feat: remove vllm distributed (#2907)

Co-authored-by: Zhangyi <1109276519@qq.com>

python/sglang/srt/models/qwen2_vl.py

@@ -30,12 +30,12 @@ import torch
 import torch.nn as nn
 import torch.nn.functional as F
 from einops import rearrange, repeat
-from vllm.distributed import parallel_state
-from vllm.distributed import utils as dist_utils
 from vllm.logger import init_logger
 from vllm.model_executor.layers.activation import QuickGELU
 
 from sglang.srt.configs import Qwen2VLConfig, Qwen2VLVisionConfig
+from sglang.srt.distributed import parallel_state
+from sglang.srt.distributed import utils as dist_utils
 from sglang.srt.hf_transformers_utils import get_processor
 from sglang.srt.layers.attention.triton_ops.prefill_attention import (
     context_attention_fwd,

python/sglang/srt/models/stablelm.py

@@ -24,9 +24,9 @@ from typing import Iterable, Optional, Tuple
 import torch
 from torch import nn
 from transformers import PretrainedConfig
-from vllm.distributed import get_tensor_model_parallel_world_size
 from vllm.model_executor.layers.rotary_embedding import get_rope
 
+from sglang.srt.distributed import get_tensor_model_parallel_world_size
 from sglang.srt.layers.activation import SiluAndMul
 from sglang.srt.layers.linear import (
     MergedColumnParallelLinear,

python/sglang/srt/models/torch_native_llama.py

@@ -47,12 +47,12 @@ import torch
 from torch import nn
 from torch.nn.parameter import Parameter
 from transformers import LlamaConfig
-from vllm.distributed import (
+from vllm.model_executor.layers.rotary_embedding import get_rope
+
+from sglang.srt.distributed import (
     get_tensor_model_parallel_rank,
     get_tensor_model_parallel_world_size,
 )
-from vllm.model_executor.layers.rotary_embedding import get_rope
-
 from sglang.srt.layers.activation import SiluAndMul
 from sglang.srt.layers.layernorm import RMSNorm
 from sglang.srt.layers.logits_processor import LogitsProcessor, LogitsProcessorOutput

python/sglang/srt/models/xverse.py

@@ -21,7 +21,6 @@ from typing import Any, Dict, Iterable, Optional, Tuple
 import torch
 from torch import nn
 from transformers import LlamaConfig
-from vllm.distributed import get_tensor_model_parallel_world_size
 from vllm.model_executor.layers.activation import SiluAndMul
 from vllm.model_executor.layers.layernorm import RMSNorm
 from vllm.model_executor.layers.linear import (
@@ -31,6 +30,7 @@ from vllm.model_executor.layers.linear import (
 )
 from vllm.model_executor.layers.rotary_embedding import get_rope
 
+from sglang.srt.distributed import get_tensor_model_parallel_world_size
 from sglang.srt.layers.logits_processor import LogitsProcessor
 from sglang.srt.layers.quantization.base_config import QuantizationConfig
 from sglang.srt.layers.radix_attention import RadixAttention

python/sglang/srt/models/xverse_moe.py

@@ -18,11 +18,6 @@ from typing import Any, Dict, Iterable, Optional, Tuple
 import torch
 from torch import nn
 from transformers import PretrainedConfig
-from vllm.distributed import (
-    get_tensor_model_parallel_rank,
-    get_tensor_model_parallel_world_size,
-    tensor_model_parallel_all_reduce,
-)
 from vllm.model_executor.layers.activation import SiluAndMul
 from vllm.model_executor.layers.layernorm import RMSNorm
 from vllm.model_executor.layers.linear import (
@@ -33,6 +28,11 @@ from vllm.model_executor.layers.linear import (
 )
 from vllm.model_executor.layers.rotary_embedding import get_rope
 
+from sglang.srt.distributed import (
+    get_tensor_model_parallel_rank,
+    get_tensor_model_parallel_world_size,
+    tensor_model_parallel_all_reduce,
+)
 from sglang.srt.layers.logits_processor import LogitsProcessor
 from sglang.srt.layers.moe.fused_moe_triton import fused_moe
 from sglang.srt.layers.quantization.base_config import QuantizationConfig