Today, we’re excited to share Orpheus, an open-source TTS model that exceeds the capabilities of existing open and closed source speech generation models! Blog post: https://lnkd.in/grJyAgTE Orpheus is capable of expressing empathy, consistent with the emotional intelligence of a human, being able to produce non-textual cues like sighing, laughing, chuckling, etc, which was not previously accessible. It has a great understanding of disfluencies (like ums) and other artefacts that make speech actually feel human. Under the hood, Orpheus is built on top of an Llama-3.2, so it has a much better understanding of how humans think and express themselves than existing approaches. We’ve shared more details about this release in our blog post!
关于我们
Virtual humans indistinguishable from real ones
- 网站
-
https://canopylabs.ai
Canopy Labs的外部链接
- 所属行业
- 研究服务
- 规模
- 11-50 人
- 总部
- San Francisco
- 类型
- 私人持股
地点
-
主要
获取路线
US，San Francisco