Feature Request: 8-Speaker (8 spk) version of `diar_streaming_sortformer`
Hello NVIDIA NeMo Team,
Thank you for your excellent work on the diar_streaming_sortformer_4spk-v2 model. The performance and streaming capabilities for 4-speaker diarization are very impressive.
We are writing to express strong interest in, and inquire about the possibility of, an 8-speaker (8 spk) version of this model.
In many of our practical use cases, such as large business meetings, panel discussions, or multi-participant calls, we frequently encounter scenarios that exceed the 4-speaker limit.
A model capable of handling up to 8 speakers would be incredibly useful and significantly expand the model's applicability to these more complex, real-world scenarios.
Could you share if an 8-speaker version is on your roadmap? Any information on its potential availability would be greatly appreciated by the community.
Thank you for your consideration!
Best,
Hi @FIT17 ,
Thank you for your kind words about the 4-speaker model and for the great feedback. We really appreciate hearing about your use cases.
To answer your question: yes, an 8-speaker version is on our roadmap.
We are currently working on it, and the release is targeted for the first half of 2026.
We value your input, and please stay tuned for future announcements.