Update app.py

app.py

@@ -17,15 +17,15 @@ def format_prompt(input_text, history):
     <start_of_turn>model"""    
     return prompt
 
-def generate(prompt, history):
+def generate(prompt, history, temperature=0.95, max_new_tokens=512, top_p=0.9, repetition_penalty=1.0):
     if not history:
         history = []
 
     kwargs = dict(
-        temperature=1.0,
-        max_new_tokens=512,
-        top_p=0.9,
-        repetition_penalty=1,
+        temperature=temperature,
+        max_new_tokens=max_new_tokens,
+        top_p=top_p,
+        repetition_penalty=repetition_penalty,
         do_sample=True,
     )
 
@@ -40,6 +40,45 @@ def generate(prompt, history):
     return output
 
 
+additional_inputs=[
+    gr.Slider(
+        label="Temperature",
+        value=0.85,
+        minimum=0.1,
+        maximum=1.0,
+        step=0.05,
+        interactive=True,
+        info="A higher value (> 1) will generate randomness and variability in the model response",
+    ),
+    gr.Slider(
+        label="Max new tokens",
+        value=512,
+        minimum=128,
+        maximum=1048,
+        step=64,
+        interactive=True,
+        info="The maximum numbers of new tokens generated in the model response",
+    ),
+    gr.Slider(
+        label="Top-p (random sampling)",
+        value=0.80,
+        minimum=0.1,
+        maximum=1,
+        step=0.05,
+        interactive=True,
+        info="A smaller value generates the highest probability tokens, a higher value (~ 1) allows low-probability tokens",
+    ),
+    gr.Slider(
+        label="Repetition penalty",
+        value=1.0,
+        minimum=0.5,
+        maximum=2.0,
+        step=0.05,
+        interactive=True,
+        info="Penalizes repeated tokens in model response",
+    )
+]
+
 chatbot = gr.Chatbot(height=500)
 with gr.Blocks(theme=gr.themes.Soft()) as demo:
     gr.HTML("<center><h1>Google Gemma 7B IT</h1><center>")
@@ -50,6 +89,7 @@ with gr.Blocks(theme=gr.themes.Soft()) as demo:
         undo_btn=None,
         clear_btn="Clear",
         description="This chatbot is using a Hugging Face Inference Client for the google/gemma-7b-it model.",
+        additional_inputs=additional_inputs,
         examples=[["Explain artificial intelligence in a few lines."]]
     )
 demo.queue().launch()