Hi,
I am developing a VRChat avatar tool called KoEmo, which integrates SenseVoice for real-time emotion recognition. The detected emotions (Happy, Sad, Angry, Neutral) are used to automatically drive avatar facial expressions.
I am currently using the original pre-trained model weights without any fine-tuning, and I am planning to release this tool commercially in the future.
Could you please clarify the following?
- Which license applies to the original SenseVoice model weights — MIT or the FunASR Model Open Source License Agreement?
- Is commercial use permitted for a tool that bundles or calls the original model weights?
- Is any attribution required (e.g., in documentation or an about screen)?
Thank you for your great work on SenseVoice!
Hi,
I am developing a VRChat avatar tool called KoEmo, which integrates SenseVoice for real-time emotion recognition. The detected emotions (Happy, Sad, Angry, Neutral) are used to automatically drive avatar facial expressions.
I am currently using the original pre-trained model weights without any fine-tuning, and I am planning to release this tool commercially in the future.
Could you please clarify the following?
Thank you for your great work on SenseVoice!