Update app.py

app.py

@@ -38,6 +38,17 @@ def load_models(inp):
         out_box[z]=(gr.update(label=models[inp[z]]))
     return out_box[0],out_box[1],out_box[2],out_box[3]
 
+def format_prompt_default(message, history):
+    prompt = ""
+    if history:
+        #<start_of_turn>userHow does the brain work?<end_of_turn><start_of_turn>model
+        for user_prompt, bot_response in history:
+            prompt += f"{user_prompt}\n"
+            print(prompt)
+            prompt += f"{bot_response}\n"
+            print(prompt)
+    prompt += f"{message}\n"
+    return prompt
 
 def format_prompt_gemma(message, history):
     prompt = ""
@@ -60,13 +71,14 @@ def format_prompt_mixtral(message, history):
             prompt += f" {bot_response}</s> "
     prompt += f"[INST] {message} [/INST]"
     return prompt
+
 def format_prompt_choose(message, history, model_name):
-    if "gemma" in models[model_name].lower():
+    if "gemma" in models[model_name].lower() and "it" in models[model_name].lower():
         return format_prompt_gemma(message,history)
     if "mixtral" in models[model_name].lower():
         return format_prompt_mixtral(message,history)
     else:
-        return format_prompt_mixtral(message,history)
+        return format_prompt_default(message,history)