Update app.py

app.py

@@ -1,4 +1,11 @@
 from transformers import pipeline, AutoTokenizer, T5ForConditionalGeneration
+import torch
+from pyannote.audio import Pipeline
+import os
+from faster_whisper import WhisperModel
+from pydub import AudioSegment
+from random import randint
+import gradio as gr
 
 
 model_name = "IlyaGusev/rut5_base_sum_gazeta"
@@ -23,21 +30,16 @@ def summ_mT5_G(text):
     summary = tokenizer.decode(output_ids, skip_special_tokens=True)
     return summary
 
-import torch
-
 
 # punctuation
 model_punc, example_texts, languages, punct, apply_te = torch.hub.load(repo_or_dir='snakers4/silero-models', model='silero_te')
 
+
 def punct(text):
     # print(text)
     return apply_te(text.lower(), lan='ru')
 
 
-from pyannote.audio import Pipeline
-import os
-
-
 pipeline_a = Pipeline.from_pretrained(
     "pyannote/speaker-diarization-3.1",
     use_auth_token=str(os.getenv("s1")))
@@ -71,15 +73,11 @@ def speackers_list(audio_f : str):
 
 # speackers = speackers_list(name_of_file)
 
-
-
-from faster_whisper import WhisperModel
-
-
 model_size = "large-v3"
 # Run on GPU with FP16
 model_tts = WhisperModel(model_size)  #, lan = "ru") #,  device="cpu", compute_type="int8") #, device="cuda", compute_type="float16")
 
+
 def speach_to_text(file_name):
 
     segments, info = model_tts.transcribe(file_name, beam_size=5)
@@ -124,7 +122,7 @@ class Segment_text:
     def get_speacker(self):
         return self.speacker
 
-from pydub import AudioSegment
+
 def init_segments(speackers, name_of_file):
     list_of_segments = []
     audio = AudioSegment.from_file(name_of_file)
@@ -147,15 +145,12 @@ def get_text_to_out(list_of_segments : list):
         res_sum += seg.get_speacker() + ":  " + seg.get_summarization() + "\n"
     return res_text, res_sum
 
-from random import randint
-
 
 def do_smth(file):
     audio = AudioSegment.from_wav(file)
     name_of_file = "f"+str(randint(1,10**8))
     audio.export(name_of_file, format="mp3")
 
-
     speackers = speackers_list(name_of_file)
 
     list_of_segments = init_segments(speackers, name_of_file)
@@ -164,8 +159,6 @@ def do_smth(file):
 
     return out_text, out_sum
 
-import gradio as gr
-
 
 demo = gr.Interface(
     do_smth,