Interactive platform device with multiple voice modes

Abstract

This invention is an interactive platform device with multiple voice modes, comprising a terminal device, a computing device, and a data storage device. The terminal device has a user interface, receives input phrases, and displays system responses. The computing device, coupled to the terminal device, processes the input phrases and generates system responses. The computing device includes a local large-scale language model, a request processing module, and a prompting engineering module. The local large-scale language model performs inference in a local environment to generate system responses based on the input phrases or prompts. The request processing module receives the input phrases, and the prompting engineering module combines the input phrases with search results to form prompts. The data storage device has a retrieval module that performs similarity searches in a vector database based on the input phrases to obtain search results.

Claims

1. An interactive platform device with multiple voice modes, comprising: a terminal device having a user interface, the terminal device receiving an input phrase and displaying a system response through the user interface; a computing device coupled to the terminal device, the computing device processing the input phrase and generating the system response, the computing device including a local large language model, a request processing module and a prompting engineering module, wherein the local large language model performs inference in a local environment to generate the system response based on the input phrase or a prompt, the request processing module receiving the input phrase, and the prompting engineering module combining the input phrase with a search result to form the prompt; a data storage device having a search module performing a similarity search in a vector database based on the input phrase to obtain the search result.

2. The interactive platform device with multiple voice modes as described in claim 1, wherein the terminal device further includes a voice synthesis module that converts the system response into a voice output, the voice synthesis module being coupled to an audio output device that plays the voice output.

3. The interactive platform device with multiple voice modes as described in claim 2, wherein the voice output can support at least one of the following voices: Chinese, English or Taiwanese.

4. The interactive platform device with multiple voice modes as described in claim 2, wherein the terminal device can select a voice mode, and the prompting module combines the input phrase with the search result to form the prompt word according to the voice mode selected by the terminal device.

5. The multi-voice mode interactive platform device as described in claim 1, wherein the computing device further includes a model accelerator that accelerates the inference process of the local large language model.

Images