[Deepseek v3.2] Optimize top_k_per_row (#26763)

Signed-off-by: Daniel Campora <961215+dcampora@users.noreply.github.com>
This commit is contained in:
Daniel Cámpora
2025-10-21 08:30:07 +00:00
committed by GitHub
parent c3a2c6ac5f
commit 80e9452984
5 changed files with 13 additions and 49 deletions
+1 -2
View File
@@ -99,8 +99,7 @@ void apply_repetition_penalties_(torch::Tensor& logits,
void top_k_per_row(const torch::Tensor& logits, const torch::Tensor& rowStarts,
const torch::Tensor& rowEnds, torch::Tensor& indices,
torch::Tensor& values, int64_t numRows, int64_t stride0,
int64_t stride1);
int64_t numRows, int64_t stride0, int64_t stride1);
void rms_norm_static_fp8_quant(torch::Tensor& out, torch::Tensor& input,
torch::Tensor& weight, torch::Tensor& scale,