Skip to content

Support Half/BFloat16 in native_layer_norm #7752

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
merged 1 commit into from
Jan 22, 2025
Merged

Support Half/BFloat16 in native_layer_norm #7752

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.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion kernels/optimized/cpu/op_native_layer_norm.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -155,7 +155,7 @@ std::tuple<Tensor&, Tensor&, Tensor&> opt_native_layer_norm_out(
InvalidArgument,
ret_val);

ET_SWITCH_FLOAT_TYPES(
ET_SWITCH_FLOATHBF16_TYPES(
input.scalar_type(), ctx, "native_layer_norm.out", CTYPE, [&]() {
layer_norm<CTYPE>(
input,
Expand Down
11 changes: 6 additions & 5 deletions kernels/portable/cpu/op_native_layer_norm.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -66,15 +66,16 @@ void layer_norm(
bias_data = nullptr;
}

const CTYPE ct_normalized = static_cast<CTYPE>(normalized);
for (int i = 0; i < leading; ++i) {
const CTYPE* x = input_data + i * normalized;
CTYPE* y = out_data + i * normalized;

// compute E[X] and Var[x] = E[x^2] - E[x]^2
CTYPE sum = reduce_add(x, normalized);
CTYPE sq_sum = vec_powerf(x, normalized);
CTYPE mean_value = sum / normalized;
CTYPE variance = sq_sum / normalized - mean_value * mean_value;
CTYPE sum = reduce_add(x, ct_normalized);
CTYPE sq_sum = vec_powerf(x, ct_normalized);
CTYPE mean_value = sum / ct_normalized;
CTYPE variance = sq_sum / ct_normalized - mean_value * mean_value;
CTYPE std = std::sqrt(variance + eps);

// Calculate the elements of output
Expand Down Expand Up @@ -167,7 +168,7 @@ std::tuple<Tensor&, Tensor&, Tensor&> native_layer_norm_out(
InvalidArgument,
ret_val);

ET_SWITCH_FLOAT_TYPES(
ET_SWITCH_FLOATHBF16_TYPES(
input.scalar_type(), ctx, "native_layer_norm.out", CTYPE, [&]() {
layer_norm<CTYPE>(
input,
Expand Down
12 changes: 11 additions & 1 deletion kernels/test/op_native_layer_norm_test.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -95,7 +95,15 @@ class OpNativeLayerNormTest : public OperatorTest {
EXPECT_TENSOR_CLOSE(out0, std::get<0>(result));

Tensor expected = tf.make(test_case.sizes, test_case.expected_data);
EXPECT_TENSOR_CLOSE(out0, expected);
if constexpr (DTYPE == ScalarType::BFloat16) {
EXPECT_TENSOR_CLOSE_WITH_TOL(
out0,
expected,
1e-2,
executorch::runtime::testing::internal::kDefaultBFloat16Atol);
} else {
EXPECT_TENSOR_CLOSE(out0, expected);
}
}
}

Expand Down Expand Up @@ -393,6 +401,8 @@ std::vector<int64_t> vector_32_to_64(std::vector<int32_t> vector_32) {
TEST_F(OpNativeLayerNormTest, FloatTensors) {
run_floating_point_test_cases<ScalarType::Float>();
run_floating_point_test_cases<ScalarType::Double>();
run_floating_point_test_cases<ScalarType::Half>();
run_floating_point_test_cases<ScalarType::BFloat16>();
}

TEST_F(OpNativeLayerNormTest, IntTensorsDies) {
Expand Down
Loading