predict using soundnet features

[jacobsn]

Idea: Use features extracted from the soundnet network in a traditional ML pipeline. Use lots of bagging and boosting like the typical winning solutions to Kaggle challenges.

[jacobsn]

@tsalem is working on this right now. He has the conv7 (I think) features extracted and is working on a scikit learn implementation. I am hoping that this will be an easier way for some, especially those that took my ML course, to get in the mix.

[wderekjones]

I can take a crack at this

[jacobsn]

Tawfiq might have started with this. It is probably worthwhile to coordinate with him.

…

On Fri, Dec 23, 2016, 12:50 PM Derek Jones ***@***.***> wrote: I can take a crack at this — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#32 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AA8G0K5beCDewTt0MVtMRPc79OwHxbk4ks5rLBfrgaJpZM4LQI4i> .

[jacobsn]

@williamdjones I just merged an initial version of this in fe28c61.

Work with @tsalem if you want access to the raw features (I am pretty sure I already uploaded them to google drive, but I suggest you chat with him).

[tsalem]

@jacobsn
@williamdjones already emailed me and I gave him access to the code.