Google opens MusicLM: after 280000 hours of audio data training, 5-minute songs can be generated based on the text.

CTOnews.com May 11, Google launched the MusicLM model in January this year, which can automatically generate music works according to the text requirements entered by users. In its blog post today, Google announced the opening of the MusicLM model to all users.

Users can access the MusicLM model through AI Test Kitchen applications on Android and iOS platforms.

According to the blog post, the MusicLM model can create two versions of music works for users to choose from. The generated lyrics can not only contain information about the contents of the repertoire and musical instruments, but also explain the emotional information about the creation of the lyrics.

Google is trying to avoid criticism from copyright holders, so MusicLM will not create music that contains the voice of a particular artist.

CTOnews.com reported in April that the MusicLM model can generate 5-minute songs based on text after 280000 hours of audio data training.