academic torrent and availability of missing files

[benearnthof]

Hello and thank you for this extremely useful repository!

I'm currently downloading all of the raw files to compile a .torrent to make this dataset available at https://academictorrents.com/ and truly future proof this work and was wondering if the missing files (videos that have been deleted etc.) are available for download somewhere. I've read through #8 but am unable to use the chinese download provider there and also don't believe the raw files are hosted on there.

Could you provide raw files or link to a backup somewhere?

I'll compile the .torrent with raw and processed files in the coming week but my download is slow and may take another 4-5 days.

Please let me know if the missing files are available anywhere else apart from #8 Best regards.

[benearnthof]

I've created a complete academic torrent here:
https://academictorrents.com/details/843b5adb0358124d388c4e9836654c246b988ff4 and will be seeding this for the next year. Please mention me should you be unable to download the torrent and I will reseed if needed. Upload is confirmed working by the public Academic Torrent mirror in Florida. My rate is limited to 60mbit/second and I will only seed at night starting 22pm Central European Time until 6am the following day.

[fenghe12]

Thanks! but i can't download from your onedrive link .Is there anything wrong with it?

[johndpope]

magnet:?xt=urn:btih:843b5adb0358124d388c4e9836654c246b988ff4&dn=CelebV-HQ&tr=https%3A%2F%2Facademictorrents.com%2Fannounce.php&tr=https%3A%2F%2Fipv6.academictorrents.com%2Fannounce.php&tr=udp%3A%2F%2Ftracker.opentrackr.org%3A1337%2Fannounce&tr=udp%3A%2F%2Ftracker.openbittorrent.com%3A80%2Fannounce try this @fenghe12 - downloaded fine the other day. fyi - I'm doing some ground work for EMO paper - (slow progress) https://github.com/johndpope/Emote-hack

[fenghe12]

ok thanks a lot！ By the way,i'm also studying emo,maybe I can learn from your repo

[aixiaodewugege]

Thanks for your link. But some videos don't have audio.... Anybody notice that?

[triton99]

Hi @johndpope ,
I downloaded it and found that some videos don't have audio (like _CyaF_EymBU_38_0). Can you recheck it?
Thank you.

[johndpope]

hi @triton99 - it takes me a few days - i'm away on holiday. in mean time
@francqz31 dug up celebvox2 -
johndpope/VASA-1-hack#5 (comment)

the way to fix no audio - is to create an errata gist - and list all the videos - i saw this done on another dataset.

[triton99]

Hi @johndpope ,

I found that it has 6896/35666 (=19.34%) videos that don't have audio. Did you mean we remove non-audio videos from training and validation? And could you show me which dataset or project that did this way?

Thank you very much for the information.

[johndpope]

mdeff/fma#70

@triton99 - throw list into a gist
. i susepect it maybe due to licensing of music. it maybe possible to redownload these files 6896.

[triton99]

@johndpope . Here is list of the non-audio videos. Can you show me how to redownload it with the audio? I don't understand how to throw list into a gist. Thank you for your help.

no_audio_files.txt

[johndpope]

Gist.github.com
Each file name is extension for YouTube

YouTube.com/v=filename

if you use yt-dlp you can just pass the code in

https://github.com/yt-dlp/yt-dlp