Skip to content

chore: refine the default value by using pydantic default instead of … #5695

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Jul 3, 2025
Merged

chore: refine the default value by using pydantic default instead of … #5695

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
22 changes: 14 additions & 8 deletions tensorrt_llm/llmapi/llm_args.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1905,11 +1905,14 @@ def get_pytorch_backend_config(self) -> "PyTorchConfig":
extra_resource_managers=self.extra_resource_managers,
use_cuda_graph=bool(self.cuda_graph_config is not None),
cuda_graph_batch_sizes=self.cuda_graph_config.batch_sizes
if self.cuda_graph_config else None,
if self.cuda_graph_config else
CudaGraphConfig.model_fields['batch_sizes'].default,
cuda_graph_max_batch_size=self.cuda_graph_config.max_batch_size
if self.cuda_graph_config else 0,
if self.cuda_graph_config else
CudaGraphConfig.model_fields['max_batch_size'].default,
cuda_graph_padding_enabled=self.cuda_graph_config.padding_enabled
if self.cuda_graph_config else False,
if self.cuda_graph_config else
CudaGraphConfig.model_fields['padding_enabled'].default,
disable_overlap_scheduler=self.disable_overlap_scheduler,
moe_max_num_tokens=self.moe_max_num_tokens,
moe_load_balancer=self.moe_load_balancer,
Expand All @@ -1923,15 +1926,18 @@ def get_pytorch_backend_config(self) -> "PyTorchConfig":
print_iter_log=self.print_iter_log,
torch_compile_enabled=bool(self.torch_compile_config is not None),
torch_compile_fullgraph=self.torch_compile_config.enable_fullgraph
if self.torch_compile_config is not None else True,
if self.torch_compile_config is not None else
TorchCompileConfig.model_fields['enable_fullgraph'].default,
torch_compile_inductor_enabled=self.torch_compile_config.
enable_inductor if self.torch_compile_config is not None else False,
enable_inductor if self.torch_compile_config is not None else
TorchCompileConfig.model_fields['enable_inductor'].default,
torch_compile_piecewise_cuda_graph=self.torch_compile_config.
enable_piecewise_cuda_graph
if self.torch_compile_config is not None else False,
if self.torch_compile_config is not None else TorchCompileConfig.
model_fields['enable_piecewise_cuda_graph'].default,
torch_compile_enable_userbuffers=self.torch_compile_config.
enable_userbuffers
if self.torch_compile_config is not None else True,
enable_userbuffers if self.torch_compile_config is not None else
TorchCompileConfig.model_fields['enable_userbuffers'].default,
autotuner_enabled=self.autotuner_enabled,
enable_layerwise_nvtx_marker=self.enable_layerwise_nvtx_marker,
load_format=self.load_format,
Expand Down