Function arm_nn_vec_mat_mult_t_fp16

Function Documentation

arm_cmsis_nn_status arm_nn_vec_mat_mult_t_fp16(const float16_t *lhs, const float16_t *rhs, const float16_t *bias, float16_t *dst, const int32_t rhs_cols, const int32_t rhs_rows, const float16_t activation_min, const float16_t activation_max)

fp16 elementwise multiplication with fp16 output

Supported framework: TensorFlow Lite micro

Parameters:

  • lhs – [in] pointer to input vector 1

  • rhs – [in] pointer to input vector 2

  • bias – [in] pointer to bias

  • dst – [inout] pointer to output vector

  • rhs_cols – [in] number of samples per batch

  • rhs_rows – [in] number of samples per batch

  • activation_min – [in] minimum value to clamp output to

  • activation_max – [in] maximum value to clamp output to

Returns:

The function returns ARM_CMSIS_NN_SUCCESS