[V1][Mamba] - Enable V1 by default for Mamba Models

docs/usage/v1_guide.md

@@ -111,11 +111,10 @@ Models that use Mamba-2 and Mamba-1 layers (e.g., `Mamba2ForCausalLM`, `MambaFor
 
 Models that combine Mamba-2 and Mamba-1 layers with standard attention layers are also supported (e.g., `BambaForCausalLM`,
 `Zamba2ForCausalLM`, `NemotronHForCausalLM`, `FalconH1ForCausalLM` and `GraniteMoeHybridForCausalLM`, `JambaForCausalLM`). Please note that
-these models currently require disabling prefix caching and using the FlashInfer attention backend in V1.
+these models currently require disabling prefix caching in V1.
 
 Hybrid models with mechanisms different to Mamba are also supported (e.g, `MiniMaxText01ForCausalLM`, `MiniMaxM1ForCausalLM`).
-Please note that these models currently require disabling prefix caching, enforcing eager mode, and using the FlashInfer
-attention backend in V1.
+Please note that these models currently require disabling prefix caching and enforcing eager mode in V1.
 
 #### Encoder-Decoder Models
 

vllm/engine/arg_utils.py

@@ -1464,11 +1464,6 @@ def _is_v1_supported_oracle(self, model_config: ModelConfig) -> bool:
                                recommend_to_remove=False)
             return False
 
-        # V1 mamba models are unoptimized.
-        if model_config.has_inner_state and _warn_or_fallback(
-                feature_name="Mamba"):
-            return False
-
         # No Concurrent Partial Prefills so far.
         if (self.max_num_partial_prefills
                 != SchedulerConfig.max_num_partial_prefills