Add sampling defaults for generate

llama_cpp/llama.py

@@ -446,10 +446,10 @@ class Llama:
     def generate(
         self,
         tokens: Sequence[llama_cpp.llama_token],
-        top_k: int,
-        top_p: float,
-        temp: float,
-        repeat_penalty: float,
+        top_k: int = 40,
+        top_p: float = 0.95,
+        temp: float = 0.80,
+        repeat_penalty: float = 1.1,
         reset: bool = True,
         frequency_penalty: float = 0.0,
         presence_penalty: float = 0.0,