Update app.py

app.py

@@ -63,7 +63,7 @@ def evaluate(instruction, input, model, tokenizer):
         result.append( output.split("### Response:")[1].strip())
     return ' '.join(el for el in result)            
 
-def inference(model_name, text, input):
+def inference(model_name, text, input, temperature, top_p, num_beams):
     model = load_model(model_name)
     tokenizer = load_tokenizer(model_name)
     output = evaluate(instruction = text, input = input, model = model, tokenizer = tokenizer)
@@ -72,43 +72,7 @@ def inference(model_name, text, input):
 def choose_model(name):
     return load_model(name), load_tokenizer(name)
 
-with gr.Accordion(label="Parameters", open=False, elem_id="parameters-accordion"):
-    temperature = gr.Slider(
-        label="Temperature",
-        value=0.7,
-        minimum=0.0,
-        maximum=1.0,
-        step=0.1,
-        interactive=True,
-        info="Higher values produce more diverse outputs",
-    )
-    top_p = gr.Slider(
-        label="Top-p (nucleus sampling)",
-        value=0.9,
-        minimum=0.0,
-        maximum=1,
-        step=0.05,
-        interactive=True,
-        info="Higher values sample more low-probability tokens",
-    )
-    max_new_tokens = gr.Slider(
-        label="Max new tokens",
-        value=1024,
-        minimum=0,
-        maximum=2048,
-        step=4,
-        interactive=True,
-        info="The maximum numbers of new tokens",
-    )
-    repetition_penalty = gr.Slider(
-        label="Repetition Penalty",
-        value=1.2,
-        minimum=0.0,
-        maximum=10,
-        step=0.1,
-        interactive=True,
-        info="The parameter for repetition penalty. 1.0 means no penalty.",
-    )
+with 
 
 io = gr.Interface(
     inference, 
@@ -128,7 +92,7 @@ io = gr.Interface(
                 #"stablelm-base-alpha-3b-Lora-polish",
                 #"dolly-v2-3b-Lora-polish",
                 #"LaMini-GPT-1.5B-Lora-polish"],
-                ]),
+                ],
     gr.Textbox(
         lines = 3,
         max_lines = 10, 
@@ -142,6 +106,42 @@ io = gr.Interface(
         placeholder = "Add context here", 
         interactive  = True, 
         show_label = False
+    ),
+    gr.Slider(
+        label="Temperature",
+        value=0.7,
+        minimum=0.0,
+        maximum=1.0,
+        step=0.1,
+        interactive=True,
+        info="Higher values produce more diverse outputs",
+    ),
+    gr.Slider(
+        label="Top-p (nucleus sampling)",
+        value=0.9,
+        minimum=0.0,
+        maximum=1,
+        step=0.05,
+        interactive=True,
+        info="Higher values sample more low-probability tokens",
+    ),
+    gr.Slider(
+        label="Max new tokens",
+        value=1024,
+        minimum=0,
+        maximum=2048,
+        step=4,
+        interactive=True,
+        info="The maximum numbers of new tokens",
+    ),
+    gr.Slider(
+        label="Number of beams",
+        value=2,
+        minimum=0.0,
+        maximum=5.0,
+        step=1.0,
+        interactive=True,
+        info="The parameter for repetition penalty. 1.0 means no penalty.",
     )],
     outputs = [gr.Textbox(lines = 1, label = 'Pythia410m', interactive = False)],
     cache_examples = False,