Tiangong SkyMusic music model launch invitation test

Tiangong SkyMusic music model

DIYuan | 2024-04-02 10:38

【數(shù)據(jù)猿導(dǎo)讀】 Tiangong SkyMusic music model launch invitation test

On April 2, Kunlun Wanwei AI music generation model "Tiangong SkyMusic" opened a free invitation test for the society from now on. This round of invitation will be open to 1,000 free places for industry media, experts, and interested music practitioners. According to reports, "Tiangong SkyMusic" is based on the Kunlun "Tiangong 3.0" super large model, which is the only publicly available AI music generation large model in China. "Tiangong SkyMusic" adopts Sora model architecture in the field of Music and audio, and Large-scale Transformer is responsible for composing music to learn the context dependence of Music Patches and achieve music controllability at the same time. Diffusion Transformer is responsible for singing, and the Music Patches can be restored to high quality audio through LDM, so that "SkyMusic" can support the generation of 80 seconds of 44100Hz sample rate two-channel stereo songs. This model architecture works well in the areas of video, audio and music. The Kunlun team will also gradually iterate and add new capabilities in the future, so that the model has multi-modal emotional understanding and expression capabilities.

來(lái)源：DIYuan

收藏分享

聲明：數(shù)據(jù)猿尊重媒體行業(yè)規(guī)范，相關(guān)內(nèi)容都會(huì)注明來(lái)源與作者；轉(zhuǎn)載我們?cè)瓌?chuàng)內(nèi)容時(shí)，也請(qǐng)務(wù)必注明“來(lái)源：數(shù)據(jù)猿”與作者名稱，否則將會(huì)受到數(shù)據(jù)猿追責(zé)。