add more paremters to the chat generation

app.py

@@ -18,9 +18,6 @@ subprocess.run(
 )
 
 
-token = os.environ["HF_TOKEN"]
-
-
 model = AutoModelForCausalLM.from_pretrained(
     "microsoft/Phi-3-small-8k-instruct",
     torch_dtype="auto", 
@@ -43,8 +40,10 @@ model = model.to(device)
 
 
 @spaces.GPU(duration=60)
-def chat(message, history, temperature, do_sample, max_tokens):
-    chat = []
+def chat(message, history,system_prompt, temperature, do_sample, max_tokens, top_k, repetition_penalty, top_p):
+    chat = [
+        {"role": "assistant", "content": system_prompt}
+    ]
     for item in history:
         chat.append({"role": "user", "content": item[0]})
         if item[1] is not None:
@@ -62,6 +61,9 @@ def chat(message, history, temperature, do_sample, max_tokens):
         do_sample=True,
         temperature=temperature,
         eos_token_id=terminators,
+        top_k=top_k,
+        repetition_penalty=repetition_penalty,
+        top_p=top_p
     )
 
     if temperature == 0:
@@ -80,12 +82,16 @@ def chat(message, history, temperature, do_sample, max_tokens):
 
 demo = gr.ChatInterface(
     fn=chat,
-    examples=[["Write me a poem about Machine Learning."]],
-    # multimodal=False,
+    examples=[["Write me a poem about Machine Learning."],
+              ["write fibonacci sequence in python"],
+              ["who won the world cup in 2018?"],
+              ["when was the first computer invented?"],
+              ],
     additional_inputs_accordion=gr.Accordion(
         label="⚙️ Parameters", open=False, render=False
     ),
     additional_inputs=[
+        gr.Textbox("Perform the task to the best of your ability.", label="System prompt"),
         gr.Slider(
             minimum=0, maximum=1, step=0.1, value=0.9, label="Temperature", render=False
         ),
@@ -98,9 +104,12 @@ demo = gr.ChatInterface(
             label="Max new tokens",
             render=False,
         ),
+        gr.Slider(1, 80, 40, label="Top K sampling"),
+        gr.Slider(0, 2, 1.1, label="Repetition penalty"),
+        gr.Slider(0, 1, 0.95, label="Top P sampling"),
     ],
     stop_btn="Stop Generation",
-    title="Chat With Phi-3-Small-8k-8b-Instruct",
+    title="Chat With Phi-3-Small-8k-7b-Instruct",
     description="[microsoft/Phi-3-small-8k-instruct](https://huggingface.co/microsoft/Phi-3-small-8k-instruct)",
 )
 demo.launch()