MCPcopy Create free account
hub / github.com/NVIDIA/TensorRT / engine_get_profile_shape_str

Function engine_get_profile_shape_str

python/src/infer/pyCore.cpp:220–223  ·  view source on GitHub ↗

Overload to allow using binding names instead of indices.

Source from the content-addressed store, hash-verified

218};
219// Overload to allow using binding names instead of indices.
220std::vector<Dims> engine_get_profile_shape_str(ICudaEngine& self, int32_t profileIndex, std::string const& bindingName)
221{
222 return engine_get_profile_shape(self, profileIndex, self.getBindingIndex(bindingName.c_str()));
223};
224
225std::vector<Dims> get_tensor_profile_shape(ICudaEngine& self, std::string const& tensorName, int32_t profileIndex)
226{

Callers

nothing calls this directly

Calls 3

engine_get_profile_shapeFunction · 0.85
getBindingIndexMethod · 0.80
c_strMethod · 0.80

Tested by

no test coverage detected