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

Method _get_graph_impl

tensorrt_llm/network.py:826–829  ·  view source on GitHub ↗
(self, network_hash: bytes)

Source from the content-addressed store, hash-verified

824
825 #TODO: using one LRU cache here can cause the Network object to be leaked, need a way to speed this function w/o using global lru cache.
826 def _get_graph_impl(self, network_hash: bytes) -> "Network._GraphState":
827 graph = Network._GraphState()
828 graph.build(self)
829 return graph
830
831 @dataclass
832 class _GraphState:

Callers 1

_get_graphMethod · 0.95

Calls 1

buildMethod · 0.45

Tested by

no test coverage detected