2

I will be working on an implementation of DQNs and recent extensions from Google DeepMind in Caffe.

For this I will write a simulator (in place of the Atari emulator) for creating training experiences for the agent.

My question is: Which of the Matlab or Python interfaces to Caffe is the most mature and well functioning?
Any other considerations for choosing between Python and Matlab for this task?

It should be noted that I'm much more experienced with Python development than with Matlab development (apart from using it as a tool for math classes etc.), however a team I'm collaborating with seems to mostly use Matlab.

Shai
  • 111,146
  • 38
  • 238
  • 371
Skeppet
  • 931
  • 1
  • 9
  • 17
  • 1
    Interesting question (I would really appreciate an answer), but I fear any answer would be primarily opinion-based, which would make the question rather off-topic for SO. – hbaderts Oct 12 '15 at 14:19
  • 2
    by all means, use python. – Shai Oct 13 '15 at 09:07
  • @Shai any reasons apart from personal preference? I might face that decision in the near future and will have to decide about switching from MATLAB to Python. – hbaderts Oct 13 '15 at 10:10
  • 1
    @hbaderts AFAIK caffe is developing a more extensive python interface: you can create layers using python code, you can build networks from python (without using prototxt...) I don't see this kind of progress on the matlab side of things. – Shai Oct 13 '15 at 10:41
  • @Shai Thank you for this input, I wasn't aware of that. – hbaderts Oct 13 '15 at 11:26
  • 1
    @Skeppet you can see an example [in this question](http://stackoverflow.com/q/33140000/1714410) how to define a net in python without using prototxt! – Shai Oct 15 '15 at 07:43

1 Answers1

2

IMHO, it is better to use python to interface with caffe.
From my understanding caffe developers and community intend on developing the python interface more than the matlab one. One such direction is allowing for a net to be constructed purely in python (not using prototxt file at all), you can see such an example in this question.

Community
  • 1
  • 1
Shai
  • 111,146
  • 38
  • 238
  • 371