IVGCVSW-8468: DTS Fixing floating point exception
* Added check to prevent modulo by 0 in PerAxisIterator
function SetIndexOnMem
* Added m_Offset property to QSymm8PerAxisEncoder and
QSymm8PerAxisDecoder, to be used instead of a literal '0'
as the quantization offset
* Refactored PerAxisEncoder/PerAxisDecoder creation in functions
MakeEncoder/MakeDecoder in Encoders.hpp/Decoders.hpp
* Modified functions GetTensorInfoForTfLite(Opaque)Tensor
in opaque and classic delegate utils to include quantization
offset for per-axis quantization
Signed-off-by:
Martin Klacer <martin.klacer@arm.com>
Change-Id: I9f450ade65db53247238fb2d109701273398c83e
Loading
Please register or sign in to comment