[v1] Support allowed_token_ids in v1 Sampler (#13210)

Signed-off-by: Lu Fang <lufang@fb.com>

vllm/v1/sample/metadata.py

@@ -37,3 +37,7 @@ class SamplingMetadata:
     min_tokens: Dict[int, Tuple[int, Set[int]]]
 
     logit_bias: List[Optional[Dict[int, float]]]
+
+    # `allowed_token_ids_mask` is a 2D bool tensor of shape (max batch size,
+    # vocab size).
+    allowed_token_ids_mask: Optional[torch.Tensor]