diff --git a/lmdeploy/pytorch/backends/dlinfer/ascend/op_backend.py b/lmdeploy/pytorch/backends/dlinfer/ascend/op_backend.py
index 484cbd1b72..f528bab2ca 100644
--- a/lmdeploy/pytorch/backends/dlinfer/ascend/op_backend.py
+++ b/lmdeploy/pytorch/backends/dlinfer/ascend/op_backend.py
@@ -382,11 +382,20 @@ def get_moe_group_name(group):
                 AscendKVQuantMeta.set_value(step_context.block_offsets.device, step_context.model_config.dtype,
                                             record_file, total_layers)
 
+        cu_seqlens = None
+        has_initial_state = None
+
+        if step_context.state_offsets is not None:
+            q_start_loc = step_context.q_start_loc
+            cu_seqlens = torch.cat((q_start_loc, step_context.q_seqlens.sum().unsqueeze(0))).int()
+            if not step_context.is_decoding:
+                has_initial_state = ~(step_context.q_seqlens == step_context.kv_seqlens)
+
         attn_meta_cls = cls.get_attention_metadata_cls()
         attn_metadata = attn_meta_cls(
             step_context.is_decoding,
             step_context.block_offsets,
-            q_start_loc=None,
+            q_start_loc=cu_seqlens,
             q_seqlens=q_seqlens_cpu,
             # kv_seqlens_expanded is only expanded in paged prefill,
             # otherwise it equals kv_seqlens_cpu
@@ -399,6 +408,7 @@ def get_moe_group_name(group):
             max_kv_seq_len=max_kv_seq_len,
             quant_policy=step_context.kv_quant_policy,
             quant_meta=AscendKVQuantMeta.quant_meta,
+            has_initial_state=has_initial_state,
         )
         step_context.attn_metadata = attn_metadata
 
@@ -442,6 +452,8 @@ def init():
         """Initialize Ascend backend."""
         try:
             from torch_npu.contrib import transfer_to_npu  # noqa: F401
+            from dlinfer.vendor.ascend.triton_ops.fla.triton_utils import init_device_properties_triton
+            init_device_properties_triton()
         except ImportError:
             logger.warning('Failed to import torch_npu. Please make sure torch_npu is installed correctly. '
                            'Ascend initialization skipped.')
diff --git a/lmdeploy/pytorch/backends/dlinfer/attention.py b/lmdeploy/pytorch/backends/dlinfer/attention.py
index 8566187021..8b481730e5 100644
--- a/lmdeploy/pytorch/backends/dlinfer/attention.py
+++ b/lmdeploy/pytorch/backends/dlinfer/attention.py
@@ -18,6 +18,7 @@ class DlinferAttentionMetadata(AttentionMetadata):
     max_kv_seq_len: int = 1
     quant_meta: Dict = None
     cu_seq_lens_kv: Optional[Tensor] = None
+    has_initial_state: Optional[Tensor] = None
 
 
 class DlinferAttentionImpl(AttentionImpl[DlinferAttentionMetadata]):