I wanted to select audios only containing one type of instrument but the data in "autotagging_intrument.tsv" is wrong... For example taking a random audio (1269902) supposed to contain only trumpet contains a lot of other instruments.
You could retag them with another tool like https://github.com/qiuqiangkong/panns_inference
I wanted to select audios only containing one type of instrument but the data in "autotagging_intrument.tsv" is wrong... For example taking a random audio (1269902) supposed to contain only trumpet contains a lot of other instruments.
You could retag them with another tool like https://github.com/qiuqiangkong/panns_inference