import gradio as gr import whisper from transformers import pipeline # Load Whisper model for transcription whisper_model = whisper.load_model("base") # Load summarization and question-answering pipelines summarizer = pipeline("summarization", model="facebook/bart-large-cnn") qa_pipeline = pipeline("question-answering", model="distilbert-base-cased-distilled-squad") # Function to transcribe video def transcribe_video(video_path): result = whisper_model.transcribe(video_path) return result["text"] # Function to summarize text def summarize_text(text): summary = summarizer(text, max_length=130, min_length=30, do_sample=False) return summary[0]["summary_text"] # Function to answer questions def answer_question(summary, question): answer = qa_pipeline(question=question, context=summary) return answer["answer"] # Main function for Gradio interface def process_video(video_path): # Step 1: Transcribe the video transcript = transcribe_video(video_path) # Step 2: Summarize the transcript summary = summarize_text(transcript) return transcript, summary def ask_question(summary, question): # Step 3: Answer the question answer = answer_question(summary, question) return answer # Gradio interface with gr.Blocks() as demo: gr.Markdown("# Video Transcription, Summarization, and Q&A") with gr.Tab("Transcribe and Summarize"): video_input = gr.Video(label="Upload Video") transcript_output = gr.Textbox(label="Transcript", lines=10) summary_output = gr.Textbox(label="Summary", lines=5) transcribe_button = gr.Button("Transcribe and Summarize") with gr.Tab("Ask a Question"): summary_input = gr.Textbox(label="Paste Summary Here", lines=5) question_input = gr.Textbox(label="Your Question") answer_output = gr.Textbox(label="Answer", lines=2) ask_button = gr.Button("Ask Question") # Set up button actions transcribe_button.click( process_video, inputs=video_input, outputs=[transcript_output, summary_output] ) ask_button.click( ask_question, inputs=[summary_input, question_input], outputs=answer_output ) # Launch the app demo.launch()