[Core][Refactor] move parallel_utils into vllm/distributed (#3950)

[WIP][Core][Refactor] move vllm/model_executor/parallel_utils into vllm/distributed and vllm/device_communicators (#3950)
This commit is contained in:
youkaichao
2024-04-10 15:33:30 -07:00
committed by GitHub
parent 934d3662f7
commit 63e7176f26
52 changed files with 111 additions and 141 deletions
+3 -4
View File
@@ -7,13 +7,12 @@ import torch.distributed
from vllm.attention import get_attn_backend
from vllm.config import (CacheConfig, DeviceConfig, LoRAConfig, ModelConfig,
ParallelConfig, SchedulerConfig)
from vllm.distributed import (broadcast_tensor_dict,
ensure_model_parallel_initialized,
init_distributed_environment)
from vllm.logger import init_logger
from vllm.model_executor import set_random_seed
from vllm.model_executor.model_loader import get_model
from vllm.model_executor.parallel_utils.communication_op import (
broadcast_tensor_dict)
from vllm.model_executor.parallel_utils.parallel_state import (
ensure_model_parallel_initialized, init_distributed_environment)
from vllm.sequence import SamplerOutput, SequenceGroupMetadata
from vllm.utils import STR_DTYPE_TO_TORCH_DTYPE
from vllm.worker.model_runner import ModelRunner