[5/N][Attention] Finish eliminating vllm/attention folder (#32064)
Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>
This commit is contained in:
@@ -9,7 +9,6 @@ from itertools import islice
|
||||
import torch
|
||||
from torch import nn
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig, get_current_vllm_config
|
||||
from vllm.distributed import (
|
||||
@@ -18,6 +17,7 @@ from vllm.distributed import (
|
||||
get_tensor_model_parallel_world_size,
|
||||
)
|
||||
from vllm.logger import init_logger
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.fused_moe.shared_fused_moe import SharedFusedMoE
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import (
|
||||
|
||||
@@ -11,7 +11,7 @@ import torch.nn as nn
|
||||
from vllm.distributed import get_tensor_model_parallel_world_size
|
||||
from vllm.distributed.utils import divide
|
||||
from vllm.model_executor.layers.activation import SiluAndMul
|
||||
from vllm.model_executor.layers.attention.mm_encoder_attention import MMEncoderAttention
|
||||
from vllm.model_executor.layers.attention import MMEncoderAttention
|
||||
from vllm.model_executor.layers.conv import Conv2dLayer
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import (
|
||||
|
||||
@@ -32,12 +32,12 @@ import torch
|
||||
from torch import nn
|
||||
from transformers import ApertusConfig
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
|
||||
from vllm.model_executor.layers.activation import XIELU
|
||||
from vllm.model_executor.layers.attention.encoder_only_attention import (
|
||||
from vllm.model_executor.layers.attention import (
|
||||
Attention,
|
||||
EncoderOnlyAttention,
|
||||
)
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
|
||||
@@ -8,7 +8,6 @@ from itertools import islice
|
||||
import torch
|
||||
from torch import nn
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.distributed import (
|
||||
@@ -19,6 +18,7 @@ from vllm.distributed import (
|
||||
)
|
||||
from vllm.logger import init_logger
|
||||
from vllm.model_executor.layers.activation import SiluAndMul
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.fused_moe import fused_experts, fused_topk
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import (
|
||||
|
||||
@@ -29,7 +29,6 @@ import torch
|
||||
from torch import nn
|
||||
from transformers import PretrainedConfig
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.distributed import (
|
||||
@@ -38,6 +37,7 @@ from vllm.distributed import (
|
||||
get_tensor_model_parallel_world_size,
|
||||
)
|
||||
from vllm.model_executor.layers.activation import SiluAndMul
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import (
|
||||
MergedColumnParallelLinear,
|
||||
|
||||
@@ -32,7 +32,6 @@ import torch.nn.functional as F
|
||||
from torch import nn
|
||||
from transformers.configuration_utils import PretrainedConfig
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.distributed import (
|
||||
@@ -41,6 +40,7 @@ from vllm.distributed import (
|
||||
get_tensor_model_parallel_world_size,
|
||||
)
|
||||
from vllm.model_executor.layers.activation import SiluAndMul
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.fused_moe import SharedFusedMoE
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import (
|
||||
|
||||
@@ -9,12 +9,12 @@ import torch
|
||||
from torch import nn
|
||||
from transformers import BambaConfig
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, ModelConfig, VllmConfig
|
||||
from vllm.distributed import get_tensor_model_parallel_world_size
|
||||
from vllm.distributed.parallel_state import get_pp_group
|
||||
from vllm.model_executor.layers.activation import SiluAndMul
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import (
|
||||
MergedColumnParallelLinear,
|
||||
|
||||
@@ -11,7 +11,7 @@ from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, ModelConfig, PoolerConfig, VllmConfig
|
||||
from vllm.distributed import get_tensor_model_parallel_world_size
|
||||
from vllm.model_executor.layers.activation import get_act_fn
|
||||
from vllm.model_executor.layers.attention.encoder_only_attention import (
|
||||
from vllm.model_executor.layers.attention import (
|
||||
EncoderOnlyAttention,
|
||||
)
|
||||
from vllm.model_executor.layers.linear import (
|
||||
|
||||
@@ -15,7 +15,7 @@ from vllm.distributed import (
|
||||
tensor_model_parallel_all_reduce,
|
||||
)
|
||||
from vllm.model_executor.layers.activation import get_act_and_mul_fn, get_act_fn
|
||||
from vllm.model_executor.layers.attention.encoder_only_attention import (
|
||||
from vllm.model_executor.layers.attention import (
|
||||
EncoderOnlyAttention,
|
||||
)
|
||||
from vllm.model_executor.layers.fused_moe import activation_without_mul, fused_topk
|
||||
|
||||
@@ -11,7 +11,7 @@ from transformers import Blip2VisionConfig, BlipVisionConfig
|
||||
|
||||
from vllm.distributed import divide, get_tensor_model_parallel_world_size
|
||||
from vllm.model_executor.layers.activation import get_act_fn
|
||||
from vllm.model_executor.layers.attention.mm_encoder_attention import MMEncoderAttention
|
||||
from vllm.model_executor.layers.attention import MMEncoderAttention
|
||||
from vllm.model_executor.layers.conv import Conv2dLayer
|
||||
from vllm.model_executor.layers.linear import (
|
||||
ColumnParallelLinear,
|
||||
|
||||
@@ -27,7 +27,6 @@ import torch
|
||||
from torch import nn
|
||||
from transformers import BloomConfig
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.distributed import (
|
||||
@@ -36,6 +35,7 @@ from vllm.distributed import (
|
||||
get_tensor_model_parallel_world_size,
|
||||
)
|
||||
from vllm.model_executor.layers.activation import get_act_fn
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.linear import (
|
||||
ColumnParallelLinear,
|
||||
QKVParallelLinear,
|
||||
|
||||
@@ -16,12 +16,12 @@ from transformers import (
|
||||
ChameleonVQVAEConfig,
|
||||
)
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.config.multimodal import BaseDummyOptions
|
||||
from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
|
||||
from vllm.logger import init_logger
|
||||
from vllm.model_executor.layers.activation import SiluAndMul
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.conv import Conv2dLayer
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import (
|
||||
|
||||
@@ -12,11 +12,11 @@ import torch
|
||||
from torch import nn
|
||||
from torch.nn import LayerNorm
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
|
||||
from vllm.model_executor.layers.activation import SiluAndMul
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import (
|
||||
MergedColumnParallelLinear,
|
||||
|
||||
@@ -14,12 +14,11 @@ from transformers import (
|
||||
CLIPVisionConfig,
|
||||
)
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.config import VllmConfig
|
||||
from vllm.config.multimodal import BaseDummyOptions
|
||||
from vllm.distributed import divide, get_tensor_model_parallel_world_size
|
||||
from vllm.model_executor.layers.activation import get_act_fn
|
||||
from vllm.model_executor.layers.attention.mm_encoder_attention import MMEncoderAttention
|
||||
from vllm.model_executor.layers.attention import Attention, MMEncoderAttention
|
||||
from vllm.model_executor.layers.conv import Conv2dLayer
|
||||
from vllm.model_executor.layers.linear import (
|
||||
ColumnParallelLinear,
|
||||
|
||||
@@ -30,11 +30,11 @@ import torch
|
||||
from torch import nn
|
||||
from transformers import Cohere2Config, CohereConfig
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
|
||||
from vllm.model_executor.layers.activation import SiluAndMul
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.linear import (
|
||||
MergedColumnParallelLinear,
|
||||
QKVParallelLinear,
|
||||
|
||||
@@ -8,13 +8,13 @@ import torch
|
||||
import torch.nn as nn
|
||||
from transformers import DbrxConfig
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.distributed import (
|
||||
get_pp_group,
|
||||
get_tensor_model_parallel_rank,
|
||||
get_tensor_model_parallel_world_size,
|
||||
)
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.fused_moe import FusedMoE
|
||||
from vllm.model_executor.layers.linear import (
|
||||
QKVParallelLinear,
|
||||
|
||||
@@ -18,7 +18,7 @@ import torch.nn as nn
|
||||
import torch.nn.functional as F
|
||||
from transformers import CLIPVisionConfig
|
||||
|
||||
from vllm.model_executor.layers.attention.mm_encoder_attention import MMEncoderAttention
|
||||
from vllm.model_executor.layers.attention import MMEncoderAttention
|
||||
from vllm.model_executor.layers.conv import Conv2dLayer
|
||||
from vllm.model_executor.layers.quantization import QuantizationConfig
|
||||
from vllm.model_executor.model_loader.weight_utils import default_weight_loader
|
||||
|
||||
@@ -33,7 +33,6 @@ from torch import nn
|
||||
from transformers import DeepseekV2Config, DeepseekV3Config
|
||||
|
||||
from vllm._aiter_ops import rocm_aiter_ops
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, ParallelConfig, VllmConfig, get_current_vllm_config
|
||||
from vllm.distributed import (
|
||||
@@ -45,6 +44,7 @@ from vllm.distributed import (
|
||||
)
|
||||
from vllm.logger import init_logger
|
||||
from vllm.model_executor.layers.activation import SiluAndMul
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.attention_layer_base import AttentionLayerBase
|
||||
from vllm.model_executor.layers.fused_moe import SharedFusedMoE
|
||||
from vllm.model_executor.layers.layernorm import LayerNorm, RMSNorm
|
||||
|
||||
@@ -32,7 +32,6 @@ import torch
|
||||
from torch import nn
|
||||
from transformers import Dots1Config
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, ModelConfig, VllmConfig
|
||||
from vllm.distributed import (
|
||||
@@ -41,6 +40,7 @@ from vllm.distributed import (
|
||||
tensor_model_parallel_all_reduce,
|
||||
)
|
||||
from vllm.model_executor.layers.activation import SiluAndMul
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.fused_moe import SharedFusedMoE
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import (
|
||||
|
||||
@@ -16,7 +16,7 @@ from vllm.distributed.parallel_state import (
|
||||
get_tensor_model_parallel_world_size,
|
||||
)
|
||||
from vllm.model_executor.layers.activation import SiluAndMul
|
||||
from vllm.model_executor.layers.attention.mm_encoder_attention import (
|
||||
from vllm.model_executor.layers.attention import (
|
||||
MMEncoderAttention,
|
||||
)
|
||||
from vllm.model_executor.layers.conv import Conv2dLayer
|
||||
|
||||
@@ -32,7 +32,6 @@ import torch
|
||||
from torch import nn
|
||||
from transformers import PretrainedConfig
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig, get_current_vllm_config
|
||||
from vllm.distributed import (
|
||||
@@ -42,6 +41,7 @@ from vllm.distributed import (
|
||||
)
|
||||
from vllm.logger import init_logger
|
||||
from vllm.model_executor.layers.activation import SiluAndMul
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.fused_moe import SharedFusedMoE
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import (
|
||||
|
||||
@@ -42,7 +42,7 @@ 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 vllm.model_executor.layers.attention.mm_encoder_attention import (
|
||||
from vllm.model_executor.layers.attention import (
|
||||
MMEncoderAttention,
|
||||
)
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
|
||||
@@ -31,12 +31,11 @@ import torch
|
||||
from torch import nn
|
||||
from transformers import PretrainedConfig
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
|
||||
# from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
|
||||
from vllm.logger import init_logger
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.fused_moe import SharedFusedMoE
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import (
|
||||
|
||||
@@ -32,11 +32,11 @@ import torch
|
||||
from torch import nn
|
||||
from transformers import PretrainedConfig
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
|
||||
from vllm.model_executor.layers.activation import SiluAndMul
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import (
|
||||
MergedColumnParallelLinear,
|
||||
|
||||
@@ -28,11 +28,11 @@ import torch
|
||||
from torch import nn
|
||||
from transformers import Exaone4Config
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
|
||||
from vllm.model_executor.layers.activation import SiluAndMul
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import (
|
||||
MergedColumnParallelLinear,
|
||||
|
||||
@@ -30,7 +30,6 @@ from torch import nn
|
||||
from torch.nn import LayerNorm
|
||||
from transformers import FalconConfig as HF_FalconConfig
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.distributed import (
|
||||
@@ -40,6 +39,7 @@ from vllm.distributed import (
|
||||
tensor_model_parallel_all_reduce,
|
||||
)
|
||||
from vllm.model_executor.layers.activation import get_act_fn
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.linear import (
|
||||
ColumnParallelLinear,
|
||||
QKVParallelLinear,
|
||||
|
||||
@@ -9,12 +9,12 @@ import torch
|
||||
from torch import nn
|
||||
from transformers import FalconH1Config
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, ModelConfig, VllmConfig
|
||||
from vllm.distributed import get_tensor_model_parallel_world_size
|
||||
from vllm.distributed.parallel_state import get_pp_group
|
||||
from vllm.model_executor.layers.activation import SiluAndMul
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import (
|
||||
MergedColumnParallelLinear,
|
||||
|
||||
@@ -26,12 +26,12 @@ import torch
|
||||
from torch import nn
|
||||
from transformers import GemmaConfig
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
|
||||
from vllm.logger import init_logger
|
||||
from vllm.model_executor.layers.activation import GeluAndMul
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.layernorm import GemmaRMSNorm
|
||||
from vllm.model_executor.layers.linear import (
|
||||
MergedColumnParallelLinear,
|
||||
|
||||
@@ -23,12 +23,12 @@ import torch
|
||||
from torch import nn
|
||||
from transformers import Gemma2Config
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
|
||||
from vllm.logger import init_logger
|
||||
from vllm.model_executor.layers.activation import GeluAndMul
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.layernorm import GemmaRMSNorm
|
||||
from vllm.model_executor.layers.linear import (
|
||||
MergedColumnParallelLinear,
|
||||
|
||||
@@ -22,13 +22,13 @@ import torch
|
||||
from torch import nn
|
||||
from transformers import Gemma3TextConfig
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
|
||||
from vllm.logger import init_logger
|
||||
from vllm.model_executor.layers.activation import GeluAndMul
|
||||
from vllm.model_executor.layers.attention.encoder_only_attention import (
|
||||
from vllm.model_executor.layers.attention import (
|
||||
Attention,
|
||||
EncoderOnlyAttention,
|
||||
)
|
||||
from vllm.model_executor.layers.layernorm import GemmaRMSNorm
|
||||
|
||||
@@ -21,7 +21,6 @@ import torch
|
||||
from torch import nn
|
||||
from transformers.models.gemma3n.configuration_gemma3n import Gemma3nTextConfig
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.distributed import get_tensor_model_parallel_world_size
|
||||
@@ -32,6 +31,7 @@ from vllm.model_executor.layers.activation import (
|
||||
GeluAndMul,
|
||||
GeluAndMulSparse,
|
||||
)
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import (
|
||||
ColumnParallelLinear,
|
||||
|
||||
@@ -29,10 +29,10 @@ import torch
|
||||
from torch import nn
|
||||
from transformers import Glm4Config
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import QKVParallelLinear, RowParallelLinear
|
||||
from vllm.model_executor.layers.logits_processor import LogitsProcessor
|
||||
|
||||
@@ -52,7 +52,7 @@ from vllm.config.multimodal import BaseDummyOptions, VideoDummyOptions
|
||||
from vllm.distributed import get_tensor_model_parallel_world_size, parallel_state
|
||||
from vllm.distributed import utils as dist_utils
|
||||
from vllm.logger import init_logger
|
||||
from vllm.model_executor.layers.attention.mm_encoder_attention import (
|
||||
from vllm.model_executor.layers.attention import (
|
||||
MMEncoderAttention,
|
||||
)
|
||||
from vllm.model_executor.layers.conv import Conv2dLayer, Conv3dLayer
|
||||
|
||||
@@ -32,7 +32,6 @@ import torch
|
||||
from torch import nn
|
||||
from transformers.models.glm4_moe import Glm4MoeConfig
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig, get_current_vllm_config
|
||||
from vllm.distributed import (
|
||||
@@ -42,6 +41,7 @@ from vllm.distributed import (
|
||||
)
|
||||
from vllm.logger import init_logger
|
||||
from vllm.model_executor.layers.activation import SiluAndMul
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.fused_moe import SharedFusedMoE
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import (
|
||||
|
||||
@@ -23,7 +23,7 @@ from vllm.config import VllmConfig
|
||||
from vllm.config.multimodal import BaseDummyOptions
|
||||
from vllm.distributed import get_tensor_model_parallel_world_size
|
||||
from vllm.model_executor.layers.activation import SiluAndMul, get_act_fn
|
||||
from vllm.model_executor.layers.attention.mm_encoder_attention import MMEncoderAttention
|
||||
from vllm.model_executor.layers.attention import MMEncoderAttention
|
||||
from vllm.model_executor.layers.conv import Conv2dLayer
|
||||
from vllm.model_executor.layers.linear import (
|
||||
ColumnParallelLinear,
|
||||
|
||||
@@ -16,7 +16,7 @@ from vllm.config.multimodal import BaseDummyOptions
|
||||
from vllm.distributed.parallel_state import get_tensor_model_parallel_world_size
|
||||
from vllm.inputs.data import PromptType
|
||||
from vllm.model_executor.layers.activation import get_act_fn
|
||||
from vllm.model_executor.layers.attention.mm_encoder_attention import MMEncoderAttention
|
||||
from vllm.model_executor.layers.attention import MMEncoderAttention
|
||||
from vllm.model_executor.layers.linear import (
|
||||
ColumnParallelLinear,
|
||||
QKVParallelLinear,
|
||||
|
||||
@@ -27,7 +27,6 @@ import torch
|
||||
from torch import nn
|
||||
from transformers import GPT2Config
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.distributed.parallel_state import (
|
||||
@@ -35,6 +34,7 @@ from vllm.distributed.parallel_state import (
|
||||
get_tensor_model_parallel_world_size,
|
||||
)
|
||||
from vllm.model_executor.layers.activation import get_act_fn
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.linear import (
|
||||
ColumnParallelLinear,
|
||||
QKVParallelLinear,
|
||||
|
||||
@@ -28,11 +28,11 @@ import torch
|
||||
from torch import nn
|
||||
from transformers import GPTBigCodeConfig
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
|
||||
from vllm.model_executor.layers.activation import get_act_fn
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.linear import (
|
||||
ColumnParallelLinear,
|
||||
QKVParallelLinear,
|
||||
|
||||
@@ -26,11 +26,11 @@ import torch
|
||||
from torch import nn
|
||||
from transformers import GPTJConfig
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
|
||||
from vllm.model_executor.layers.activation import get_act_fn
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.linear import (
|
||||
ColumnParallelLinear,
|
||||
QKVParallelLinear,
|
||||
|
||||
@@ -26,11 +26,11 @@ import torch
|
||||
from torch import nn
|
||||
from transformers import GPTNeoXConfig
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
|
||||
from vllm.model_executor.layers.activation import get_act_fn
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.linear import (
|
||||
ColumnParallelLinear,
|
||||
QKVParallelLinear,
|
||||
|
||||
@@ -7,7 +7,6 @@ import torch.distributed as dist
|
||||
from torch import nn
|
||||
from transformers import GptOssConfig
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.distributed import (
|
||||
@@ -19,6 +18,7 @@ from vllm.distributed import (
|
||||
get_tensor_model_parallel_world_size,
|
||||
tensor_model_parallel_all_gather,
|
||||
)
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.fused_moe import FusedMoE
|
||||
from vllm.model_executor.layers.fused_moe.config import FusedMoEParallelConfig
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
|
||||
@@ -31,11 +31,11 @@ import torch
|
||||
from torch import nn
|
||||
from transformers import GraniteConfig
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
|
||||
from vllm.model_executor.layers.activation import SiluAndMul
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import (
|
||||
MergedColumnParallelLinear,
|
||||
|
||||
@@ -31,7 +31,6 @@ from typing import Any
|
||||
import torch
|
||||
from torch import nn
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.distributed import (
|
||||
@@ -39,6 +38,7 @@ from vllm.distributed import (
|
||||
get_tensor_model_parallel_world_size,
|
||||
tensor_model_parallel_all_gather,
|
||||
)
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.fused_moe import FusedMoE
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import (
|
||||
|
||||
@@ -9,11 +9,11 @@ import torch
|
||||
from torch import nn
|
||||
from transformers import GraniteMoeHybridConfig
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, ModelConfig, VllmConfig
|
||||
from vllm.distributed import get_tensor_model_parallel_world_size
|
||||
from vllm.distributed.parallel_state import get_pp_group
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import QKVParallelLinear, RowParallelLinear
|
||||
from vllm.model_executor.layers.logits_processor import LogitsProcessor
|
||||
|
||||
@@ -32,12 +32,12 @@ import torch
|
||||
import torch.nn.functional as F
|
||||
from torch import nn
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
|
||||
from vllm.logger import init_logger
|
||||
from vllm.model_executor.layers.activation import GeluAndMul
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.fused_moe import FusedMoE
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import (
|
||||
|
||||
@@ -33,7 +33,6 @@ import torch
|
||||
from torch import nn
|
||||
from transformers import PretrainedConfig
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig, get_current_vllm_config
|
||||
from vllm.distributed import (
|
||||
@@ -43,6 +42,7 @@ from vllm.distributed import (
|
||||
tensor_model_parallel_all_reduce,
|
||||
)
|
||||
from vllm.model_executor.layers.activation import SiluAndMul
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.fused_moe import SharedFusedMoE
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import (
|
||||
|
||||
@@ -39,7 +39,7 @@ 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 get_act_fn
|
||||
from vllm.model_executor.layers.attention.mm_encoder_attention import MMEncoderAttention
|
||||
from vllm.model_executor.layers.attention import MMEncoderAttention
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import (
|
||||
ColumnParallelLinear,
|
||||
|
||||
@@ -29,7 +29,7 @@ from transformers.models.idefics2.configuration_idefics2 import (
|
||||
|
||||
from vllm.distributed import get_tensor_model_parallel_world_size
|
||||
from vllm.model_executor.layers.activation import get_act_fn
|
||||
from vllm.model_executor.layers.attention.mm_encoder_attention import MMEncoderAttention
|
||||
from vllm.model_executor.layers.attention import MMEncoderAttention
|
||||
from vllm.model_executor.layers.conv import Conv2dLayer
|
||||
from vllm.model_executor.layers.linear import (
|
||||
ColumnParallelLinear,
|
||||
|
||||
@@ -23,7 +23,7 @@ from vllm.distributed import (
|
||||
tensor_model_parallel_all_gather,
|
||||
)
|
||||
from vllm.model_executor.layers.activation import get_act_fn
|
||||
from vllm.model_executor.layers.attention.mm_encoder_attention import MMEncoderAttention
|
||||
from vllm.model_executor.layers.attention import MMEncoderAttention
|
||||
from vllm.model_executor.layers.conv import Conv2dLayer
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import (
|
||||
|
||||
@@ -10,7 +10,6 @@ import torch
|
||||
from torch import nn
|
||||
from transformers import PretrainedConfig
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.distributed import (
|
||||
@@ -21,6 +20,7 @@ from vllm.distributed import (
|
||||
tensor_model_parallel_all_gather,
|
||||
)
|
||||
from vllm.model_executor.layers.activation import SiluAndMul
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import (
|
||||
MergedColumnParallelLinear,
|
||||
|
||||
@@ -15,7 +15,7 @@ from transformers import PretrainedConfig
|
||||
from transformers.utils import torch_int
|
||||
|
||||
from vllm.model_executor.layers.activation import get_act_fn
|
||||
from vllm.model_executor.layers.attention.mm_encoder_attention import MMEncoderAttention
|
||||
from vllm.model_executor.layers.attention import MMEncoderAttention
|
||||
from vllm.model_executor.layers.conv import Conv2dLayer
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import ColumnParallelLinear, RowParallelLinear
|
||||
|
||||
@@ -24,10 +24,10 @@ import torch
|
||||
from torch import nn
|
||||
from transformers import PretrainedConfig
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.distributed import get_tensor_model_parallel_world_size
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import (
|
||||
ColumnParallelLinear,
|
||||
|
||||
@@ -20,7 +20,7 @@ from vllm.config import VllmConfig
|
||||
from vllm.config.model import ModelConfig
|
||||
from vllm.distributed import parallel_state
|
||||
from vllm.distributed import utils as dist_utils
|
||||
from vllm.model_executor.layers.attention.mm_encoder_attention import MMEncoderAttention
|
||||
from vllm.model_executor.layers.attention import MMEncoderAttention
|
||||
from vllm.model_executor.layers.linear import (
|
||||
ColumnParallelLinear,
|
||||
QKVParallelLinear,
|
||||
|
||||
@@ -28,7 +28,6 @@ from itertools import islice
|
||||
import torch
|
||||
from torch import nn
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.distributed import (
|
||||
@@ -36,6 +35,7 @@ from vllm.distributed import (
|
||||
get_tensor_model_parallel_rank,
|
||||
get_tensor_model_parallel_world_size,
|
||||
)
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.linear import (
|
||||
ColumnParallelLinear,
|
||||
QKVParallelLinear,
|
||||
|
||||
@@ -31,7 +31,6 @@ import torch
|
||||
from torch import nn
|
||||
from transformers import Jais2Config
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.distributed import (
|
||||
@@ -39,6 +38,7 @@ from vllm.distributed import (
|
||||
get_tensor_model_parallel_world_size,
|
||||
)
|
||||
from vllm.model_executor.layers.activation import ReLUSquaredActivation
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.linear import (
|
||||
ColumnParallelLinear,
|
||||
QKVParallelLinear,
|
||||
|
||||
@@ -9,11 +9,11 @@ import torch
|
||||
from torch import nn
|
||||
from transformers import JambaConfig
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, ModelConfig, VllmConfig
|
||||
from vllm.distributed import get_tensor_model_parallel_world_size
|
||||
from vllm.distributed.parallel_state import get_pp_group
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.fused_moe import FusedMoE
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import (
|
||||
|
||||
@@ -20,7 +20,7 @@ from vllm.config import VllmConfig
|
||||
from vllm.config.multimodal import BaseDummyOptions
|
||||
from vllm.distributed import get_tensor_model_parallel_world_size
|
||||
from vllm.logger import init_logger
|
||||
from vllm.model_executor.layers.attention.mm_encoder_attention import (
|
||||
from vllm.model_executor.layers.attention import (
|
||||
MMEncoderAttention,
|
||||
)
|
||||
from vllm.model_executor.layers.conv import Conv2dLayer
|
||||
|
||||
@@ -7,11 +7,11 @@ import torch
|
||||
import torch.nn as nn
|
||||
from transformers import Lfm2Config
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, ModelConfig, VllmConfig
|
||||
from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
|
||||
from vllm.model_executor.layers.activation import SiluAndMul
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import (
|
||||
MergedColumnParallelLinear,
|
||||
|
||||
@@ -6,7 +6,6 @@ from itertools import islice
|
||||
import torch
|
||||
import torch.nn as nn
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, ModelConfig, VllmConfig, get_current_vllm_config
|
||||
from vllm.distributed import (
|
||||
@@ -15,6 +14,7 @@ from vllm.distributed import (
|
||||
get_tensor_model_parallel_world_size,
|
||||
)
|
||||
from vllm.model_executor.layers.activation import SiluAndMul
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.fused_moe import FusedMoE
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import (
|
||||
|
||||
@@ -13,7 +13,7 @@ from transformers import Siglip2VisionConfig
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.distributed import get_tensor_model_parallel_world_size
|
||||
from vllm.model_executor.layers.activation import get_act_fn
|
||||
from vllm.model_executor.layers.attention.mm_encoder_attention import MMEncoderAttention
|
||||
from vllm.model_executor.layers.attention import MMEncoderAttention
|
||||
from vllm.model_executor.layers.linear import (
|
||||
ColumnParallelLinear,
|
||||
QKVParallelLinear,
|
||||
|
||||
@@ -31,12 +31,12 @@ import torch
|
||||
from torch import nn
|
||||
from transformers import LlamaConfig
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
|
||||
from vllm.model_executor.layers.activation import SiluAndMul
|
||||
from vllm.model_executor.layers.attention.encoder_only_attention import (
|
||||
from vllm.model_executor.layers.attention import (
|
||||
Attention,
|
||||
EncoderOnlyAttention,
|
||||
)
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
|
||||
@@ -24,7 +24,6 @@ import torch
|
||||
from torch import nn
|
||||
from transformers import Llama4TextConfig
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.distributed import (
|
||||
@@ -33,7 +32,8 @@ from vllm.distributed import (
|
||||
tensor_model_parallel_all_gather,
|
||||
)
|
||||
from vllm.logger import init_logger
|
||||
from vllm.model_executor.layers.attention.chunked_local_attention import (
|
||||
from vllm.model_executor.layers.attention import (
|
||||
Attention,
|
||||
ChunkedLocalAttention,
|
||||
)
|
||||
from vllm.model_executor.layers.fused_moe import SharedFusedMoE
|
||||
|
||||
@@ -6,7 +6,6 @@ from itertools import islice
|
||||
import torch
|
||||
from torch import nn
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.config import (
|
||||
CacheConfig,
|
||||
VllmConfig,
|
||||
@@ -22,6 +21,7 @@ from vllm.distributed import (
|
||||
)
|
||||
from vllm.logger import init_logger
|
||||
from vllm.model_executor.layers.activation import SiluAndMul
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.fused_moe import FusedMoE
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import (
|
||||
|
||||
@@ -33,7 +33,6 @@ import torch
|
||||
from torch import nn
|
||||
from transformers import PretrainedConfig
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.distributed import (
|
||||
@@ -43,6 +42,7 @@ from vllm.distributed import (
|
||||
tensor_model_parallel_all_reduce,
|
||||
)
|
||||
from vllm.model_executor.layers.activation import FatreluAndMul, SiluAndMul
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.fused_moe import fused_experts, fused_topk
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import (
|
||||
|
||||
@@ -29,9 +29,9 @@ import torch
|
||||
from torch import nn
|
||||
from transformers import PretrainedConfig
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.distributed import get_tensor_model_parallel_world_size
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import (
|
||||
ColumnParallelLinear,
|
||||
|
||||
@@ -30,7 +30,6 @@ import torch
|
||||
from torch import nn
|
||||
from transformers import PretrainedConfig
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, ModelConfig, VllmConfig
|
||||
from vllm.distributed import (
|
||||
@@ -38,6 +37,7 @@ from vllm.distributed import (
|
||||
get_tensor_model_parallel_world_size,
|
||||
tensor_model_parallel_all_reduce,
|
||||
)
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.fused_moe import FusedMoE
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import (
|
||||
|
||||
@@ -14,7 +14,6 @@ import torch
|
||||
from torch import nn
|
||||
from transformers import MiniMaxConfig
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, ModelConfig, VllmConfig
|
||||
from vllm.distributed.parallel_state import (
|
||||
@@ -24,6 +23,7 @@ from vllm.distributed.parallel_state import (
|
||||
)
|
||||
from vllm.forward_context import get_forward_context
|
||||
from vllm.model_executor.layers.activation import SiluAndMul
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.fused_moe import FusedMoE
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import (
|
||||
|
||||
@@ -32,7 +32,6 @@ import torch
|
||||
from torch import nn
|
||||
from transformers import MixtralConfig
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig, get_current_vllm_config
|
||||
from vllm.distributed import (
|
||||
@@ -40,6 +39,7 @@ from vllm.distributed import (
|
||||
get_pp_group,
|
||||
get_tensor_model_parallel_world_size,
|
||||
)
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.fused_moe import FusedMoE
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import (
|
||||
|
||||
@@ -36,7 +36,7 @@ from vllm.config import VllmConfig, set_current_vllm_config
|
||||
from vllm.config.multimodal import BaseDummyOptions
|
||||
from vllm.distributed import get_tensor_model_parallel_world_size
|
||||
from vllm.forward_context import set_forward_context
|
||||
from vllm.model_executor.layers.attention.mm_encoder_attention import MMEncoderAttention
|
||||
from vllm.model_executor.layers.attention import MMEncoderAttention
|
||||
from vllm.model_executor.layers.fused_moe import FusedMoE
|
||||
from vllm.model_executor.layers.linear import (
|
||||
ColumnParallelLinear,
|
||||
|
||||
@@ -10,7 +10,7 @@ from transformers.activations import ACT2FN
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import ModelConfig, VllmConfig
|
||||
from vllm.distributed import get_tensor_model_parallel_world_size
|
||||
from vllm.model_executor.layers.attention.encoder_only_attention import (
|
||||
from vllm.model_executor.layers.attention import (
|
||||
EncoderOnlyAttention,
|
||||
)
|
||||
from vllm.model_executor.layers.linear import QKVParallelLinear, RowParallelLinear
|
||||
|
||||
@@ -17,7 +17,6 @@ from transformers import BatchFeature, PretrainedConfig, ProcessorMixin, TensorT
|
||||
from transformers.image_utils import ImageInput
|
||||
from transformers.tokenization_utils_base import TextInput
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.config.multimodal import BaseDummyOptions
|
||||
@@ -29,7 +28,7 @@ from vllm.distributed import (
|
||||
tensor_model_parallel_all_gather,
|
||||
)
|
||||
from vllm.model_executor.layers.activation import MulAndSilu, QuickGELU, SiluAndMul
|
||||
from vllm.model_executor.layers.attention.mm_encoder_attention import MMEncoderAttention
|
||||
from vllm.model_executor.layers.attention import Attention, MMEncoderAttention
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import (
|
||||
ColumnParallelLinear,
|
||||
|
||||
@@ -23,7 +23,6 @@ from transformers.image_utils import ImageInput
|
||||
from transformers.tokenization_utils_base import TextInput
|
||||
from transformers.video_utils import VideoInput, VideoMetadata
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.config.multimodal import BaseDummyOptions, VideoDummyOptions
|
||||
@@ -36,7 +35,7 @@ from vllm.distributed import (
|
||||
)
|
||||
from vllm.logger import init_logger
|
||||
from vllm.model_executor.layers.activation import MulAndSilu, SiluAndMul, get_act_fn
|
||||
from vllm.model_executor.layers.attention.mm_encoder_attention import MMEncoderAttention
|
||||
from vllm.model_executor.layers.attention import Attention, MMEncoderAttention
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import (
|
||||
ColumnParallelLinear,
|
||||
|
||||
@@ -53,7 +53,7 @@ from transformers.activations import ACT2FN
|
||||
from transformers.modeling_utils import PreTrainedModel
|
||||
|
||||
from vllm.distributed import divide, get_tensor_model_parallel_world_size
|
||||
from vllm.model_executor.layers.attention.mm_encoder_attention import MMEncoderAttention
|
||||
from vllm.model_executor.layers.attention import MMEncoderAttention
|
||||
from vllm.model_executor.layers.conv import Conv2dLayer
|
||||
from vllm.model_executor.layers.linear import (
|
||||
ColumnParallelLinear,
|
||||
|
||||
@@ -10,7 +10,6 @@ import torch
|
||||
import torch.nn as nn
|
||||
from transformers import MptConfig
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.distributed import (
|
||||
@@ -19,6 +18,7 @@ from vllm.distributed import (
|
||||
get_tensor_model_parallel_world_size,
|
||||
)
|
||||
from vllm.model_executor.layers.activation import get_act_fn
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.linear import (
|
||||
ColumnParallelLinear,
|
||||
QKVParallelLinear,
|
||||
|
||||
@@ -30,11 +30,11 @@ from itertools import islice
|
||||
import torch
|
||||
from torch import nn
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
|
||||
from vllm.model_executor.layers.activation import get_act_fn
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.linear import (
|
||||
ColumnParallelLinear,
|
||||
QKVParallelLinear,
|
||||
|
||||
@@ -25,7 +25,6 @@ from itertools import islice
|
||||
import torch
|
||||
from torch import nn
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, ModelConfig, VllmConfig
|
||||
from vllm.config.parallel import ParallelConfig
|
||||
@@ -33,6 +32,7 @@ from vllm.distributed import get_ep_group, get_tensor_model_parallel_world_size
|
||||
from vllm.distributed.communication_op import tensor_model_parallel_all_gather
|
||||
from vllm.distributed.parallel_state import get_pp_group
|
||||
from vllm.model_executor.layers.activation import ReLUSquaredActivation
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.fused_moe import FusedMoE, SharedFusedMoE
|
||||
from vllm.model_executor.layers.fused_moe.utils import activation_without_mul
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
|
||||
@@ -31,11 +31,11 @@ import torch
|
||||
from torch import nn
|
||||
from transformers import OlmoConfig
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
|
||||
from vllm.model_executor.layers.activation import SiluAndMul
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.linear import (
|
||||
MergedColumnParallelLinear,
|
||||
QKVParallelLinear,
|
||||
|
||||
@@ -32,7 +32,6 @@ import torch
|
||||
from torch import nn
|
||||
from transformers import Olmo2Config
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import VllmConfig
|
||||
from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
|
||||
@@ -40,6 +39,7 @@ from vllm.distributed.communication_op import tensor_model_parallel_all_gather
|
||||
from vllm.distributed.parallel_state import get_tensor_model_parallel_rank
|
||||
from vllm.distributed.utils import split_tensor_along_last_dim
|
||||
from vllm.model_executor.layers.activation import SiluAndMul
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import (
|
||||
MergedColumnParallelLinear,
|
||||
|
||||
@@ -21,7 +21,6 @@ from itertools import islice
|
||||
import torch
|
||||
from torch import nn
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import VllmConfig
|
||||
from vllm.distributed import (
|
||||
@@ -32,6 +31,7 @@ from vllm.distributed import (
|
||||
)
|
||||
from vllm.distributed.utils import split_tensor_along_last_dim
|
||||
from vllm.logger import init_logger
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.fused_moe import FusedMoE
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import (
|
||||
|
||||
@@ -29,7 +29,6 @@ import torch
|
||||
from torch import nn
|
||||
from transformers import PretrainedConfig
|
||||
|
||||
from vllm.attention.layer import Attention, AttentionType
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, ParallelConfig, VllmConfig
|
||||
from vllm.distributed import (
|
||||
@@ -41,7 +40,8 @@ from vllm.distributed import (
|
||||
tensor_model_parallel_all_gather,
|
||||
)
|
||||
from vllm.model_executor.layers.activation import SiluAndMul
|
||||
from vllm.model_executor.layers.attention.static_sink_attention import (
|
||||
from vllm.model_executor.layers.attention import (
|
||||
Attention,
|
||||
StaticSinkAttention,
|
||||
)
|
||||
from vllm.model_executor.layers.fused_moe import SharedFusedMoE
|
||||
@@ -84,6 +84,7 @@ from vllm.model_executor.utils import set_weight_attrs
|
||||
from vllm.platforms import current_platform
|
||||
from vllm.sequence import IntermediateTensors
|
||||
from vllm.transformers_utils.config import set_default_rope_theta
|
||||
from vllm.v1.attention.backend import AttentionType
|
||||
from vllm.v1.attention.backends.flash_attn_diffkv import FlashAttentionDiffKVBackend
|
||||
|
||||
|
||||
|
||||
@@ -27,11 +27,11 @@ import torch
|
||||
from torch import nn
|
||||
from transformers import OPTConfig
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
|
||||
from vllm.model_executor.layers.activation import get_act_fn
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.linear import (
|
||||
ColumnParallelLinear,
|
||||
QKVParallelLinear,
|
||||
|
||||
@@ -15,11 +15,11 @@ import torch
|
||||
from torch import nn
|
||||
from transformers import PretrainedConfig
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
|
||||
from vllm.model_executor.layers.activation import SiluAndMul
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.linear import (
|
||||
MergedColumnParallelLinear,
|
||||
QKVParallelLinear,
|
||||
|
||||
@@ -33,11 +33,11 @@ import torch
|
||||
from torch import nn
|
||||
from transformers import PretrainedConfig
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.distributed import get_tensor_model_parallel_world_size
|
||||
from vllm.model_executor.layers.activation import SiluAndMul
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import (
|
||||
MergedColumnParallelLinear,
|
||||
|
||||
@@ -34,7 +34,7 @@ from vllm.config import VllmConfig
|
||||
from vllm.config.multimodal import BaseDummyOptions
|
||||
from vllm.distributed import parallel_state
|
||||
from vllm.distributed import utils as dist_utils
|
||||
from vllm.model_executor.layers.attention.mm_encoder_attention import (
|
||||
from vllm.model_executor.layers.attention import (
|
||||
MMEncoderAttention,
|
||||
)
|
||||
from vllm.model_executor.layers.conv import Conv2dLayer
|
||||
|
||||
@@ -30,11 +30,11 @@ import torch
|
||||
from torch import nn
|
||||
from transformers import PersimmonConfig
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
|
||||
from vllm.model_executor.layers.activation import get_act_fn
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.linear import (
|
||||
ColumnParallelLinear,
|
||||
QKVParallelLinear,
|
||||
|
||||
@@ -45,11 +45,11 @@ import torch
|
||||
from torch import nn
|
||||
from transformers import PhiConfig
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
|
||||
from vllm.model_executor.layers.activation import get_act_fn
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.linear import (
|
||||
ColumnParallelLinear,
|
||||
QKVParallelLinear,
|
||||
|
||||
@@ -31,10 +31,10 @@ import torch
|
||||
from torch import nn
|
||||
from transformers.configuration_utils import PretrainedConfig
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.fused_moe import FusedMoE
|
||||
from vllm.model_executor.layers.linear import (
|
||||
QKVParallelLinear,
|
||||
|
||||
@@ -9,7 +9,6 @@ import torch
|
||||
from torch import nn
|
||||
from transformers import PretrainedConfig
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import VllmConfig, get_current_vllm_config
|
||||
from vllm.distributed import divide, get_tensor_model_parallel_world_size
|
||||
@@ -17,6 +16,7 @@ from vllm.distributed.parallel_state import get_pp_group
|
||||
from vllm.forward_context import ForwardContext, get_forward_context
|
||||
from vllm.model_executor.custom_op import CustomOp
|
||||
from vllm.model_executor.layers.activation import SiluAndMul
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import (
|
||||
ColumnParallelLinear,
|
||||
|
||||
@@ -10,12 +10,12 @@ import torch
|
||||
from torch import nn
|
||||
from transformers import PretrainedConfig
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import VllmConfig
|
||||
from vllm.distributed import get_tensor_model_parallel_world_size
|
||||
from vllm.distributed.parallel_state import get_pp_group
|
||||
from vllm.model_executor.layers.activation import SiluAndMul
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import (
|
||||
MergedColumnParallelLinear,
|
||||
|
||||
@@ -16,11 +16,11 @@ import torch
|
||||
from torch import nn
|
||||
from transformers import PretrainedConfig
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
|
||||
from vllm.model_executor.layers.activation import SiluAndMul
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import (
|
||||
MergedColumnParallelLinear,
|
||||
|
||||
@@ -33,12 +33,12 @@ import torch
|
||||
from torch import nn
|
||||
from transformers import Qwen2Config
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
|
||||
from vllm.model_executor.layers.activation import SiluAndMul
|
||||
from vllm.model_executor.layers.attention.encoder_only_attention import (
|
||||
from vllm.model_executor.layers.attention import (
|
||||
Attention,
|
||||
EncoderOnlyAttention,
|
||||
)
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
|
||||
@@ -49,7 +49,7 @@ from vllm.distributed import utils as dist_utils
|
||||
from vllm.forward_context import set_forward_context
|
||||
from vllm.logger import init_logger
|
||||
from vllm.model_executor.layers.activation import get_act_and_mul_fn
|
||||
from vllm.model_executor.layers.attention.mm_encoder_attention import MMEncoderAttention
|
||||
from vllm.model_executor.layers.attention import MMEncoderAttention
|
||||
from vllm.model_executor.layers.conv import Conv3dLayer
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import (
|
||||
|
||||
@@ -34,12 +34,12 @@ import torch.nn.functional as F
|
||||
from torch import nn
|
||||
from transformers import Qwen2MoeConfig
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
|
||||
from vllm.logger import init_logger
|
||||
from vllm.model_executor.layers.activation import SiluAndMul
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.fused_moe import SharedFusedMoE
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import (
|
||||
|
||||
@@ -49,7 +49,7 @@ from vllm.distributed import parallel_state, tensor_model_parallel_all_gather
|
||||
from vllm.distributed import utils as dist_utils
|
||||
from vllm.logger import init_logger
|
||||
from vllm.model_executor.layers.activation import QuickGELU
|
||||
from vllm.model_executor.layers.attention.mm_encoder_attention import MMEncoderAttention
|
||||
from vllm.model_executor.layers.attention import MMEncoderAttention
|
||||
from vllm.model_executor.layers.conv import Conv3dLayer
|
||||
from vllm.model_executor.layers.linear import (
|
||||
ColumnParallelLinear,
|
||||
|
||||
@@ -30,11 +30,11 @@ import torch
|
||||
from torch import nn
|
||||
from transformers import Qwen3Config
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
|
||||
from vllm.logger import init_logger
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import QKVParallelLinear, RowParallelLinear
|
||||
from vllm.model_executor.layers.logits_processor import LogitsProcessor
|
||||
|
||||
@@ -32,7 +32,6 @@ import torch
|
||||
import torch.nn.functional as F
|
||||
from torch import nn
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig, get_current_vllm_config
|
||||
from vllm.distributed import (
|
||||
@@ -43,6 +42,7 @@ from vllm.distributed import (
|
||||
)
|
||||
from vllm.logger import init_logger
|
||||
from vllm.model_executor.layers.activation import SiluAndMul
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.fused_moe import SharedFusedMoE
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import (
|
||||
|
||||
@@ -10,7 +10,6 @@ from einops import rearrange
|
||||
from torch import nn
|
||||
from transformers.activations import ACT2FN
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import (
|
||||
CacheConfig,
|
||||
@@ -29,6 +28,7 @@ from vllm.distributed import (
|
||||
)
|
||||
from vllm.forward_context import ForwardContext, get_forward_context
|
||||
from vllm.logger import init_logger
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.fla.ops import (
|
||||
chunk_gated_delta_rule,
|
||||
fused_recurrent_gated_delta_rule,
|
||||
|
||||
@@ -30,12 +30,12 @@ import torch
|
||||
from torch import nn
|
||||
from transformers import PretrainedConfig as SeedOssConfig
|
||||
|
||||
from vllm.attention.layer import Attention
|
||||
from vllm.compilation.decorators import support_torch_compile
|
||||
from vllm.config import CacheConfig, VllmConfig
|
||||
from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
|
||||
from vllm.logger import init_logger
|
||||
from vllm.model_executor.layers.activation import SiluAndMul
|
||||
from vllm.model_executor.layers.attention import Attention
|
||||
from vllm.model_executor.layers.layernorm import RMSNorm
|
||||
from vllm.model_executor.layers.linear import (
|
||||
MergedColumnParallelLinear,
|
||||
|
||||
@@ -19,10 +19,10 @@ from vllm.config import VllmConfig
|
||||
from vllm.config.multimodal import BaseDummyOptions
|
||||
from vllm.distributed import divide, get_tensor_model_parallel_world_size
|
||||
from vllm.model_executor.layers.activation import get_act_fn
|
||||
from vllm.model_executor.layers.attention.encoder_only_attention import (
|
||||
from vllm.model_executor.layers.attention import (
|
||||
EncoderOnlyAttention,
|
||||
MMEncoderAttention,
|
||||
)
|
||||
from vllm.model_executor.layers.attention.mm_encoder_attention import MMEncoderAttention
|
||||
from vllm.model_executor.layers.conv import Conv2dLayer
|
||||
from vllm.model_executor.layers.linear import (
|
||||
ColumnParallelLinear,
|
||||
|
||||
@@ -13,7 +13,7 @@ from transformers.configuration_utils import PretrainedConfig
|
||||
|
||||
from vllm.distributed import divide, get_tensor_model_parallel_world_size
|
||||
from vllm.model_executor.layers.activation import get_act_fn
|
||||
from vllm.model_executor.layers.attention.mm_encoder_attention import MMEncoderAttention
|
||||
from vllm.model_executor.layers.attention import MMEncoderAttention
|
||||
from vllm.model_executor.layers.conv import Conv2dLayer
|
||||
from vllm.model_executor.layers.linear import (
|
||||
ColumnParallelLinear,
|
||||
|
||||
Some files were not shown because too many files have changed in this diff Show More
Reference in New Issue
Block a user