Spaces:
Sleeping
Sleeping
| # -*- coding: utf-8 -*- | |
| """Untitled | |
| Automatically generated by Colab. | |
| Original file is located at | |
| https://colab.research.google.com/drive/12GhPKbBzxei0ZhB0r-m5kvNOaCRyCxiM | |
| """ | |
| !pip install gradio openai gtts pydub numpy requests groq openai-whisper transformers | |
| !apt-get install -y ffmpeg | |
| import os | |
| os.environ["GROQ_API_KEY"] = "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" | |
| import os | |
| import gradio as gr | |
| import whisper | |
| from gtts import gTTS | |
| import io | |
| from transformers import pipeline | |
| from groq import Groq | |
| # Initialize the Groq client | |
| client = Groq(api_key=os.environ.get("GROQ_API_KEY")) | |
| # Load the Whisper model | |
| whisper_model = whisper.load_model("base") # You can choose other models like "small", "medium", "large" | |
| # Initialize the grammar correction pipeline | |
| corrector = pipeline("text2text-generation", model="pszemraj/flan-t5-large-grammar-synthesis") | |
| def process_audio(file_path): | |
| try: | |
| # Load the audio file | |
| audio = whisper.load_audio(file_path) | |
| # Transcribe the audio using Whisper | |
| result = whisper_model.transcribe(audio) | |
| user_text = result["text"] | |
| # Display the user input text | |
| corrected_text = corrector(user_text)[0]['generated_text'].strip() | |
| # Generate a response using Groq | |
| chat_completion = client.chat.completions.create( | |
| messages=[{"role": "user", "content": corrected_text}], | |
| model="llama3-8b-8192", # Replace with the correct model if necessary | |
| ) | |
| # Access the response using dot notation | |
| response_message = chat_completion.choices[0].message.content.strip() | |
| # Convert the response text to speech | |
| tts = gTTS(response_message) | |
| response_audio_io = io.BytesIO() | |
| tts.write_to_fp(response_audio_io) # Save the audio to the BytesIO object | |
| response_audio_io.seek(0) | |
| # Save audio to a file to ensure it's generated correctly | |
| with open("response.mp3", "wb") as audio_file: | |
| audio_file.write(response_audio_io.getvalue()) | |
| # Return the original text, corrected text, and the path to the saved audio file | |
| return user_text, corrected_text, "response.mp3" | |
| except Exception as e: | |
| return f"An error occurred: {e}", None, None | |
| # Create a Gradio interface with a submit button | |
| iface = gr.Interface( | |
| fn=process_audio, | |
| inputs=gr.Audio(type="filepath"), # Use type="filepath" | |
| outputs=[ | |
| gr.Textbox(label="User voice input into text"), # Original user input text | |
| gr.Textbox(label="Corrected version of user input"), # Corrected text | |
| gr.Audio(label="Response Audio") # Response audio | |
| ], | |
| live=False, # Ensure live mode is off to use a submit button | |
| title="Audio Processing with Grammar Correction", | |
| description="Upload an audio file, which will be transcribed, corrected for grammar, and then used to generate a response.", | |
| allow_flagging="never" | |
| ) | |
| iface.launch() | |
| # import os | |
| # import gradio as gr | |
| # import whisper | |
| # from gtts import gTTS | |
| # import io | |
| # from transformers import pipeline, AutoTokenizer, AutoModelForSeq2SeqLM | |
| # from groq import Groq | |
| # # Initialize the Groq client | |
| # client = Groq(api_key=os.environ.get("GROQ_API_KEY")) | |
| # # Load the Whisper model | |
| # whisper_model = whisper.load_model("base") # You can choose other models like "small", "medium", "large" | |
| # # Initialize the grammar correction pipeline | |
| # corrector = pipeline("text2text-generation", model="pszemraj/flan-t5-large-grammar-synthesis") | |
| # def process_audio(file_path): | |
| # try: | |
| # # Load the audio file | |
| # audio = whisper.load_audio(file_path) | |
| # # Transcribe the audio using Whisper | |
| # result = whisper_model.transcribe(audio) | |
| # user_text = result["text"] | |
| # # Display the user input text | |
| # corrected_text = corrector(user_text)[0]['generated_text'].strip() | |
| # # Generate a response using Groq | |
| # chat_completion = client.chat.completions.create( | |
| # messages=[{"role": "user", "content": corrected_text}], | |
| # model="llama3-8b-8192", # Replace with the correct model if necessary | |
| # ) | |
| # # Access the response using dot notation | |
| # response_message = chat_completion.choices[0].message.content.strip() | |
| # # Convert the response text to speech | |
| # tts = gTTS(response_message) | |
| # response_audio_io = io.BytesIO() | |
| # tts.write_to_fp(response_audio_io) # Save the audio to the BytesIO object | |
| # response_audio_io.seek(0) | |
| # # Save audio to a file to ensure it's generated correctly | |
| # with open("response.mp3", "wb") as audio_file: | |
| # audio_file.write(response_audio_io.getvalue()) | |
| # # Return the original text, corrected text, and the path to the saved audio file | |
| # return user_text, corrected_text, "response.mp3" | |
| # except Exception as e: | |
| # return f"An error occurred: {e}", None, None | |
| # iface = gr.Interface( | |
| # fn=process_audio, | |
| # inputs=gr.Audio(type="filepath"), # Use type="filepath" | |
| # outputs=[ | |
| # gr.Textbox(label="User voice input into text"), # Original user input text | |
| # gr.Textbox(label="Corrected version of user input"), # Corrected text | |
| # gr.Audio(label="Response Audio") # Response audio | |
| # ], | |
| # live=True | |
| # ) | |
| # iface.launch() | |
| # # import os | |
| # # import gradio as gr | |
| # # import whisper | |
| # # from gtts import gTTS | |
| # # import io | |
| # # from groq import Groq | |
| # # # Initialize the Groq client | |
| # # client = Groq(api_key=os.environ.get("GROQ_API_KEY")) | |
| # # # Load the Whisper model | |
| # # model = whisper.load_model("base") # You can choose other models like "small", "medium", "large" | |
| # # def process_audio(file_path): | |
| # # try: | |
| # # # Load the audio file | |
| # # audio = whisper.load_audio(file_path) | |
| # # # Transcribe the audio using Whisper | |
| # # result = model.transcribe(audio) | |
| # # text = result["text"] | |
| # # # Generate a response using Groq | |
| # # chat_completion = client.chat.completions.create( | |
| # # messages=[{"role": "user", "content": text}], | |
| # # model="llama3-8b-8192", # Replace with the correct model if necessary | |
| # # ) | |
| # # # Access the response using dot notation | |
| # # response_message = chat_completion.choices[0].message.content.strip() | |
| # # # Convert the response text to speech | |
| # # tts = gTTS(response_message) | |
| # # response_audio_io = io.BytesIO() | |
| # # tts.write_to_fp(response_audio_io) # Save the audio to the BytesIO object | |
| # # response_audio_io.seek(0) | |
| # # # Save audio to a file to ensure it's generated correctly | |
| # # with open("response.mp3", "wb") as audio_file: | |
| # # audio_file.write(response_audio_io.getvalue()) | |
| # # # Return the response text and the path to the saved audio file | |
| # # return response_message, "response.mp3" | |
| # # except Exception as e: | |
| # # return f"An error occurred: {e}", None | |
| # # iface = gr.Interface( | |
| # # fn=process_audio, | |
| # # inputs=gr.Audio(type="filepath"), # Use type="filepath" | |
| # # outputs=[gr.Textbox(label="Response Text"), gr.Audio(label="Response Audio")], | |
| # # live=True | |
| # # ) | |
| # # iface.launch() | |