ZeroGPU: raise GPU budget 120->180s, cap max_new_tokens 512->256

generate.py

@@ -80,8 +80,8 @@ def _render(messages, tok) -> str:
         messages, tokenize=False, add_generation_prompt=True)
 
 
-@GPU(duration=120)
-def generate(messages: list[dict], max_new_tokens: int = 512,
+@GPU(duration=180)
+def generate(messages: list[dict], max_new_tokens: int = 256,
              temperature: float = 0.2) -> str:
     """Generate an assistant reply for chat-format ``messages``."""
     if STUB: