Update app.py

app.py

@@ -54,7 +54,7 @@ global_data = {
     'eot': {},
     'eom': {},
     'lf': {},
-    'max_tokens': 999999,
+    'max_tokens': {},
     'tokenizers': {},
     'model_params': {},
     'model_size': {},
@@ -252,6 +252,7 @@ model_configs = [
         "name": "Meta-Llama-3.1-70B"
     }
 ]
+
 class ModelManager:
     def __init__(self):
         self.models = {}
@@ -307,7 +308,7 @@ def cache_response(func):
 @cache_response
 def generate_model_response(model, inputs):
     try:
-        response = model(inputs)
+        response = model(inputs, max_tokens=9999999)
         return remove_duplicates(response['choices'][0]['text'])
     except Exception as e:
         return ""
@@ -331,9 +332,7 @@ async def process_message(message):
             for model_name, future in zip(global_data['models'].keys(), as_completed(futures))
         ]
     unique_responses = remove_repetitive_responses(responses)
-    formatted_response = ""
-    for model, response in unique_responses.items():
-        formatted_response += f"**{model}:**\n{response}\n\n"
+    formatted_response = next(iter(unique_responses.values()))  # Get a single response
     return formatted_response
 
 app = FastAPI()