The parameters problem(num of priorbox)

[wjxiz1992]

MobileNet-SSD/gen.py

Line 748 in daef68a

self.mbox("conv11", 3)

Hi, chuanqi305, why the parameter here is 3 ,not 4? Because according to weiliu89's source code in "prior_box_layer.cpp", if aspect ratio = [2.0], there should be 4 default boxes, their width are:

1. aspect = 1.0 width = min_size
   2.aspect = 1.0 width = sqrt(min_size*max_size)
2. aspect = 2.0 width = min_size * sqrt(2.0)
3. aspect = 1/2.0 width = min_size * sqrt(1/2.0)

Thanks for your explanation here.

[chuanqi305]

I followed tensorflow's config, 3 anchors for conv11, without
2.aspect = 1.0 width = sqrt(min_size*max_size)

[wjxiz1992]

Hi chuanqi305, I am a deep learning engineer in NVIDIA, could you please share your WeChat number and discuss more? 2017-08-28 8:55 GMT+08:00 chuanqi305 <notifications@github.com>:

…

I followed tensorflow's config, 3 anchors for conv11, without 2.aspect = 1.0 width = sqrt(min_size*max_size) — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#23 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ATh0GmYkHyGJnXXqIVqZCWDIBdRckNzGks5scg_kgaJpZM4O_lnP> .

[Aspirinkb]

Hi @chuanqi305 where can I find tensorflow's mobilenet(v2) ssd config? I downloaded its source code (tensorflow-models) but did not find the config details about conv11 layer's anchors.