try to reduce loading model time

app.py

@@ -26,6 +26,9 @@ import argparse
 import langid
 from transformers import pipeline, AutoModelForSeq2SeqLM, AutoTokenizer
 
+tokenizer = AutoTokenizer.from_pretrained("facebook/nllb-200-distilled-600M")
+model = AutoModelForSeq2SeqLM.from_pretrained("facebook/nllb-200-distilled-600M")
+
 
 class myTheme(Base):
     def __init__(
@@ -128,8 +131,6 @@ def opus_trans(article, target_language):
 def nllb_trans(article, target_language):
     result_lang = detect_lang(article)
 
-    tokenizer = AutoTokenizer.from_pretrained("facebook/nllb-200-distilled-600M")
-    model = AutoModelForSeq2SeqLM.from_pretrained("facebook/nllb-200-distilled-600M")
     inputs = tokenizer(article, return_tensors="pt")
 
     if target_language == "English":