Joakim Rishaug

I believe the biggest problem for having a flexible PriorBox implementation is mostly related to [this line in the python implementation](https://github.com/rykov8/ssd_keras/blob/master/ssd_layers.py#L153) which should really be [like this](https://github.com/weiliu89/caffe/blob/ssd/src/caffe/layers/prior_box_layer.cpp#L39) including the lines...

@rykov8 yes, it works perfectly in the implementation you have now. Thanks for making it :) Based on @oarriaga 's comment and what I've read so far, it seems like...

The tests shouldn't fail from what I changed here, so I suppose this is related to branch-builds or something else?

That's great @ataymano, thanks for the prompt response :slightly_smiling_face: If you have some time though, can you explain quickly how the example I provided ends up with empty decision elements?...

Hi @ataymano , I was wondering if there's any progress on this? It seemed to be going so smoothly with the PR. Just pinging in, in case it fell by...

Could I trouble you guys to do a release of the library as well? We prefer not to build from source unless absolutely necessary.

@lalo , Python and Java bindings mainly, why?

+1 Not having this makes this otherwise perfect integration almost unusable in our case. It would be great to hear an update on when we could expect someone to start...

@azuredream I don't mean for this to sound as nagging/pestering, but just for adjusting my expectations: How long do you think it'll take you to implement? I was also hacking...

@azuredream I'm sorry if I offended, and made you think I was _**expecting**_ things to get solved very quickly. That was not my intention. The end of the week would...