Update app.py

app.py

@@ -1,5 +1,5 @@
 import gradio as gr
-from transformers import pipeline, AutoProcessor, BarkModel
+from transformers import pipeline
 import cv2
 from PIL import Image
 import io
@@ -11,8 +11,7 @@ import torch
 #Commit
 def video_to_descriptions(video, target_language="en"):
 
-    processor = AutoProcessor.from_pretrained("suno/bark-small")
-    model = BarkModel.from_pretrained("suno/bark-small")
+    
     
     # Load the image-to-text and summarization pipelines
     ImgToText = pipeline("image-to-text", model="Salesforce/blip-image-captioning-large")
@@ -20,6 +19,7 @@ def video_to_descriptions(video, target_language="en"):
     
     # Load the translation pipeline for the target language
     translator = pipeline("translation", model=f"Helsinki-NLP/opus-mt-en-{target_language}")
+    audio = pipeline("text-to-speech", model="suno/bark-small")
     voice_preset = f"v2/{target_language}_speaker_1"
     
     
@@ -59,16 +59,10 @@ def video_to_descriptions(video, target_language="en"):
     translated_text = translator(summarized_description)[0]["translation_text"]
     print("TRANSLATION : " + translated_text)
     
-    inputs = processor(translated_text, voice_preset=voice_preset)
-    #audio_file = audio("translated_text", voice_preset=voice_preset)
-
-    
-    audio_array = model.generate(**inputs)
-    audio_array = audio_array.squeeze()
-    sample_rate = model.generation_config.sample_rate
+    audio_file = audio(translated_text)
 
     output_path = "./bark_out.wav"
-    scipy.io.wavfile.write(output_path, rate=sample_rate, data=audio_array)
+    scipy.io.wavfile.write(output_path, rate=audio_file["audio"][0], data=audio_file["sampling_rate"])
     
     return output_path