OpenAI인증된 계정

Some of the pieces sounds nice. Is a code or paper release planned? Or should we lower expectations :)

Yep, code & pre-trained model is coming!

Thanks Greg. Congrats @mcleavey on the great release.

@OpenAI @gdb @ilyasut is it trained on audio wave data, or midi data? will a paper be released?

Trained on MIDI! From the post: "MuseNet was not explicitly programmed with our understanding of music, but instead discovered patterns of harmony, rhythm, and style by learning to predict the next token in hundreds of thousands of MIDI files"

Unsupervised learning is pretty amazing!!

But aren't you worried about FakeMuse

not really from a sort of malicious/abusive use perspective. I think generative models will lead to us being able to make enough good music that at some point there will be major work around IP of it. (Eg, what happens if a radio station wants to broadcast stuff?)

@robinhanson if people will genuinely enjoy these soundtracks, what happens to the argument of "we love art because we love to admire the people behind it" in signalling theory? Will they admire the developers or the executive team of OpenAI, the Internet itself - what else?

cc: @teslamanaf

Wuihhh

Gw otw ngoprek2 + ngoding nih bang wkwkwk

Sabtu dateng ya ;)