Reproduce Vidtext with distilled whisper

app.py

@@ -4,25 +4,17 @@ from transformers import pipeline
 from pytube import YouTube
 from pydub import AudioSegment
 from audio_extract import extract_audio
-import google.generativeai as google_genai
-
 import os
 from dotenv import load_dotenv
 
-
-
 load_dotenv()
 
-GOOGLE_API_KEY =os.getenv("GOOGLE_API_KEY")
-
-google_genai.configure(api_key=GOOGLE_API_KEY)
-
 st.set_page_config(
-    page_title="VidText"
+    page_title="VidText_distilled"
 )
 
-st.title('Vidtext_whisper')
-st.write('A web app for video/audio transcription(Youtube, mp4, mp3)')
+st.title('Vidtext_distilwhisper')
+st.write('A web app for video/audio transcription(Youtube, mp4, mp3). Using distilled Whisper')
 
 
 def youtube_video_downloader(url):
@@ -57,7 +49,7 @@ def audio_processing(mp3_audio):
 
 @st.cache_resource
 def load_asr_model():
-    asr_model = pipeline(task="automatic-speech-recognition", model="openai/whisper-small")
+    asr_model = pipeline(task="automatic-speech-recognition", model="distil-whisper/distil-large-v3")
     return asr_model
 
 transcriber_model = load_asr_model()
@@ -66,11 +58,6 @@ def transcriber_pass(processed_audio):
     text_extract = transcriber_model(processed_audio)
     return text_extract['text']
 
-def generate_ai_summary(transcript):
-    model = google_genai.GenerativeModel('gemini-pro')
-    model_response = model.generate_content([f"Give a summary of the text {transcript}"], stream=True)
-    return model_response.text
-
     
 
 # Streamlit UI