feat: remove vllm get_rope (#2964)

python/sglang/srt/models/olmo2.py

@@ -21,7 +21,6 @@ from typing import Iterable, Optional, Tuple
 import torch
 from torch import nn
 from transformers import PretrainedConfig
-from vllm.model_executor.layers.rotary_embedding import get_rope
 
 from sglang.srt.distributed import (
     get_tensor_model_parallel_rank,
@@ -39,6 +38,7 @@ from sglang.srt.layers.linear import (
 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
+from sglang.srt.layers.rotary_embedding import get_rope
 from sglang.srt.layers.vocab_parallel_embedding import (
     ParallelLMHead,
     VocabParallelEmbedding,

python/sglang/srt/models/olmoe.py

@@ -22,7 +22,6 @@ from typing import Any, Dict, Iterable, Optional, Tuple
 import torch
 from torch import nn
 from transformers import PretrainedConfig
-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.layernorm import RMSNorm
@@ -35,6 +34,7 @@ from sglang.srt.layers.logits_processor import LogitsProcessor
 from sglang.srt.layers.moe.fused_moe_triton import FusedMoE
 from sglang.srt.layers.quantization.base_config import QuantizationConfig
 from sglang.srt.layers.radix_attention import RadixAttention
+from sglang.srt.layers.rotary_embedding import get_rope
 from sglang.srt.layers.vocab_parallel_embedding import (
     ParallelLMHead,
     VocabParallelEmbedding,

python/sglang/srt/models/phi3_small.py

@@ -5,7 +5,6 @@ import torch
 from torch import nn
 from transformers import Phi3Config
 from transformers.configuration_utils import PretrainedConfig
-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.linear import (
@@ -17,6 +16,7 @@ from sglang.srt.layers.logits_processor import LogitsProcessor, LogitsProcessorO
 from sglang.srt.layers.pooler import Pooler, PoolingType
 from sglang.srt.layers.quantization.base_config import QuantizationConfig
 from sglang.srt.layers.radix_attention import RadixAttention
+from sglang.srt.layers.rotary_embedding import get_rope
 from sglang.srt.layers.vocab_parallel_embedding import (
     DEFAULT_VOCAB_PADDING_SIZE,
     ParallelLMHead,

python/sglang/srt/models/qwen.py

@@ -20,7 +20,6 @@ from typing import Any, Dict, Iterable, Optional, Tuple
 import torch
 from torch import nn
 from transformers import PretrainedConfig
-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
@@ -33,6 +32,7 @@ from sglang.srt.layers.linear import (
 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
+from sglang.srt.layers.rotary_embedding import get_rope
 from sglang.srt.layers.vocab_parallel_embedding import (
     ParallelLMHead,
     VocabParallelEmbedding,

python/sglang/srt/models/qwen2.py

@@ -20,7 +20,6 @@ from typing import Any, Dict, Iterable, Optional, Tuple
 
 import torch
 from torch import nn
-from vllm.model_executor.layers.rotary_embedding import get_rope
 
 from sglang.srt.distributed import (
     get_tensor_model_parallel_rank,
@@ -37,6 +36,7 @@ from sglang.srt.layers.logits_processor import LogitsProcessor
 from sglang.srt.layers.pooler import Pooler, PoolingType
 from sglang.srt.layers.quantization.base_config import QuantizationConfig
 from sglang.srt.layers.radix_attention import RadixAttention
+from sglang.srt.layers.rotary_embedding import get_rope
 from sglang.srt.layers.vocab_parallel_embedding import (
     ParallelLMHead,
     VocabParallelEmbedding,

python/sglang/srt/models/qwen2_moe.py

@@ -22,7 +22,6 @@ import torch
 import torch.nn.functional as F
 from torch import nn
 from transformers import PretrainedConfig
-from vllm.model_executor.layers.rotary_embedding import get_rope
 
 from sglang.srt.distributed import (
     get_tensor_model_parallel_world_size,
@@ -40,6 +39,7 @@ from sglang.srt.layers.logits_processor import LogitsProcessor
 from sglang.srt.layers.moe.fused_moe_triton import FusedMoE
 from sglang.srt.layers.quantization.base_config import QuantizationConfig
 from sglang.srt.layers.radix_attention import RadixAttention
+from sglang.srt.layers.rotary_embedding import get_rope
 from sglang.srt.layers.vocab_parallel_embedding import (
     ParallelLMHead,
     VocabParallelEmbedding,

python/sglang/srt/models/stablelm.py

@@ -24,7 +24,6 @@ from typing import Iterable, Optional, Tuple
 import torch
 from torch import nn
 from transformers import PretrainedConfig
-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
@@ -36,6 +35,7 @@ from sglang.srt.layers.linear import (
 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
+from sglang.srt.layers.rotary_embedding import get_rope
 from sglang.srt.layers.vocab_parallel_embedding import (
     ParallelLMHead,
     VocabParallelEmbedding,

python/sglang/srt/models/torch_native_llama.py

@@ -47,7 +47,6 @@ import torch
 from torch import nn
 from torch.nn.parameter import Parameter
 from transformers import LlamaConfig
-from vllm.model_executor.layers.rotary_embedding import get_rope
 
 from sglang.srt.distributed import (
     get_tensor_model_parallel_rank,
@@ -58,6 +57,7 @@ from sglang.srt.layers.layernorm import RMSNorm
 from sglang.srt.layers.logits_processor import LogitsProcessor, LogitsProcessorOutput
 from sglang.srt.layers.quantization.base_config import QuantizationConfig
 from sglang.srt.layers.radix_attention import RadixAttention
+from sglang.srt.layers.rotary_embedding import get_rope
 from sglang.srt.layers.vocab_parallel_embedding import (
     ParallelLMHead,
     VocabParallelEmbedding,

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.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
@@ -34,6 +33,7 @@ from sglang.srt.layers.linear import (
 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
+from sglang.srt.layers.rotary_embedding import get_rope
 from sglang.srt.layers.vocab_parallel_embedding import (
     ParallelLMHead,
     VocabParallelEmbedding,

python/sglang/srt/models/xverse_moe.py

@@ -18,7 +18,6 @@ from typing import Any, Dict, Iterable, Optional, Tuple
 import torch
 from torch import nn
 from transformers import PretrainedConfig
-from vllm.model_executor.layers.rotary_embedding import get_rope
 
 from sglang.srt.distributed import (
     get_tensor_model_parallel_rank,
@@ -37,6 +36,7 @@ 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
 from sglang.srt.layers.radix_attention import RadixAttention
+from sglang.srt.layers.rotary_embedding import get_rope
 from sglang.srt.layers.vocab_parallel_embedding import (
     ParallelLMHead,
     VocabParallelEmbedding,