Closed liyunlongaaa closed 1 year ago
Hi there, due to copyright issues, the dataset collectors didn't release the raw audio/video, you will need to use some tool like youtube-dl to download data, which is a non-trivial process.
-Yuan
Thank you!
Is there a way to use tf-record to do this experiment, because the official website says "Frame-level features are stored as tensorflow.SequenceExample protocol buffers.", can this avoid downloading the original audio file?
@liyunlongaaa I have the whole audioset files in netdisk, do you still need now? (About 100GB, I don't remember)
@liyunlongaaa I have the whole audioset files in netdisk, do you still need now? (About 100GB, I don't remember)
oh, Thank you very much for your help, but I don't need it anymore. Thank you very much.
Hello @thelou1s, Sorry for commenting out old. Is there a direct link to download the full dataset?
@ASMIftekhar link: https://pan.baidu.com/s/1mD52damicvRCl5r9hbIRZA?pwd=m1jt password: m1jt
I only downloaded the tfrecord file on the audioset official website, and there is no audio file, or do I need to pay to download it?