On October 20th, the ants will announce todayOpen Source Ming-3.0 Voice Series, which provides voice synthesis, audio understanding and dialogue, and voice recognition, respectively, with five models. Developers can choose between tasks or combine different competencies into their own work streams。
The Voice Design of the Ming-3.0-ttts series can describe the target sound in text while also supporting the audio setting of the reference audio. More specifically, creators can use director models to describe the performance of lines, how emotions change, which words need to be read again, and how the rhythm should be addressed。
In ants, from the main conversation and voice call of Alfred App to the doctor's voice-cloning, Tiger's smart office, Ming-3.0-ttts have entered a variety of business. It is able to carry both an independent voice and a combination of Ming-3.0-asr-tiny, linking both the voice input and the answers。
Ming-3.0-audio-flash/tiny supports the complex understanding and structured output of transliteration, summary and analysis of views directly on long audio. The relatively traditional process of first-speech recognition and then analysis preserves emotions in dialogue, talking person identities, environmental information, etc., as well as the need for a model for rapid reasoning of long context sequences。
Ming-3.0-asr-tiny provides a comprehensive voice recognition for the real business scene by building high-quality language materials such as Mandarin, English, dialects, field hot words, and rationally controlling distribution, combined with a strong context understanding of the language model base, using candidate hot word lists, dialogue context information。
In ants, Ming-3.0-asr-tiny has been used for Dr. Afu ' s assistant, payment treasure search, real-time voice dialogue with Ant Xiaobao and safe core。
1AI WITH THE FOLLOWING RELEVANT LINKS:
Online Experience:https://ling.tbox.cn
-
Online experience: access to Ling Studio Chat, selection of Ming-3.0-udio-flash to experience audio understanding and dialogue
-
API Call: Ling Studio provides API services for Ming-3.0-tts-0.5B, Ming-3.0-udio-flash and Ming-3.0-asr-tiny
Min-3.0-ttts-0.5B
-
Hugging Face:https://huggingface.co/inclusionAI/Ming-3.0-tts-0.5B
-
ModelScope:https://www.modelscope.cn/models/inclusionAI/Ming-3.0-tts-0.5B
Min-3.0-tts-tiny
-
Hugging Face:https://huggingface.co/inclusionAI/Ming-3.0-tts-tiny
-
ModelScope:https://www.modelscope.cn/models/inclusionAI/Ming-3.0-tts-tiny
Ming-3.0-udio-flash
-
Hugging Face:https://huggingface.co/inclusionAI/Ming-3.0-audio-flash
-
ModelScope:https://www.modelscope.cn/models/inclusionAI/Ming-3.0-audio-flash
Min-3.0-Audio-tiny
-
Hugging Face:https://huggingface.co/inclusionAI/Ming-3.0-audio-tiny
-
ModelScope:https://www.modelscope.cn/models/inclusionAI/Ming-3.0-audio-tiny
Min-3.0-asr-tiny
-
Hugging Face:https://huggingface.co/inclusionAI/Ming-3.0-asr-tiny
-
ModelScope:https://www.modelscope.cn/models/inclusionAI/Ming-3.0-asr-tiny