Update app.py

app.py

@@ -4,31 +4,31 @@ import json
 import time
 
 # Initialize the pipeline with the new model
-pipe = pipeline("text-generation", model="Blexus/Quble_test_model_v1_INSTRUCT_v2")
+pipe = pipeline("text-generation", model="Blexus/originos-icn-savant-instruct")
 
-def format_prompt(message, system, history):
-    prompt = f"SYSTEM: {system} <|endofsystem|>"
+def format_prompt(message, history):
+    prompt = ""
     
-    for entry in history:
-        if len(entry) == 2:
-            user_prompt, bot_response = entry
-            prompt += f"USER: {user_prompt} <|endofuser|>\nASSISTANT: {bot_response}<|endoftext|>\n"
+    #for entry in history:
+    #    if len(entry) == 2:
+    #        user_prompt, bot_response = entry
+    #        prompt += f"USER: {user_prompt} <|endofuser|>\nASSISTANT: {bot_response}<|endoftext|>\n"
     
-    prompt += f"USER: {message}<|endofuser|>\nASSISTANT:"
+    prompt += f"<|in|> {message} <|out|>"
     return prompt
 
-def generate(prompt, system, history, temperature=0.9, max_new_tokens=4096, top_p=0.9, repetition_penalty=1.2):
+def generate(prompt, history, temperature=0.9, max_new_tokens=4096, top_p=0.9, repetition_penalty=1.2):
     temperature = float(temperature)
     if temperature < 1e-2:
         temperature = 1e-2
     top_p = float(top_p)
 
-    formatted_prompt = format_prompt(prompt, system, history)
-    response_text = "We are sorry but Quble doesn't know how to answer."
+    formatted_prompt = format_prompt(prompt, history)
+    response_text = "We are sorry but ICN savant doesn't know how to answer."
     # Generate the response without streaming
     try:
             response = pipe(formatted_prompt, max_new_tokens=max_new_tokens, temperature=temperature, top_p=top_p, repetition_penalty=repetition_penalty)[0]["generated_text"]
-            response_text = response.split("ASSISTANT:")[-1].strip()
+            response_text = response.split("<|end|>")[-1].strip()
 
             # Simulate streaming by yielding parts of the response
             accumulated_response = ""  # To keep track of the full response
@@ -48,12 +48,6 @@ customCSS = """
 """
 
 additional_inputs = [
-    gr.Textbox(
-        label="System prompt",
-        value="You are a helpful intelligent assistant. Your name is Quble.",
-        info="System prompt",
-        interactive=True,
-    ),
     gr.Slider(
         label="Temperature",
         value=0.9,