返回全部动态
Maya Research 发布开源 TTS 模型 Maya-1
原标题:maya-research/maya1
AI 摘要
Maya Research 在 Hugging Face 上发布了开源权重文本转语音模型 Maya-1,采用 Apache 2.0 许可。该模型约 33 亿参数,基于 decoder-only LlamaForCausalLM 架构,通过自然语言描述设计音色并支持内联表情控制,可生成 20 多种情绪与语音风格。模型以 SNAC 离散音频码表示语音,解码为 24 kHz 单声道波形,并公开了 checkpoint、tokenizer 与参考推理代码。
以上摘要由 AI 生成,可能存在误差。事实请以原文为准。
正文节选
--- license: apache-2.0 library_name: transformers pipeline_tag: text-to-speech tags: - voice-design - expressive-speech - audio-generation --- # Maya-1 **Maya-1 is an open-weight text-to-speech model from Maya Research that combines natural-language voice descriptions with inline expressive controls.** It generates speech as discrete audio tokens, which are decoded into a 24 kHz waveform. Developers can describe a voice and its delivery in text instead of choosing only from a fixed speaker me
发布时间:2026-10-14 05:09
抓取时间:2026-10-11 21:10
来源机构:Hugging Face