Drop 0.12.0 support (#5146)

We decided to release v0.13.0 soon. So no need to support 0.12.0 now.
Let's drop it.

- vLLM version: v0.12.0
- vLLM main:
ad32e3e19c

Signed-off-by: wangxiyuan <wangxiyuan1007@gmail.com>

vllm_ascend/worker/npu_input_batch.py

@@ -22,6 +22,7 @@ import torch
 from vllm.lora.request import LoRARequest
 from vllm.pooling_params import PoolingParams
 from vllm.v1.outputs import LogprobsTensors
+from vllm.v1.pool.metadata import PoolingStates
 from vllm.v1.sample.logits_processor import (BatchUpdateBuilder,
                                              LogitsProcessors)
 from vllm.v1.worker.gpu_input_batch import InputBatch
@@ -29,16 +30,6 @@ from vllm.v1.worker.gpu_input_batch import InputBatch
 from vllm_ascend.worker.block_table import MultiGroupBlockTable
 
 
-class PoolingStates:
-    # NOTE: This should be removed after we drop support of vLLM v0.12.0
-    def __init__(self):
-        # for chunked prefill with ALL pooling
-        self.hidden_states_cache: list[torch.Tensor] = []
-
-    def clean(self):
-        self.hidden_states_cache.clear()
-
-
 class NPUInputBatch(InputBatch):
 
     def __init__(