Failed to load files.
Diff view
Diff view
43 changes: 26 additions & 17 deletions tensorflow/core/kernels/sparse_fill_empty_rows_op_gpu.cu.cc
View file
Open in desktop
Expand Up @@ -297,9 +297,12 @@ struct SparseFillEmptyRows<GPUDevice, T, Tindex> {
empty_row_indicator = empty_row_indicator_t.vec<bool>().data();
}

TF_RETURN_IF_ERROR(wrap_kernel_call(ComputeEmptyRowIndicatorKernel<Tindex>,
/*device=*/device, /*size=*/dense_rows,
elements_per_row, empty_row_indicator));
if (dense_rows > 0) {
TF_RETURN_IF_ERROR(
wrap_kernel_call(ComputeEmptyRowIndicatorKernel<Tindex>,
/*device=*/device, /*size=*/dense_rows,
elements_per_row, empty_row_indicator));
}

// For each row, the number of empty rows up to and including that row.
Tensor num_empty_rows_through_t;
Expand Down Expand Up @@ -405,14 +408,16 @@ struct SparseFillEmptyRows<GPUDevice, T, Tindex> {
done);
}

OP_REQUIRES_OK_ASYNC(
context,
wrap_kernel_call(ScatterNewElementsKernel<T, Tindex>,
/*device=*/device, /*size=*/dense_rows, rank,
default_value, num_empty_rows_through,
input_row_ends, empty_row_indicator, output_indices,
output_values),
done);
if (dense_rows > 0) {
OP_REQUIRES_OK_ASYNC(
context,
wrap_kernel_call(ScatterNewElementsKernel<T, Tindex>,
/*device=*/device, /*size=*/dense_rows, rank,
default_value, num_empty_rows_through,
input_row_ends, empty_row_indicator,
output_indices, output_values),
done);
}

done();
};
Expand Down Expand Up @@ -461,9 +466,11 @@ struct SparseFillEmptyRows<GPUDevice, T, Tindex> {
TF_RETURN_IF_ERROR(
context->allocate_temp(index_type, TensorShape({N}), &row_indices_t));
auto row_indices = row_indices_t.flat<Tindex>();
TF_RETURN_IF_ERROR(wrap_kernel_call(CopyRowIndicesKernel<Tindex>,
/*device=*/device, /*size=*/N, rank,
indices, row_indices));
if (N > 0) {
TF_RETURN_IF_ERROR(wrap_kernel_call(CopyRowIndicesKernel<Tindex>,
/*device=*/device, /*size=*/N, rank,
indices, row_indices));
}
// Allocate input_index_map.
TF_RETURN_IF_ERROR(context->allocate_temp(index_type, TensorShape({N}),
input_index_map_t));
Expand Down Expand Up @@ -528,9 +535,11 @@ struct SparseFillEmptyRowsGrad<GPUDevice, T, Tindex> {
auto visited = visited_t.vec<bool>();
visited.device(device) = visited.constant(false);

TF_RETURN_IF_ERROR(wrap_kernel_call(
GatherOriginalGradValuesKernel<T, Tindex>, /*device=*/device,
/*size=*/N, reverse_index_map, grad_values, d_values, visited));
if (N > 0) {
TF_RETURN_IF_ERROR(wrap_kernel_call(
GatherOriginalGradValuesKernel<T, Tindex>, /*device=*/device,
/*size=*/N, reverse_index_map, grad_values, d_values, visited));
}

// Now we mask out the visited values and sum the remaining ones (which
// correspond to the empty rows in the forward input) to compute
Expand Down
7 changes: 7 additions & 0 deletions tensorflow/python/kernel_tests/sparse_ops/sparse_ops_test.py
View file
Open in desktop
Expand Up @@ -514,6 +514,13 @@ def testFillNumber(self):
self.assertAllEqual(empty_row_indicator_out,
np.array([0, 0, 1, 0, 1]).astype(np.bool_))

def testSparseFillEmptyRowsGradEmpty(self):
with test_util.use_gpu():
grad, _ = self.evaluate(
sparse_ops.sparse_fill_empty_rows_grad(
reverse_index_map=[], grad_values=[]))
self.assertAllEqual(grad, [])

@test_util.run_deprecated_v1
def testFillFloat(self):
with self.session():
Expand Down