only using one model now

app.py

@@ -36,36 +36,35 @@ def translation(text,dest_lang,dest_lang_code, src_lang_code):
 
     headers = {"Authorization": f"Bearer {api_token_header}"}
 
-    # Bengali Done
-    if(dest_lang == "Bengali" and src_lang_code == "en_XX"):
-        API_URL = "https://api-inference.huggingface.co/models/csebuetnlp/banglat5_nmt_en_bn"
-        def query(payload):
-            response = requests.post(API_URL, headers=headers, json=payload)
-            return response.json()
-        output = query({
-            "inputs": text,
-        })
-        print(output)
-        return output[0]['translation_text']
+    # # Bengali Done
+    # if(dest_lang == "Bengali" and src_lang_code == "en_XX"):
+    #     API_URL = "https://api-inference.huggingface.co/models/csebuetnlp/banglat5_nmt_en_bn"
+    #     def query(payload):
+    #         response = requests.post(API_URL, headers=headers, json=payload)
+    #         return response.json()
+    #     output = query({
+    #         "inputs": text,
+    #     })
+    #     print(output)
+    #     return output[0]['translation_text']
+    # else:
+    global model
+    if model:
+        pass
     else:
-        global model
-        if model:
-            pass
-        else:
-            model = load_model()
-        loaded_model = model
-        tokenizer = MBart50TokenizerFast.from_pretrained("SnypzZz/Llama2-13b-Language-translate", src_lang=src_lang_code)
-        #model_inputs = tokenizer(text, return_tensors="pt")
-        loaded_model_inputs = tokenizer(text, return_tensors="pt")
-
-        # translate
-        generated_tokens = loaded_model.generate(
-            **loaded_model_inputs,
-            forced_bos_token_id=tokenizer.lang_code_to_id[dest_lang_code]
-        )
-        output = tokenizer.batch_decode(generated_tokens, skip_special_tokens=True)
-        print(output)
-        return output[0]
+        model = load_model()
+    loaded_model = model
+    tokenizer = MBart50TokenizerFast.from_pretrained("SnypzZz/Llama2-13b-Language-translate", src_lang=src_lang_code)
+    #model_inputs = tokenizer(text, return_tensors="pt")
+    loaded_model_inputs = tokenizer(text, return_tensors="pt")
+    # translate
+    generated_tokens = loaded_model.generate(
+        **loaded_model_inputs,
+        forced_bos_token_id=tokenizer.lang_code_to_id[dest_lang_code]
+    )
+    output = tokenizer.batch_decode(generated_tokens, skip_special_tokens=True)
+    print(output)
+    return output[0]
     
 
 def main_translation(text,dest_lang_code,src_lang_code):