[Platform][Refactor] Extract func get_default_attn_backend to Platform (#10358)

Signed-off-by: Mengqing Cao <cmq0113@163.com>
This commit is contained in:
Mengqing Cao
2024-11-19 11:22:26 +08:00
committed by GitHub
parent 7eb719df13
commit 8c1fb50705
14 changed files with 99 additions and 69 deletions
+1 -1
View File
@@ -13,7 +13,6 @@ from torch.nn import functional as F
from transformers import PretrainedConfig
from vllm.attention import Attention, AttentionMetadata
from vllm.attention.selector import _Backend
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_rank,
@@ -38,6 +37,7 @@ from vllm.model_executor.layers.vocab_parallel_embedding import (
from vllm.model_executor.model_loader.weight_utils import default_weight_loader
from vllm.multimodal import MULTIMODAL_REGISTRY, MultiModalKwargs
from vllm.multimodal.utils import cached_get_tokenizer
from vllm.platforms import _Backend
from vllm.sequence import (VLLM_TOKEN_ID_ARRAY_TYPE, IntermediateTensors,
SequenceData)
from vllm.transformers_utils.processor import get_processor