Underdog 在浏览器游戏中演示其 32 MB 的 Woof 语音模型
Underdog shows off a 32MB speech model in new demo
- 谁
- Underdog
- 引文
One such startup, Underdog, has released a game built with its new voice model called Woof that’s only 32MB large.
发生了什么
Underdog 发布了一款基于 Woof 构建的浏览器游戏。Woof 是其 32 MB 的语音生成模型。玩家可以输入文本，让游戏朗读出来，还可以旋转调音器来生成音乐；Underdog 表示，该模型拥有 3100 万个参数，可在大多数设备上离线运行。这款演示版推出前几周，该公司刚发布了 Bark，这是一款 167 MB 的语音转文本模型。
接下来读什么
查看这件事的完整进展Underdog Releases a Browser Game Demo Powered by Its Woof Speech Model这件事到目前为止
Underdog released a browser game demo powered by Woof, its 32 MB speech generation model. Players can type text for the game to speak aloud or turn a tuner to generate music. Underdog says the 31-million-parameter model can run on most devices without an internet connection. The demo follows Underdog’s launch of Bark, a 167 MB speech-to-text model, weeks earlier.
每天一封，跟上 AI 的重要变化
模型发布、价格变化、算力进展，附原始来源。免费，确认后开始发送，随时一键退订。 看一份往期日报 · 2026-11-06