Address code rabbit comment

Signed-off-by: Jin Li <[email protected]>

Author: liji-nv

tensorrt_llm/_torch/custom_ops/trtllm_gen_custom_ops.py

@@ -540,7 +540,7 @@ def fp8_block_scale_moe_runner(
 
 @fp8_block_scale_moe_runner.register_fake
 def _(
-    rrouting_logits: torch.Tensor,
+    routing_logits: torch.Tensor,
     routing_bias: torch.Tensor,
     hidden_states: torch.Tensor,
     hidden_states_scale: torch.Tensor,

tensorrt_llm/llmapi/llm_args.py

@@ -1797,6 +1797,15 @@ class TorchCompileConfig(BaseModel):
         description=
         "The maximum number of CUDA streams to use for torch.compile.")
 
+    @field_validator('max_num_streams')
+    @classmethod
+    def validate_torch_compile_max_num_streams(cls, v):
+        """Validate torch_compile_config.max_num_streams >= 1."""
+        if v < 1:
+            raise ValueError(
+                "torch_compile_config.max_num_streams must be >= 1")
+        return v
+
 
 class TorchLlmArgs(BaseLlmArgs):
     # Just a dummy BuildConfig to allow code reuse with the TrtLlmArgs