vggvox-speaker-identification icon indicating copy to clipboard operation
vggvox-speaker-identification copied to clipboard

about the conv_bn_dynamic_apool

Open hktxt opened this issue 5 years ago • 4 comments

I read your code and found that the 9*1 is a conv layer in conv_bn_dynamic_apool() function. The paper says "replaced by two -layers-a fully connected layers of 9*1 and an average layer with 1/*8..." I stuck on this for a long time. Maybe you are right, that is a conv layer, which make sense.

hktxt avatar Nov 09 '18 03:11 hktxt

another question is why K.l2_normalize ?

hktxt avatar Nov 09 '18 07:11 hktxt

The wavreader function produce different result against with matlab.

hktxt avatar Nov 12 '18 06:11 hktxt

FileNotFoundError: File b'cfg/enroll_list.csv' does not exist ? can you help me ?

zhengqun avatar Nov 15 '18 08:11 zhengqun

Pretty sure I got the layer structure by following the Matlab model. Will check/update when I got more time.

linhdvu14 avatar Nov 15 '18 16:11 linhdvu14