support to the mobilnetv2-yolo structure of r1.13

[hejinIzhanghao]

add the mobilenetv2_yolo_body_r13, which makes the r1.14 supoort the mobilenetv2-yolo structure of r1.13

[hejinIzhanghao]

@fsx950223 i did not find squash button for this repo.

[hejinIzhanghao]

I make r1.14 support the model structure of r.1.13 instead of using r1.13, because r1.13 does NOT work well in my multi-gpu machine.

[fsx950223]

What is the value of the mAP on VOC dataset?

[hejinIzhanghao]

I did not try on voc.
it works better than the one of r1.14 on my own dataset.

[fsx950223]

You should try it and show me the mAP

[hejinIzhanghao]

OK. Once my machines have spare time.
And the model is exactly same as the one in r1.13. And the weights file trained by r1.13 can be loaded.

[fsx950223]

Ok, thanks hejinIzhanghao <[email protected]<mailto:[email protected]>> 于2019年9月5日周四 下午4:12写道： OK. Once my machines have spare time. And the model is exactly same as the one in r1.13. And the weights file trained by r1.13 can be loaded. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub<#55?email_source=notifications&email_token=AEGHB44SOC4HDATP3SP4L33QIC5NPA5CNFSM4ITZSMNKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD56H3LI#issuecomment-528252333>, or mute the thread<https://github.com/notifications/unsubscribe-auth/AEGHB4ZI2HUB4WBMNU536PLQIC5NPANCNFSM4ITZSMNA>.

[hejinIzhanghao]

@fsx950223
BTW:
backbone of mobilenetv2 + loss of GIOU/MSE worked on my dataset.
backbone of darknet53 + loss of MSE worked, but the loss of GIOU did not work with all loss of nan.

I did not find the reaseon yet.