Try to use avx_hs() - slow and requires alignment 4096 bits < (l.size*l.siz...

AlexeyAB

2018-08-08 a284a7da8d1facbf984a22302665a2a50295a687

 src/layer.h

@@ -179,6 +179,9 @@
    float *weights;
    float *weight_updates;

    char *align_bit_weights;
    float *mean_arr;

    float *col_image;
    int   * input_layers;
    int   * input_sizes;