DataType _element_type;
Shape _shape;
AffineQuantization _quantization;
- uint8_t *_data;
+ uint8_t *_data = nullptr;
std::string _name;
- bool _data_allocated;
+ bool _data_allocated = false;
// Write of tensor is reported to registered Observers only if this tensor is observable
// This is needed for tensors used in kernel implementation, but not present in original model.
bool _is_observable = true;