Update app.py

app.py

@@ -1,5 +1,4 @@
 import gradio as gr
-import os
 import whisper
 from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
 from gtts import gTTS
@@ -14,7 +13,7 @@ def translate_audio(input_file, target_language):
         temp_audio_path = temp_audio.name
     
     # Auto to text (STT)
-    model = whisper.load_model("base")
+    model = whisper.Whisper("base")
     audio = whisper.load_audio(temp_audio_path)
     audio = whisper.pad_or_trim(audio)
     mel = whisper.log_mel_spectrogram(audio).to(model.device)
@@ -27,7 +26,8 @@ def translate_audio(input_file, target_language):
     # Translate
     tokenizer = AutoTokenizer.from_pretrained("alirezamsh/small100")
     model = AutoModelForSeq2SeqLM.from_pretrained("alirezamsh/small100")
-    tokenizer.src_lang = target_language
+    tokenizer.src_lang = lang
+    tokenizer.tgt_lang = target_language
     encoded_bg = tokenizer(text, return_tensors="pt")
     generated_tokens = model.generate(**encoded_bg)
     translated_text = tokenizer.batch_decode(generated_tokens, skip_special_tokens=True)[0]
@@ -41,12 +41,12 @@ def translate_audio(input_file, target_language):
 
 # Define Gradio interface
 inputs = [
- gr.inputs.File(label="Upload Audio File"),
- gr.inputs.Dropdown(choices=['en', 'es', 'fr', 'de', 'ru'], label="Target Language")
+    gr.inputs.File(label="Upload Audio File"),
+    gr.inputs.Dropdown(choices=['en', 'es', 'fr', 'de', 'ru'], label="Target Language")
 ]
 
 outputs = [
- gr.outputs.File(label="Translated Audio")
+    gr.outputs.File(label="Translated Audio")
 ]
 
 title = "Audio Translation"