fixed generation_args in Sample inference code

* fixed temperature for "do_sample": False (`temperature=0.0` would require sampling, which is turned off with `"do_sample": False`. therefore better to use `temperature=1.0` without sampling)

README.md

@@ -133,7 +133,7 @@ pipe = pipeline(
 generation_args = {
     "max_new_tokens": 500,
     "return_full_text": False,
-    "temperature": 0.0,
+    "temperature": 1.0,
     "do_sample": False,
 }