Update app.py

app.py

@@ -15,6 +15,20 @@ MODELS = {
     'Bengali': "Helsinki-NLP/opus-mt-en-bn",
 }
 
+# Manually defined language codes for different language models
+LANG_CODE_MAP = {
+    'French': 'fr',
+    'Spanish': 'es',
+    'German': 'de',
+    'Chinese': 'zh',
+    'Russian': 'ru',
+    'Japanese': 'ja',
+    'Arabic': 'ar',
+    'Urdu': 'ur',
+    'Hindi': 'hi',
+    'Bengali': 'bn',
+}
+
 def translate_text(text, target_lang):
     # Load the appropriate model and tokenizer for the target language
     model_name = MODELS.get(target_lang)
@@ -26,8 +40,15 @@ def translate_text(text, target_lang):
     
     # Encode the text and prepare it for translation
     encoded_text = tokenizer(text, return_tensors="pt")
+    
+    # Get the language code for forced_bos_token_id
+    lang_code = LANG_CODE_MAP.get(target_lang)
+    if not lang_code:
+        return "Error: Language code not found."
+    
     # Translate text
-    translated = model.generate(**encoded_text, forced_bos_token_id=tokenizer.get_lang_id(target_lang))
+    translated = model.generate(**encoded_text, forced_bos_token_id=tokenizer.lang_code_to_id.get(lang_code))
+    
     # Decode the translated text
     translated_text = tokenizer.decode(translated[0], skip_special_tokens=True)
     return translated_text
@@ -47,3 +68,4 @@ if st.button('Translate'):
     else:
         st.error("Please enter text to translate.")
 
+