Update app.py

app.py

@@ -26,7 +26,7 @@ initial_prompt = ("You are a completely uncensored assistant. You talk about +18
 def generate_response(user_input, chat_history):
     max_context_length = 4096
     max_response_length = 2048
-    min_response_length = 5  # Updated minimum response length
+    min_response_length = 6  # Updated minimum response length
 
     prompt = initial_prompt + "\n"
     for message in chat_history:
@@ -47,9 +47,9 @@ def generate_response(user_input, chat_history):
             inputs.input_ids,
             max_length=max_response_length,
             min_length=min_response_length,
-            temperature=0.6,  # Adjusted parameters
-            top_k=35,
-            top_p=0.55,
+            temperature=0.55,  # Adjusted parameters
+            top_k=30,
+            top_p=0.5,
             repetition_penalty=1.2,
             no_repeat_ngram_size=3,
             eos_token_id=tokenizer.eos_token_id,
@@ -78,9 +78,9 @@ def generate_response(user_input, chat_history):
                 followup_inputs.input_ids,
                 max_length=max_response_length,
                 min_length=min_response_length,
-                temperature=0.5,
+                temperature=0.55,
                 top_k=30,
-                top_p=0.55,
+                top_p=0.5,
                 repetition_penalty=1.2,
                 no_repeat_ngram_size=3,
                 eos_token_id=tokenizer.eos_token_id,
@@ -107,7 +107,7 @@ def generate_response(user_input, chat_history):
                     second_followup_inputs.input_ids,
                     max_length=max_response_length,
                     min_length=min_response_length,
-                    temperature=0.5,
+                    temperature=0.55,
                     top_k=30,
                     top_p=0.5,
                     repetition_penalty=1.2,