Recently, Tongyi Laboratory officially open-sourced its innovative music generation technology—InspireMusic. This technology aims to create an open-source AIGC toolkit that integrates music, song, and audio generation capabilities, providing a comprehensive creative platform for researchers, developers, and music enthusiasts.
InspireMusic not only offers researchers and developers a wealth of tools for training and fine-tuning music/song/audio generation models but also equips them with efficient models to optimize generation results. At the same time, this toolkit significantly lowers the barrier to music creation, allowing music enthusiasts to easily generate diverse musical works through simple text descriptions or audio prompts.
The text-to-music creation mode of InspireMusic is particularly noteworthy. It encompasses various musical styles, emotional expressions, and complex music structure controls, offering users great creative freedom and flexibility. Users can generate music works that meet their needs by inputting text descriptions based on their preferences, whether it's relaxing jazz or playful melodies, InspireMusic can present them all.
Additionally, InspireMusic features a flexible reasoning mode design, including fast models and high-quality models, to meet the diverse needs of users. Whether users seek quick generation or high-quality output, they can find a suitable creative approach within this toolkit.
Currently, InspireMusic has open-sourced the training and reasoning code for music generation, and users can access and use it through platforms such as GitHub repository, ModelScope, and HuggingFace Spaces.
In the future, Tongyi Laboratory plans to further open up the foundational models for singing generation and audio generation to attract more researchers, developers, and users to actively participate in experience and development. With everyone's joint efforts, InspireMusic will continue to improve and bring more surprises to the field of music creation.
GitHub Repository: InspireMusic (https://github.com/FunAudioLLM/InspireMusic)
Online Demo:
ModelScope: https://modelscope.cn/studios/iic/InspireMusic/summary
HuggingFace Spaces: https://huggingface.co/spaces/FunAudioLLM/InspireMusic
