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

chore: Remove build config from Pytorch kwargs. (#3210)

Browse Source 
Signed-off-by: Frank Di Natale <3429989+FrankD412@users.noreply.github.com>
This commit is contained in:
Frank 2025-04-03 00:00:29 -07:00 committed by GitHub
parent 7f03125098
commit 2d80db4c36
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 1 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
tensorrt_llm/bench/benchmark/throughput.py
View File

@ -23,7 +23,6 @@ from tensorrt_llm.bench.dataclasses.general import BenchmarkEnvironment
from tensorrt_llm.bench.dataclasses.reporting import ReportUtility
from tensorrt_llm.bench.utils.data import (create_dataset_from_stream,
initialize_tokenizer)
from tensorrt_llm.builder import BuildConfig
from tensorrt_llm.llmapi.llm import LLM
from tensorrt_llm.logger import logger
from tensorrt_llm.sampling_params import SamplingParams
@ -227,7 +226,7 @@ def throughput_command(
bench_env.checkpoint_path)
kwargs_max_sql = max_seq_len or metadata.max_sequence_length
logger.info(f"Setting PyTorch max sequence length to {kwargs_max_sql}")
kwargs["build_config"] = BuildConfig(max_seq_len=kwargs_max_sql, )
kwargs["max_seq_len"] = kwargs_max_sql
else:
assert max_seq_len is None, (
"max_seq_len is not a runtime parameter for C++ backend")