Skip to content

Revert 2:3 read optimization for 1x1 convolution operator #159

Revert 2:3 read optimization for 1x1 convolution operator

Revert 2:3 read optimization for 1x1 convolution operator #159

Triggered via pull request October 31, 2023 13:53
Status Success
Total duration 1m 5s
Artifacts 2

build.yaml

on: pull_request
Generate pack
55s
Generate pack
Fit to window
Zoom out
Zoom in

Annotations

2 warnings
Generate pack: Source/SVDFunctions/arm_svdf_s8.c#L50
argument 'ctx' from the argument list of arm_svdf_s8 has multiple @param documentation sections
Generate pack: Source/NNSupportFunctions/arm_nn_vec_mat_mult_t_s8.c#L58
The following parameter of arm_nn_vec_mat_mult_t_s8(const int8_t *lhs, const int8_t *rhs, const int32_t *kernel_sum, const int32_t *bias, int8_t *dst, const int32_t lhs_offset, const int32_t dst_offset, const int32_t dst_multiplier, const int32_t dst_shift, const int32_t rhs_cols, const int32_t rhs_rows, const int32_t activation_min, const int32_t activation_max, const int32_t address_offset) is not documented:

Artifacts

Produced during runtime
Name Size
artifact Expired
968 KB
documentation Expired
2.73 MB