Skip to content

Fix HiFi relu for int8 #8423

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

Open
wants to merge 1 commit into
base: main
Choose a base branch
from
Open

Fix HiFi relu for int8 #8423

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
12 changes: 8 additions & 4 deletions backends/cadence/hifi/operators/op_quantized_relu_out.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -26,11 +26,11 @@ void quantized_relu_per_tensor_out(
const int64_t out_multiplier,
const int64_t out_shift,
Tensor& output) {
const uint8_t _in_zero_point = static_cast<uint8_t>(in_zero_point);
const uint8_t _out_zero_point = static_cast<uint8_t>(out_zero_point);
const int32_t _out_multiplier = static_cast<int32_t>(out_multiplier);
Copy link
Contributor

@zonglinpeng zonglinpeng Feb 12, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

do these two need to move?

Copy link
Contributor Author

@mcremon-meta mcremon-meta Feb 12, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

they don't you're right

const int32_t _out_shift = static_cast<int32_t>(out_shift);
if (input.scalar_type() == executorch::aten::ScalarType::Byte) {
const uint8_t _in_zero_point = static_cast<uint8_t>(in_zero_point);
const uint8_t _out_zero_point = static_cast<uint8_t>(out_zero_point);
const int32_t _out_multiplier = static_cast<int32_t>(out_multiplier);
const int32_t _out_shift = static_cast<int32_t>(out_shift);
const uint8_t* p_in = input.const_data_ptr<uint8_t>();
uint8_t* p_out = output.mutable_data_ptr<uint8_t>();

Expand All @@ -48,6 +48,10 @@ void quantized_relu_per_tensor_out(
ET_CHECK_MSG(ret_val == 0, "An internal error occured");

} else if (input.scalar_type() == executorch::aten::ScalarType::Char) {
const int8_t _in_zero_point = static_cast<int8_t>(in_zero_point);
const int8_t _out_zero_point = static_cast<int8_t>(out_zero_point);
const int32_t _out_multiplier = static_cast<int32_t>(out_multiplier);
const int32_t _out_shift = static_cast<int32_t>(out_shift);
const int8_t* p_in = input.const_data_ptr<int8_t>();
int8_t* p_out = output.mutable_data_ptr<int8_t>();

Expand Down
Loading