Update app.py

app.py

@@ -12,9 +12,7 @@ model2api = [
              "timdettmers/guanaco-33b-merged",
 ]
 
-BOT_NAME = "Assistant"
-
-STOP_SEQUENCES = ["\nUser:", " User:", "###", "</s>"]
+STOP_SEQUENCES = ["\nUser:", " User:", "<|endoftext|>", "</s>"]
 
 EXAMPLES = [
     ["Hey LLAMA! Any recommendations for my holidays in Abu Dhabi?"],
@@ -24,13 +22,13 @@ EXAMPLES = [
     ["Can you write a short tweet about the release of our latest AI model, LLAMA LLM?"]
     ]
 
-def format_prompt(message, history, system_prompt):
+def format_prompt(message, history, system_prompt, bot_name):
   prompt = ""
   if system_prompt:
     prompt += f"System: {system_prompt}\n"
   for user_prompt, bot_response in history:
     prompt += f"User: {user_prompt}\n"
-    prompt += f"{BOT_NAME}: {bot_response}\n"
+    prompt += f"{bot_name}: {bot_response}\n"
   prompt += f"""User: {message}\n{BOT_NAME}:"""
   return prompt
 
@@ -54,7 +52,6 @@ def generate(
         seed=seed,
     )
     seed = seed + 1
-    formatted_prompt = format_prompt(prompt, history, system_prompt)
 
     client = InferenceClient()
     clientList = (client.list_deployed_models('text-generation-inference'))['text-generation']
@@ -65,6 +62,13 @@ def generate(
             print(f"Choosen model: {model}")
             break
 
+    if model == model2api[0]:
+        bot_name = "Falcon"
+    else:
+        bot_name = "Assistant"
+    
+    formatted_prompt = format_prompt(prompt, history, system_prompt, bot_name)
+    
     try:
         stream = client.text_generation(formatted_prompt, **generate_kwargs, stream=True, details=True, return_full_text=False)
         output = ""