Add comment for MlasConvDepthwise new parameter

microsoft · Jul 27, 2024 · 7725154 · 7725154
1 parent 3df8960
commit 7725154
Showing 1 changed file with 1 addition and 0 deletions.
diff --git a/onnxruntime/core/mlas/inc/mlas.h b/onnxruntime/core/mlas/inc/mlas.h
@@ -1751,6 +1751,7 @@ MlasSBGemmConvertPackB(size_t N, size_t K, const float* B, size_t ldb, void* Pac
  * @brief Indirect Depthwise convolution for fp16
  * @param Input         Supplies the indirect buffer for NHWC input
  * @param Filter        Supplies the address for filter tensor
+ * @param Bias          Supplies the address for 1D bias tensor B, has size of M
  * @param Output        Supplies the address for the result tensor
  * @param Channels      # of input channels
  * @param OutputCount   # of output pixels