Live data from Hacker News

Show HN: I trained an AI model on 120M+ songs from iTunes

maroofy.com

191–200 of 444 posts

Re: Show HN: I trained an AI model on 120M+ songs from iTunes

#191

How did you get the samples for the song? iTunes allows scraping? Your project is extremely motivational .. how long did it take you? What did you train on? I do DL for work and just play with things like cifar. This is so inspired.

Apple Music subscription is $12/month and it allows full downloads of songs.

Hmmm. I would have thought that Apple would detect and block you if you try to download their entire catalog without some kind of permission. If they're the same size as the files in my Apple Music it would take in the order of a petabyte to store (compression would obviously reduce that).

120M songs would take approximately 1000 years to listen to in normal time and is way beyond normal usage.

Re: Show HN: I trained an AI model on 120M+ songs from iTunes

#192

How did you get the samples for the song? iTunes allows scraping? Your project is extremely motivational .. how long did it take you? What did you train on? I do DL for work and just play with things like cifar. This is so inspired.

Apple Music subscription is $12/month and it allows full downloads of songs.

... but certainly not ALL songs? I would think they notice if you download more then one minute of audio per minute over a long time.

Re: Show HN: I trained an AI model on 120M+ songs from iTunes

#194
post #24

This is very interesting, but unfortunately I haven't had the greatest luck in finding new songs I would enjoy listening to. It absolutely finds similar sounding tracks, but it doesn't distinguish which part of the song made it enjoyable. There's no tempo consistency or genre consistency or even main instrument/vocal timbre consistency between recommendations. I think locking one or more of those dimensions would all…

I agree with everyone's criticisms that it seems to identify similar tempo and melodic riff, irrespective of genre. But to me this is a feature, not a bug. I could see this or something like it opening my eyes to music I would never possibly have found on my own. I really like it! Spotify on the other hand seems to want to send me to the same group of artists and tracks I've listened to before, following some Collatz…

> But to me this is a feature, not a bug. I could see this or something like it opening my eyes to music I would never possibly have found on my own.

What makes it different from a big "play me a random song" button then?

Re: Show HN: I trained an AI model on 120M+ songs from iTunes

#195
I think I either found some kind of bug or Apple Music is really weird.

The similar songs for Anaal Nathrakh - Endarkenment [0] have Annal Nathrakh with エンダーケンメント at the top. So a misspelling of the band, and apparently (according to google translate) a Japanese transliteration of the song title (Endākenmento, enderkenment). The song itself is literally the same one, others listed are their other songs, or more Japanese transliterations of their songs. The weird versions can be played, but the linked page does not exist on Apple Music [1]

[0]: https://maroofy.com/songs/1522388463

[1]: https://music.apple.com/album/%E3%82%A8%E3%83%B3%E3%83%80%E3...

Re: Show HN: I trained an AI model on 120M+ songs from iTunes

#196

I find it disheartening that the comments are loaded with people saying how bad it is. The interface is awesome, the ability to have search results pop up and offer suggestions is impressive. I thought the results for a Lana Del Ray song and a song by Carlos Santana were both interesting and suggested new things. We should do a better job of supporting creators here.

Nothing plays for me.

Re: Show HN: I trained an AI model on 120M+ songs from iTunes

#200

Great initiative, but if it's sampling only the few seconds previews from the songs, then I doubt it will be useful. Also, a language filter is essential for both training the model and listing the results. I searched for a few Hindi language songs that I know of, and it listed only English, Chinese and Spanish songs as related ones - which is good to explore other music, but at times I'd just want to be in my native…

100% agree with a language filter, working on one (along with a better, improved model!)
Post reply on HN