you lost the last bottle net layer (kernel 5, out 960)

[mttbx]

you lost the last bottle net layer (kernel 5, out 960)

[frotms]

here is the last bottle net layer (kernel 5, out 960) @mttbx
https://github.com/frotms/MobilenetV3-Tensorflow/blob/master/mobilenet_v3.py#L272

[mttbx]

there are two conv layer which is kernel 5x5 and out 960, you only have one.

[mttbx]

[frotms]

@mttbx I think there is just one convolution layer that out 960.

[mttbx]

yeah, my mistake