Update app.py

app.py

@@ -1,21 +1,19 @@
 from huggingface_hub import InferenceClient
 import gradio as gr
 
-client = InferenceClient(
-    "mistralai/Mistral-7B-Instruct-v0.3"
-)
+client = InferenceClient("mistralai/Mistral-7B-Instruct-v0.3")
 
-
-def format_prompt(message, history):
-  prompt = "<s>"
-  for user_prompt, bot_response in history:
-    prompt += f"[INST] {user_prompt} [/INST]"
-    prompt += f" {bot_response}</s> "
-  prompt += f"[INST] {message} [/INST]"
-  return prompt
+def format_prompt(message, history, persona=None):
+    prompt = "<s>"
+    if persona:
+        prompt += f"[ROLE: {persona}] "
+    for user_prompt, bot_response in history:
+        prompt += f"[INST] {user_prompt} [/INST] {bot_response}</s> "
+    prompt += f"[INST] {message} [/INST]"
+    return prompt
 
 def generate(
-    prompt, history, temperature=0.9, max_new_tokens=256, top_p=0.95, repetition_penalty=1.0,
+    prompt, history, temperature=0.9, max_new_tokens=256, top_p=0.95, repetition_penalty=1.0, persona=None
 ):
     temperature = float(temperature)
     if temperature < 1e-2:
@@ -31,7 +29,7 @@ def generate(
         seed=42,
     )
 
-    formatted_prompt = format_prompt(prompt, history)
+    formatted_prompt = format_prompt(prompt, history, persona)
 
     stream = client.text_generation(formatted_prompt, **generate_kwargs, stream=True, details=True, return_full_text=False)
     output = ""
@@ -42,7 +40,7 @@ def generate(
     return output
 
 
-additional_inputs=[
+additional_inputs = [
     gr.Slider(
         label="Temperature",
         value=0.9,
@@ -78,18 +76,24 @@ additional_inputs=[
         step=0.05,
         interactive=True,
         info="Penalize repeated tokens",
-    )
+    ),
+    gr.Textbox(
+        label="Persona",
+        placeholder="Describe the role (e.g., 'wise mentor', 'friendly assistant')",
+        info="Define a persona for the model to roleplay.",
+    ),
 ]
 
 
 gr.ChatInterface(
-    fn=generate,
+    fn=lambda prompt, history, temperature, max_new_tokens, top_p, repetition_penalty, persona: generate(
+        prompt, history, temperature, max_new_tokens, top_p, repetition_penalty, persona
+    ),
     chatbot=gr.Chatbot(show_label=False, show_share_button=False, show_copy_button=True, likeable=True, layout="panel"),
     additional_inputs=additional_inputs,
-    title="""Mistral 7B v0.3"""
+    title="Mistral 7B v0.3"
 ).launch(show_api=False)
 
-
 gr.load("models/ehristoforu/dalle-3-xl-v2").launch()
 
-gr.load("models/microsoft/Phi-3-mini-4k-instruct").launch()
+gr.load("models/microsoft/Phi-3-mini-4k-instruct").launch()