Closed kimihailv closed 2 years ago
Hello. Could you please specify the size of AudioCaps test set which you used in your evaluation? test.csv file in audiocaps repo contains 4876 rows, but there is also the table with splits' sizes and according to it test set contains 975 samples.
Hi, in audiocaps, there are 5 captions for each audio. So 975*5=4875.
Thank you
Hello. Could you please specify the size of AudioCaps test set which you used in your evaluation? test.csv file in audiocaps repo contains 4876 rows, but there is also the table with splits' sizes and according to it test set contains 975 samples.