Refactoring all the code to simplify things

[fmassa]

This PR substitutes #5.

• Unified Fast-RCNN with the other frameworks (RCNN and SPP).
• Tested Fast-RCNN and RCNN, both works as in the original implementations. Should test SPP again to verify that everything works fine
• Adds a simple visualization function to plot the detections on top of the image

[fmassa]

Just a heads up on this branch.
I did several internal improvements/extensions to this branch over the past few weeks that I'll push here. I'll merge this in in about 1-2 weeks.

[erfannoury]

Hi,

What's the status of this PR? Was there a problem that you didn't merge in the main code? Are you getting correct results?
Also which script should one use to train a Fast-RCNN model based on a custom architecture?

[fmassa]

Hey @erfannoury ,
Note that this is a lua-torch implementation, and has nothing to do with my re-implementation of Faster R-CNN in PyTorch (which is not public yet).
Do you actually want to use lua-torch (which is not maintained anymore) or pytorch?

[erfannoury]

Hi @fmassa,

Yes, I understand. I'm looking for a working implementation of Fast R-CNN in either Torch or PyTorch, but unfortunately there aren't any other working implementations out there except Girshick's implementation in pycaffe.

[fmassa]

As far as I remember, the implementation of Fast R-CNN here was working.