# Solution ready on the github

This is for you while waiting for 2nd stage result :)
[https://github.com/daisukelab/freesound-audio-tagging-2019](https://github.com/daisukelab/freesound-audio-tagging-2019)
![title_picture](assets/d34c3d681572a7e20e0adc89.png)
Thank you for everybody organized this competition, and congratulations to the winners.
It was wonderful opportunity once again to dig deeper into audio data for scene understanding application, I could gain intuition and understanding for real world sound-scene-aware applications which I'm hoping to have in electric appliances around us in the future.
I'd share what I have done and how was that during this competitions other than public kernels.
Topics Summary:
- Audio preprocessing to create mel-spectrogram based features, 3 formats were tried.
- Audio preprocessing also tried background subtraction; focusing on sound events... not sure it worked or not.
- Tried to mitigate curated/noisy audio differences; inter-domain transfer regarding
frequency envelope
.
- Soft relabeling, by using co-occurrence probability of labels.
- Ensemble for multi-label problem; per-class per-model 2D weighting for better averaging.