fix

Signed-off-by: wxsIcey <[email protected]>

Author: wxsIcey

vllm_ascend/ops/rotary_embedding.py

@@ -20,7 +20,6 @@
 
 import torch
 import torch_npu
-from vllm.forward_context import get_forward_context
 from vllm.model_executor.layers.rotary_embedding import (
     DeepseekScalingRotaryEmbedding, MRotaryEmbedding, RotaryEmbedding,
     YaRNScalingRotaryEmbedding)

vllm_ascend/ops/triton/linearnorm/split_qkv_rmsnorm_rope.py

@@ -17,9 +17,8 @@
 from typing import Optional
 
 import torch
-import triton
+import triton  # type: ignore
 import triton.language as tl  # type: ignore
-import triton.runtime.driver as driver
 from vllm.utils.torch_utils import direct_register_custom_op
 
 from vllm_ascend.ops.triton.triton_utils import get_vectorcore_num