MCPcopy Index your code
hub / github.com/NVIDIA/TensorRT-LLM / LlmBuildStats

Class LlmBuildStats

tensorrt_llm/llmapi/llm_utils.py:883–897  ·  view source on GitHub ↗

LlmBuildStats is the statistics for the LLM model building.

Source from the content-addressed store, hash-verified

881
882@dataclass
883class LlmBuildStats:
884 ''' LlmBuildStats is the statistics for the LLM model building. '''
885 # Whether the cache is hit for the engine
886 cache_hitted: bool = False
887 cache_info: Optional[str] = None
888
889 model_from_hf_hub: bool = False
890
891 local_model_dir: Optional[Path] = None
892
893 # The path to the trt-llm engine
894 engine_dir: Optional[Path] = None
895
896 # The build steps information, including the step name and the latency in seconds.
897 build_steps_info: List[Tuple[str, float]] = field(default_factory=list)
898
899
900__all__ = [

Callers 3

test_CachedModelLoaderFunction · 0.85
__init__Method · 0.85
__init__Method · 0.85

Calls

no outgoing calls

Tested by 1

test_CachedModelLoaderFunction · 0.68