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

AlexeyAB

2018-08-08 a284a7da8d1facbf984a22302665a2a50295a687

 cfg/vgg-16.cfg

@@ -1,12 +1,11 @@
[net]
batch=128
subdivisions=8
subdivisions=4
height=256
width=256
channels=3
learning_rate=0.00001
momentum=0.9
seen=0
decay=0.0005

[crop]