Create app.py

app.py

@@ -0,0 +1,70 @@
+import gradio as gr
+import openai
+import whisper
+import time
+from gtts import gTTS
+from IPython.display import Audio
+model = whisper.load_model("base")
+openai.api_key = "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"
+
+res = []
+
+question = []
+
+def transcribe(audio):
+    
+    #time.sleep(3)
+    # load audio and pad/trim it to fit 30 seconds
+    audio = whisper.load_audio(audio)
+    audio = whisper.pad_or_trim(audio)
+
+    # make log-Mel spectrogram and move to the same device as the model
+    mel = whisper.log_mel_spectrogram(audio).to(model.device)
+
+    # detect the spoken language
+    _, probs = model.detect_language(mel)
+    print(f"Detected language: {max(probs, key=probs.get)}")
+
+    # decode the audio
+    options = whisper.DecodingOptions()
+    result = whisper.decode(model, mel, options)
+    res.append(result.text)
+
+    # chatgpt
+    n = len(res)
+    content = res[n-1]
+    messages.append({"role": "user", "content": content})
+
+    completion = openai.ChatCompletion.create(
+      model = "gpt-3.5-turbo",
+      messages = messages
+    )
+
+    chat_response = completion.choices[0].message.content
+
+    messages.append({"role": "assistant", "content": chat_response})
+
+    tts = gTTS(chat_response, lang='en', tld='us')
+    #tts = gTTS(chat_response, lang='zh-CN')
+
+    tts.save('1.wav')
+    sound_file = '1.wav'
+
+    question.append(chat_response)
+
+    return [result.text, chat_response, sound_file, res, question]
+
+output_1 = gr.Textbox(label="Speech to Text")
+output_2 = gr.Textbox(label="ChatGPT Output")
+output_3 = gr.Audio(label="Audio")
+
+gr.Interface(
+    title = 'TalktoAI，随时随地，谈天说地！', 
+    fn=transcribe, 
+    inputs=[
+        gr.inputs.Audio(source="microphone", type="filepath")
+    ],
+    outputs=[
+      output_1, output_2, output_3
+    ],
+    live=True).launch(share = True)