Update app.py

app.py

@@ -15,7 +15,7 @@ def format_prompt(message, history):
   return prompt
 
 def generate(
-    prompt, history, system_prompt, temperature=0.9, max_new_tokens=1000, top_p=0.95, repetition_penalty=1.0,
+    prompt, history, system_prompt, temperature=0.9, max_new_tokens=1000, top_p=0.95, repetition_penalty=1.0, seed=42
 ):
     temperature = float(temperature)
     if temperature < 1e-2:
@@ -28,7 +28,7 @@ def generate(
         top_p=top_p,
         repetition_penalty=repetition_penalty,
         do_sample=True,
-        seed=42,
+        seed=seed,
     )
 
     formatted_prompt = format_prompt(f"{system_prompt}, {prompt}", history)
@@ -46,6 +46,15 @@ additional_inputs=[
         label="System Prompt",
         interactive=True,
     ),
+    gr.Slider(
+        label="Max new tokens",
+        value=1000,
+        minimum=0,
+        maximum=32768,
+        step=64,
+        interactive=True,
+        info="The maximum numbers of new tokens, controls how long is the output",
+    ),
     gr.Slider(
         label="Temperature",
         value=0.9,
@@ -56,13 +65,13 @@ additional_inputs=[
         info="Higher values produce more diverse outputs",
     ),
     gr.Slider(
-        label="Max new tokens",
-        value=1000,
-        minimum=0,
-        maximum=32768,
-        step=64,
+        label="Repetition penalty",
+        value=1.2,
+        minimum=1.0,
+        maximum=2.0,
+        step=0.05,
         interactive=True,
-        info="The maximum numbers of new tokens",
+        info="Penalize repeated tokens, making the AI repeat less itself",
     ),
     gr.Slider(
         label="Top-p (nucleus sampling)",
@@ -74,13 +83,19 @@ additional_inputs=[
         info="Higher values sample more low-probability tokens",
     ),
     gr.Slider(
-        label="Repetition penalty",
-        value=1.2,
-        minimum=1.0,
-        maximum=2.0,
+        label="Top-k",
+        value=0.90,
+        minimum=0.0,
+        maximum=1,
         step=0.05,
         interactive=True,
-        info="Penalize repeated tokens",
+        info="Higher k means more diverse outputs by considering a range of tokens",
+    ),
+    gr.Number(
+        label="Seed",
+        value=42,
+        minimum=1,
+        info="A starting point to initiate the generation process",
     )
 ]