Google audio set

[punitagrawal32]

Hello @antoine77340 ,

I wanted to know if it would be viable to run your models on the google audio set data.

The data is stored as tfrecords. The number of classes in the data is 527. What changes do you think I should make in the script/ architecture to get your models running?

Thanks in advance.

[orprager]

@punitagrawal32
I can tell that I made a test ,In witch I used audioset for audio feature extraction and google frame feature extractor.
I tested the accuracy of the willow model I trained (I trained the model on the youtube 8m tfrecords) on the self-extracted tfrecords features (audioset + frame feature extractor) and got pretty good results.

[estathop]

did you choose little-endian or big-endian for the audioset feature extraction ? and why ?

[orprager]

I used audioset embedding op (tensor) with the default properties

…

On Tue, Nov 6, 2018, 2:20 PM estathop ***@***.*** wrote: did you choose little-endian or big-endian for the audioset feature extraction ? and why ? — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#28 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AjLcwZ37OnPq0_ba5NefmFcBpEYGQstBks5usX6dgaJpZM4W6llV> .