Fast model training with small dataset
Training set requires only 50+ audio files assigned to each label. Training time can be about an hour if training set contains 1,000 audio files.
Security monitoring：custom classification of abnormal and normal sounds, as the basis of emergency alert
Scientific research：custom recognition of sounds of individuals from the same or different species, to assist field research
Others：just brainstorm and train the sound classification model you desire