Skip to content

Fix rounding in kv transcoding #902

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
copybara-service merged 1 commit into from
Apr 28, 2026
+2 −1
Merged

Fix rounding in kv transcoding #902

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
3 changes: 2 additions & 1 deletion gemma/kv_transcoding.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,6 +4,7 @@
#include <cstddef>
#include <cstdlib>
#include <optional>
#include <cmath>

#include "compression/types.h"
#include "gemma/activations.h"
Expand Down Expand Up @@ -52,7 +53,7 @@ inline size_t VOffset(bool transposed, size_t qkv_dim, size_t dim,
}

int8_t Quantize(float v, float inv_scale) {
float scaled = v * inv_scale;
float scaled = std::nearbyint(v * inv_scale);
if (scaled > 127.0f) return 127;
if (scaled < -127.0f) return -127;
return hwy::ConvertScalarTo<int8_t>(scaled);
Expand Down
Loading