Step1o: Launching the first end-to-end speech grand model with hundreds of billions of parameters " Step1o" in China

Step-1o, the first end-to-end speech model with hundreds of billions of parameters in China, introduces the many defects of traditional speech models using cascade solutions, while end-to-end speech solutions can realize the integration of speech comprehension and generation, and improve the model's IQ and IQ upper limit. step-1o supports mixed-format inputs and outputs, can respond quickly and can be interrupted at any time, can deeply understand and imitate multiple voice features, and can improve the quality of replies through self-learning, as well as inheriting the ability to create speech models. Step-1o supports mixed forms of input and output, can respond quickly and can be interrupted at any time, can deeply understand and imitate a variety of voice characteristics, can improve the quality of response through self-learning, both the ability to provide professional advice and emotional value accompaniment and inheritance of the language large model creation capabilities, it will be accessed in the near future to provide users with real-time voice call service on the Leapfrog App.