upd: use whisper finetuned ru

app.py

@@ -21,11 +21,9 @@ model_ru2en = AutoModelForSeq2SeqLM.from_pretrained("Helsinki-NLP/opus-mt-ru-en"
 tokenizer_en2ru = AutoTokenizer.from_pretrained("Helsinki-NLP/opus-mt-en-ru")
 model_en2ru = AutoModelForSeq2SeqLM.from_pretrained("Helsinki-NLP/opus-mt-en-ru")
 
-# Load the speech recognition model
-transcriber = pipeline("automatic-speech-recognition", model="openai/whisper-base.en")
+transcriber = pipeline("automatic-speech-recognition", model="artyomboyko/whisper-base-fine_tuned-ru")
 
 
-# Functions for translation
 def translate_ru2en(text):
     inputs = tokenizer_ru2en(text, return_tensors="pt")
     outputs = model_ru2en.generate(**inputs)
@@ -40,7 +38,6 @@ def translate_en2ru(text):
     return translated_text
 
 
-# Function to generate answers
 def generate_answer_git(image, question):
     with torch.no_grad():
         encoding = processor(
@@ -90,7 +87,6 @@ def transcribe(stream, new_chunk):
     return stream, transcriber({"sampling_rate": sr, "raw": stream})["text"]
 
 
-# Gradio Interface
 interface = gr.Interface(
     fn=generate_answer,
     inputs=[