[Misc] Fix logger bug (#2024)

1. Remove useless logger
2. Fix logger bug, same problem as
https://github.com/vllm-project/vllm-ascend/pull/515

- vLLM version: v0.10.0
- vLLM main:
18cc33dd60

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

vllm_ascend/distributed/kv_transfer/simple_buffer.py

@@ -22,13 +22,11 @@ import llm_datadist  # type: ignore
 import torch
 from vllm.distributed.kv_transfer.kv_lookup_buffer.base import \
     KVLookupBufferBase
-from vllm.logger import init_logger
+from vllm.logger import logger
 
 from vllm_ascend.distributed.kv_transfer.simple_pipe import SimplePipe
 from vllm_ascend.distributed.kv_transfer.utils import TORCH_DTYPE_TO_NPU_DTYPE
 
-logger = init_logger(__name__)
-
 
 # Hash a string into a int32 value.
 def int32_hash(data):

vllm_ascend/distributed/kv_transfer/simple_pipe.py

@@ -26,14 +26,12 @@ import torch_npu
 import torchair  # type: ignore
 import zmq  # type: ignore
 from vllm.distributed.kv_transfer.kv_pipe.base import KVPipeBase
-from vllm.logger import init_logger
+from vllm.logger import logger
 from vllm.utils import get_ip
 
 import vllm_ascend.envs as envs
 from vllm_ascend.distributed.kv_transfer.utils import NPU_DTYPE_TO_TORCH_DTYPE
 
-logger = init_logger(__name__)
-
 
 class SimplePipe(KVPipeBase):
 

vllm_ascend/models/pangu_moe.py

@@ -33,7 +33,7 @@ from vllm.distributed import (divide, get_pp_group,
 from vllm.distributed.parallel_state import (get_dp_group, get_ep_group,
                                              get_tp_group, get_world_group)
 from vllm.forward_context import get_forward_context
-from vllm.logger import init_logger
+from vllm.logger import logger
 from vllm.model_executor.layers.activation import SiluAndMul
 from vllm.model_executor.layers.fused_moe import FusedMoE
 from vllm.model_executor.layers.layernorm import RMSNorm
@@ -60,8 +60,6 @@ from vllm.sequence import IntermediateTensors
 from vllm_ascend.ascend_config import get_ascend_config
 from vllm_ascend.utils import ACL_FORMAT_FRACTAL_NZ, is_310p
 
-logger = init_logger(__name__)
-
 _ROUTER_SCALE = None
 
 

vllm_ascend/multistream/decorator.py

@@ -1,10 +1,6 @@
-from vllm.logger import init_logger
-
 from .context import (get_multistream_layer_context,
                       get_multistream_microbatch_context)
 
-logger = init_logger(__name__)
-
 
 # vllm v1 use get_forward_context to get the attn_metadata,
 # we can use this decorator to update the attn metadata

vllm_ascend/sample/rejection_sampler.py

@@ -4,14 +4,11 @@ from typing import Optional
 import torch
 import torch.nn as nn
 import vllm.v1.sample.rejection_sampler as rs
-from vllm.logger import init_logger
 from vllm.v1.sample.metadata import SamplingMetadata
 from vllm.v1.sample.rejection_sampler import (RejectionSampler, compute_probs,
                                               generate_uniform_probs)
 from vllm.v1.spec_decode.metadata import SpecDecodeMetadata
 
-logger = init_logger(__name__)
-
 PLACEHOLDER_TOKEN_ID = -1
 GREEDY_TEMPERATURE = -1
 # Maximum number of speculative draft tokens allowed per request in a single

vllm_ascend/worker/eagle_proposer_v1.py

@@ -7,7 +7,7 @@ from vllm.attention.layer import Attention
 from vllm.config import (CompilationLevel, VllmConfig,
                          get_layers_from_vllm_config)
 from vllm.distributed.parallel_state import get_pp_group
-from vllm.logger import init_logger
+from vllm.logger import logger
 from vllm.model_executor.model_loader import get_model
 from vllm.model_executor.models import supports_multimodal
 from vllm.model_executor.models.llama_eagle3 import Eagle3LlamaForCausalLM
@@ -17,8 +17,6 @@ from vllm_ascend.ascend_forward_context import set_ascend_forward_context
 from vllm_ascend.attention.attention_mask import AttentionMaskBuilder
 from vllm_ascend.attention.attention_v1 import AscendAttentionState
 
-logger = init_logger(__name__)
-
 PADDING_SLOT_ID = -1