上海电子网站建设,代理加盟网站,html5 微信网站 源码,页面设计的突出主体原则WeNet语音识别Qwen-72B-Chat Bot#x1f47e;Sambert-Hifigan语音合成 简介 利用 WeNet 进行语音识别#xff0c;使用户能够通过语音输入与系统进行交互。接着#xff0c;Qwen-72B-Chat Bot作为聊天机器人接收用户的语音输入或文本输入#xff0c;提供响应并与用户进行对话…WeNet语音识别Qwen-72B-Chat BotSambert-Hifigan语音合成 简介 利用 WeNet 进行语音识别使用户能够通过语音输入与系统进行交互。接着Qwen-72B-Chat Bot作为聊天机器人接收用户的语音输入或文本输入提供响应并与用户进行对话。最后系统利用 Sambert-Hifigan 进行语音合成将机器人的响应转换为自然流畅的语音输出使用户能够以语音方式接收机器人的回复。 特点 对话记忆功能 该系统能够记忆和追踪用户和聊天机器人之间的对话历史。这使得用户能够在对话中随时回顾之前的交流内容从而实现更连贯的对话和更好的交互体验。 多语音模型切换 该系统支持多种语音模型的切换。用户可以根据需要选择不同的语音模型进行交互。这种多语音模型切换功能使得系统在不同语境下有更强的适用性和灵活性。 界面 体验一下 环境配置 完整代码 import os
os.system(pip install dashscope)
os.system(pip install modelscope)
import gradio as gr
from http import HTTPStatus
import dashscope
from dashscope import Generation
from dashscope.api_entities.dashscope_response import Role
from typing import List, Optional, Tuple, Dict
from urllib.error import HTTPError
import wenet
from modelscope.outputs import OutputKeys
from modelscope.pipelines import pipeline
from modelscope.utils.constant import Tasksdefault_system You are a helpful assistant.
chs_model wenet.load_model(chinese)
YOUR_API_TOKEN os.getenv(YOUR_API_TOKEN)
dashscope.api_key YOUR_API_TOKEN
History List[Tuple[str, str]]
Messages List[Dict[str, str]]# 加载四个不同的语音合成模型
sambert_hifigan_zh_model_id damo/speech_sambert-hifigan_tts_zh-cn_16k
sambert_hifigan_zh pipeline(taskTasks.text_to_speech, modelsambert_hifigan_zh_model_id)sambert_hifigan_ch_model_id speech_tts/speech_sambert-hifigan_tts_chuangirl_Sichuan_16k
sambert_hifigan_ch pipeline(taskTasks.text_to_speech, modelsambert_hifigan_ch_model_id)sambert_hifigan_ca_model_id speech_tts/speech_sambert-hifigan_tts_jiajia_Cantonese_16k
sambert_hifigan_ca pipeline(taskTasks.text_to_speech, modelsambert_hifigan_ca_model_id)sambert_hifigan_ws_model_id speech_tts/speech_sambert-hifigan_tts_xiaoda_WuuShanghai_16k
sambert_hifigan_ws pipeline(taskTasks.text_to_speech, modelsambert_hifigan_ws_model_id)def clear_session() - History:return []def modify_system_session(system: str) - str:if system is None or len(system) 0:system default_systemreturn system, system, []def history_to_messages(history: History, system: str) - Messages:messages [{role: Role.SYSTEM, content: system}]for h in history:messages.append({role: Role.USER, content: h[0]})messages.append({role: Role.ASSISTANT, content: h[1]})return messagesdef messages_to_history(messages: Messages) - Tuple[str, History]:assert messages[0][role] Role.SYSTEMsystem messages[0][content]history []for q, r in zip(messages[1::2], messages[2::2]):history.append([q[content], r[content]])return system, historydef model_chat(path:str, history: Optional[History], system: str,model:str,voice:str
) - Tuple[str, str, History]:if path is not None:query chs_model.transcribe(path)[text]if query is None:query if history is None:history []messages history_to_messages(history, system)messages.append({role: Role.USER, content: query})gen Generation.call(model qwen-72b-chat,messagesmessages,result_formatmessage,streamTrue)for response in gen:if response.status_code HTTPStatus.OK:role response.output.choices[0].message.roleresponse response.output.choices[0].message.contentsystem, history messages_to_history(messages [{role: role, content: response}])else:raise HTTPError(Request id: %s, Status code: %s, error code: %s, error message: %s % (response.request_id, response.status_code,response.code, response.message))outputNone# 进行语音合成sambert_hifigan_tts_model {默认: sambert_hifigan_zh,四川话: sambert_hifigan_ch,粤语: sambert_hifigan_ca,上海话: sambert_hifigan_ws}# 使用对应的语音合成模型进行合成sambert_hifigan_tts sambert_hifigan_tts_model.get(model)if model 默认:output sambert_hifigan_tts(inputresponse, voicevoice)else:output sambert_hifigan_tts(inputresponse)wav output[OutputKeys.OUTPUT_WAV]path output.wavwith open(path, wb) as f:f.write(wav)return history, system, pathdef update_dropdowns(model,voice): if model 默认: voicegr.Dropdown(choices[zhitian_emo, zhiyan_emo, zhizhe_emo, zhibei_emo], valuezhitian_emo,label声音,visibleTrue) else: voicegr.Dropdown(choices[zhitian_emo, zhiyan_emo, zhizhe_emo, zhibei_emo], valuezhitian_emo,label声音,visibleFalse)return voice
with gr.Blocks() as demo:gr.Markdown(p aligncenterimg srchttps://modelscope.cn/api/v1/models/qwen/Qwen-VL-Chat/repo?RevisionmasterFilePathassets/logo.jpgViewtrue styleheight: 80px/p)gr.Markdown(centerfont size4WeNet语音识别Qwen-72B-Chat BotSambert-Hifigan语音合成/center)textbox gr.Microphone(typefilepath,label录音)with gr.Row():with gr.Column(scale3):system_input gr.Textbox(valuedefault_system, lines1, labelSystem, visibleFalse)with gr.Column(scale1):modify_system gr.Button(️ 设置system并清除历史对话, scale2, visibleFalse)system_state gr.Textbox(valuedefault_system, visibleFalse)chatbot gr.Chatbot(labelQwen-72B-Chat, visibleFalse)modelgr.Dropdown(choices[默认, 四川话, 粤语, 上海话], value默认,label声音模型)voice gr.Dropdown(choices[zhitian_emo, zhiyan_emo, zhizhe_emo, zhibei_emo], valuezhitian_emo,label声音)audio_output gr.Audio(typefilepath,label输出音频,autoPlayTrue)with gr.Row():clear_history gr.Button( 清除记忆)sumbit gr.Button( 发送)model.change(update_dropdowns,inputs[model,voice],outputs[voice])sumbit.click(model_chat,inputs[textbox, chatbot, system_state,model,voice],outputs[chatbot, system_input,audio_output],concurrency_limit10)clear_history.click(fnclear_session,inputs[],outputs[chatbot],concurrency_limit10)modify_system.click(fnmodify_system_session,inputs[system_input],outputs[system_state, system_input, chatbot],concurrency_limit10)
demo.queue(api_openFalse).launch(height800, shareFalse)
