ONE - On-device Neural Engine
|
QuantizeWeights quantizes tensors for weights. More...
#include <QuantizeWeights.h>
Public Member Functions | |
QuantizeWeights (loco::DataType input, loco::DataType output, QuantizationGranularity gr) | |
Public Member Functions inherited from luci::CircleNodeMutableVisitor< void > | |
virtual | ~CircleNodeMutableVisitor ()=default |
Public Member Functions inherited from luci::CircleNodeMutableVisitorBase< T > | |
virtual | ~CircleNodeMutableVisitorBase ()=default |
Data Fields | |
loco::DataType | input_type |
loco::DataType | output_type |
QuantizationGranularity | granularity |
QuantizeWeights quantizes tensors for weights.
Find min/max values on the fly and then quantize
Definition at line 30 of file QuantizeWeights.h.
|
inline |
Definition at line 32 of file QuantizeWeights.h.
QuantizationGranularity luci::QuantizeWeights::granularity |
Definition at line 39 of file QuantizeWeights.h.
loco::DataType luci::QuantizeWeights::input_type |
Definition at line 37 of file QuantizeWeights.h.
loco::DataType luci::QuantizeWeights::output_type |
Definition at line 38 of file QuantizeWeights.h.