lazy import attn backends (#4200)

python/sglang/srt/layers/attention/triton_backend.py

@@ -6,9 +6,7 @@ import torch
 import triton
 
 from sglang.srt.layers.attention.base_attn_backend import AttentionBackend
-from sglang.srt.layers.attention.flashinfer_backend import (
-    create_flashinfer_kv_indices_triton,
-)
+from sglang.srt.layers.attention.utils import create_flashinfer_kv_indices_triton
 from sglang.srt.layers.dp_attention import get_attention_tp_size
 from sglang.srt.model_executor.forward_batch_info import ForwardBatch, ForwardMode
 

python/sglang/srt/model_executor/cuda_graph_runner.py

@@ -302,7 +302,7 @@ class CudaGraphRunner:
             self.stream = graph_capture_context.stream
             # Reverse the order to enable better memory sharing across cuda graphs.
             capture_range = (
-                tqdm.tqdm(reversed(self.capture_bs))
+                tqdm.tqdm(list(reversed(self.capture_bs)))
                 if get_tensor_model_parallel_rank() == 0
                 else reversed(self.capture_bs)
             )

python/sglang/srt/model_executor/model_runner.py

@@ -35,11 +35,6 @@ from sglang.srt.distributed import (
     set_custom_all_reduce,
 )
 from sglang.srt.distributed.parallel_state import monkey_patch_vllm_parallel_state
-from sglang.srt.layers.attention.double_sparsity_backend import DoubleSparseAttnBackend
-from sglang.srt.layers.attention.flashinfer_backend import FlashInferAttnBackend
-from sglang.srt.layers.attention.flashinfer_mla_backend import FlashInferMLAAttnBackend
-from sglang.srt.layers.attention.torch_native_backend import TorchNativeAttnBackend
-from sglang.srt.layers.attention.triton_backend import TritonAttnBackend
 from sglang.srt.layers.dp_attention import (
     get_attention_tp_group,
     get_attention_tp_size,
@@ -77,7 +72,6 @@ from sglang.srt.utils import (
     set_cpu_offload_max_bytes,
     set_cuda_arch,
 )
-from sglang.utils import get_exception_traceback
 
 logger = logging.getLogger(__name__)
 
@@ -779,6 +773,10 @@ class ModelRunner:
     def init_attention_backend(self):
         """Init attention kernel backend."""
         if self.server_args.attention_backend == "flashinfer":
+            from sglang.srt.layers.attention.flashinfer_backend import (
+                FlashInferAttnBackend,
+            )
+
             # Init streams
             if self.server_args.speculative_algorithm == "EAGLE":
                 self.plan_stream_for_flashinfer = torch.cuda.Stream()
@@ -794,12 +792,26 @@ class ModelRunner:
                 "Please use `--attention-backend flashinfer`."
             )
             if self.server_args.enable_double_sparsity:
+                from sglang.srt.layers.attention.double_sparsity_backend import (
+                    DoubleSparseAttnBackend,
+                )
+
                 self.attn_backend = DoubleSparseAttnBackend(self)
             else:
+                from sglang.srt.layers.attention.triton_backend import TritonAttnBackend
+
                 self.attn_backend = TritonAttnBackend(self)
         elif self.server_args.attention_backend == "torch_native":
+            from sglang.srt.layers.attention.torch_native_backend import (
+                TorchNativeAttnBackend,
+            )
+
             self.attn_backend = TorchNativeAttnBackend(self)
         elif self.server_args.attention_backend == "flashinfer_mla":
+            from sglang.srt.layers.attention.flashinfer_mla_backend import (
+                FlashInferMLAAttnBackend,
+            )
+
             self.attn_backend = FlashInferMLAAttnBackend(self)
         else:
             raise ValueError(