kanshan/TensorRT-LLMs
1
0
Fork 0
mirror of https://github.com/NVIDIA/TensorRT-LLM.git synced 2026-01-14 06:27:45 +08:00
Code Issues Actions 1 Packages Projects Releases Wiki Activity

[None][chore] Remove two unused parameters in create_py_executor (#7458)

Browse Source 
Signed-off-by: leslie-fang25 <leslief@nvidia.com>
This commit is contained in:
Leslie Fang 2025-09-04 07:31:31 +08:00 committed by GitHub
parent 5ff3a65b23
commit bd9ba97d89
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
2 changed files with 2 additions and 15 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
tensorrt_llm/_torch/pyexecutor/py_executor_creator.py
View File

@ -14,9 +14,7 @@ from tensorrt_llm._torch.pyexecutor.resource_manager import ResourceManagerType
from tensorrt_llm._utils import get_sm_version
from tensorrt_llm.bindings.executor import (CapacitySchedulerPolicy,
ContextChunkingPolicy,
ExecutorConfig,
LogitsPostProcessorConfig,
ParallelConfig)
ExecutorConfig)
from tensorrt_llm.bindings.internal.batch_manager import ContextChunkingConfig
from tensorrt_llm.llmapi.llm_args import KvCacheConnectorConfig, TorchLlmArgs
from tensorrt_llm.llmapi.tokenizer import TokenizerBase
@ -217,14 +215,9 @@ def create_py_executor(
tokenizer: Optional[TokenizerBase] = None,
lora_config: Optional[LoraConfig] = None,
kv_connector_config: Optional[KvCacheConnectorConfig] = None,
logits_post_processor_config: Optional[LogitsPostProcessorConfig] = None,
parallel_config: Optional[ParallelConfig] = None,
) -> PyExecutor:
executor_config = llm_args.get_executor_config(checkpoint_dir, tokenizer)
executor_config.logits_post_processor_config = logits_post_processor_config
executor_config.parallel_config = parallel_config
garbage_collection_gen0_threshold = llm_args.garbage_collection_gen0_threshold
_mangle_executor_config(executor_config)

8
tensorrt_llm/executor/worker.py
View File

@ -113,6 +113,7 @@ class GenerationExecutorWorker(GenerationExecutor):
assert hasattr(
self.llm_args, "backend"
), "llm_args should be with backend in _create_py_executor"
_ = _get_comm_ranks_device_id()
if self.llm_args.backend == "pytorch":
from tensorrt_llm._torch.pyexecutor.py_executor_creator import \
create_py_executor
@ -122,13 +123,6 @@ class GenerationExecutorWorker(GenerationExecutor):
args["tokenizer"] = tokenizer
args["lora_config"] = lora_config
args["kv_connector_config"] = kv_connector_config
args[
"logits_post_processor_config"] = tllm.LogitsPostProcessorConfig(
processor_batched=batched_logits_processor,
replicate=False)
comm_ranks, device_ids = _get_comm_ranks_device_id()
args["parallel_config"] = tllm.ParallelConfig(
participant_ids=comm_ranks, device_ids=device_ids)
elif self.llm_args.backend == "_autodeploy":
from tensorrt_llm._torch.auto_deploy.llm_args import \
LlmArgs as ADLlmArgs