fix(nvidia): lower default max_tokens to 1024 to fit nemotron-mini 4096 context length

apps/api/src/services/nvidia_provider.py

@@ -655,7 +655,7 @@ class NvidiaProvider:
         prompt: str,
         model: str | None = None,
         temperature: float = 0.1,
-        max_tokens: int = 2048,
+        max_tokens: int = 1024,
     ) -> tuple[str, bool, int, float]:
         """
         一般對話 (非 Tool Calling) - 用於 RCA 分析