| | |
| | | #define GEMM_H |
| | | #include "activations.h" |
| | | #include <stdint.h> |
| | | #include <stddef.h> |
| | | |
| | | void convolution_2d(int w, int h, int ksize, int n, int c, int pad, int stride, |
| | | float *weights, float *input, float *output, float *mean); |
| | | |
| | | static inline void set_bit(unsigned char *const dst, size_t index) { |
| | | size_t dst_i = index / 8; |
| | | int dst_shift = index % 8; |
| | | dst[dst_i] |= 1 << dst_shift; |
| | | //dst[dst_i] |= 1 << (8 - dst_shift); |
| | | } |
| | | |
| | | static inline unsigned char get_bit(unsigned char const*const src, size_t index) { |
| | | size_t src_i = index / 8; |
| | | int src_shift = index % 8; |
| | | unsigned char val = (src[src_i] & (1 << src_shift)) > 0; |
| | | //unsigned char val = (src[src_i] & (1 << (8 - src_shift))) > 0; |
| | | return val; |
| | | } |
| | | |
| | |
| | | void transpose_block_SSE4x4(float *A, float *B, const int n, const int m, |
| | | const int lda, const int ldb, const int block_size); |
| | | |
| | | void transpose_bin(char *A, char *B, const int n, const int m, |
| | | const int lda, const int ldb, const int block_size); |
| | | |
| | | void gemm_nn_custom_bin_mean_transposed(int M, int N, int K, float ALPHA_UNUSED, |
| | | unsigned char *A, int lda, |
| | | unsigned char *B, int ldb, |
| | |
| | | int channels, int height, int width, |
| | | int ksize, int stride, int pad, float* data_col); |
| | | |
| | | void im2col_cpu_custom_bin(float* data_im, |
| | | int channels, int height, int width, |
| | | int ksize, int stride, int pad, float* data_col, int bit_align); |
| | | |
| | | void im2col_cpu_custom_transpose(float* data_im, |
| | | int channels, int height, int width, |
| | | int ksize, int stride, int pad, float* data_col, int ldb_align); |
| | | |
| | | void activate_array_cpu_custom(float *x, const int n, const ACTIVATION a); |
| | | |
| | | |
| | |
| | | float *B, int ldb, |
| | | float *C, int ldc); |
| | | |
| | | |
| | | void forward_maxpool_layer_avx(float *src, float *dst, int *indexes, int size, int w, int h, int out_w, int out_h, int c, |
| | | int pad, int stride, int batch); |
| | | |
| | | |
| | | void gemm(int TA, int TB, int M, int N, int K, float ALPHA, |
| | | float *A, int lda, |
| | | float *B, int ldb, |