使用Python实现深度学习模型：智能语音助手与家庭管理

pip install speech_recognition pyaudio tensorflow

import speech_recognition as sr

def recognize_speech_from_mic():
    recognizer = sr.Recognizer()
    mic = sr.Microphone()

    with mic as source:
        print("请说话...")
        audio = recognizer.listen(source)

    try:
        text = recognizer.recognize_google(audio, language="zh-CN")
        print(f"你说的是: {text}")
        return text
    except sr.UnknownValueError:
        print("抱歉，我没有听懂。")
        return None
    except sr.RequestError:
        print("请求失败，请检查网络连接。")
        return None

# 测试语音识别功能
recognize_speech_from_mic()

import tensorflow as tf
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Dense, Embedding, LSTM
from tensorflow.keras.preprocessing.text import Tokenizer
from tensorflow.keras.preprocessing.sequence import pad_sequences

# 示例数据
sentences = ["打开灯", "关闭灯", "播放音乐", "暂停音乐"]
labels = [0, 1, 2, 3]  # 0: 打开灯, 1: 关闭灯, 2: 播放音乐, 3: 暂停音乐

# 数据预处理
tokenizer = Tokenizer(num_words=100)
tokenizer.fit_on_texts(sentences)
sequences = tokenizer.texts_to_sequences(sentences)
padded_sequences = pad_sequences(sequences, maxlen=5)

# 构建模型
model = Sequential([
    Embedding(input_dim=100, output_dim=16, input_length=5),
    LSTM(32),
    Dense(4, activation='softmax')
])

model.compile(optimizer='adam', loss='sparse_categorical_crossentropy', metrics=['accuracy'])

# 训练模型
model.fit(padded_sequences, labels, epochs=10)

# 保存模型
model.save("intent_model.h5")

import numpy as np
from tensorflow.keras.models import load_model

# 加载模型
model = load_model("intent_model.h5")

# 意图映射
intent_map = {0: "打开灯", 1: "关闭灯", 2: "播放音乐", 3: "暂停音乐"}

def predict_intent(text):
    sequence = tokenizer.texts_to_sequences([text])
    padded_sequence = pad_sequences(sequence, maxlen=5)
    prediction = model.predict(padded_sequence)
    intent = np.argmax(prediction)
    return intent_map[intent]

# 语音助手主程序
def voice_assistant():
    while True:
        text = recognize_speech_from_mic()
        if text:
            intent = predict_intent(text)
            print(f"执行操作: {intent}")

# 启动语音助手
voice_assistant()

import requests

def control_light(action):
    url = "http://smart-home-api/control"
    data = {"device": "light", "action": action}
    response = requests.post(url, json=data)
    if response.status_code == 200:
        print(f"灯已{action}")
    else:
        print("操作失败，请重试。")

# 在predict_intent函数中添加控制灯的逻辑
def predict_intent(text):
    sequence = tokenizer.texts_to_sequences([text])
    padded_sequence = pad_sequences(sequence, maxlen=5)
    prediction = model.predict(padded_sequence)
    intent = np.argmax(prediction)
    action = intent_map[intent]
    if "灯" in action:
        control_light(action.split("灯")[0])
    return action