[Chore] Adjust tokenizer import to avoid circular imports (#30601)

Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>

vllm/model_executor/models/pixtral.py

@@ -59,7 +59,8 @@ from vllm.multimodal.processing import (
 from vllm.multimodal.profiling import BaseDummyInputsBuilder, ProcessorInputs
 from vllm.platforms import current_platform
 from vllm.sequence import IntermediateTensors
-from vllm.tokenizers import MistralTokenizer, cached_tokenizer_from_config
+from vllm.tokenizers import cached_tokenizer_from_config
+from vllm.tokenizers.mistral import MistralTokenizer
 from vllm.utils.tensor_schema import TensorSchema, TensorShape
 
 from .interfaces import MultiModalEmbeddings, SupportsMultiModal, SupportsPP