Model files are from the Tensorflow/models Github, where they are available under an Apache 2.0 License.
Note the model checkpoint, classifier model and PCA parameters are hosted on IBM Cloud Object Storage.
class_labels_indices.csv contains all the classes from Audioset Ontology and is licensed under Creative Commons Attribution-ShareAlike 4.0 International (CC BY-SA 4.0).
All the audio samples provided as test assets here come from freesound.org and are licensed under variants of Creative Commons as specified here. They have been converted to 16-bit format wherever necessary to align with the model input format.
- birds1.wav by InspectorJ | License: Attribution
- birds2.wav by jmiddlesworth | License: Creative Commons 0
- applause.wav by joedeshon | License: Attribution
- piano.wav by Lemoncreme | License: Attribution
- crowd_cheering.wav by SoundsExciting | License: Creative Commons 0
- jazz_guitar.wav by Sub-d | License: Creative Commons 0
- typing.wav by Rodzuz | License: Creative Commons 0
- thunder.wav by InspectorJ | License: Attribution
- rain.wav by InspectorJ | License: Attribution
- train.wav by markedit | License: Creative Commons 0
- arcade_music.wav by joshuaempyre | License: Attribution
- car_engine.wav by InspectorJ | License: Attribution
- gunshots.mp3 by watupgroupie | License: Sampling+
- fireworks.wav by InspectorJ | License: Attribution
- snoring.wav by mookie182 | License: Attribution
- cinematic.wav by Hoerspielwerkstatt_HEF | License: Attribution
- siren.wav by wolferCZ | License: Creative Commons 0
- guitar.wav by User391915396 | License: Creative Commons