diff options
| author | Nina Drozd <nina.drozd@arm.com> | 2019-04-25 15:45:20 +0100 |
|---|---|---|
| committer | Nina Drozd <nina.drozd@arm.com> | 2019-05-03 14:43:50 +0100 |
| commit | 59e15b00ea51fee4baeea750dc11ab1952dfab1d (patch) | |
| tree | 97e6c9230bf153d404ad3c3e0e285acdb0b6232d /include | |
| parent | 8b194fbe79d44cba566ad8b508d1c8902987ae3c (diff) | |
| download | armnn-59e15b00ea51fee4baeea750dc11ab1952dfab1d.tar.gz | |
IVGCVSW-2834 Add dynamic quantization via datasets
* Add QuantizationDataSet class for quantization data parsed from CSV file
* Add QuantizationInput for retrieving quantization data for each layer ID
* Add unit tests for command line processor and QuantizationDataSet
Change-Id: Iaf0a747b5f25a59a766ac04f7158e8cb7909d179
Signed-off-by: Nina Drozd <nina.drozd@arm.com>
Diffstat (limited to 'include')
| -rw-r--r-- | include/armnn/INetworkQuantizer.hpp | 4 |
1 files changed, 4 insertions, 0 deletions
diff --git a/include/armnn/INetworkQuantizer.hpp b/include/armnn/INetworkQuantizer.hpp index 54c1c889d3..89548d1057 100644 --- a/include/armnn/INetworkQuantizer.hpp +++ b/include/armnn/INetworkQuantizer.hpp @@ -7,6 +7,7 @@ #include <armnn/INetwork.hpp> #include <armnn/Types.hpp> +#include <armnn/Tensor.hpp> namespace armnn { @@ -37,6 +38,9 @@ public: /// Overrides the default quantization values for the input layer with the given id virtual void OverrideInputRange(LayerBindingId layerId, float min, float max) = 0; + /// Refine input network with a set of refinement data for specified LayerBindingId + virtual void Refine(const InputTensors& inputTensors) = 0; + /// Extract final quantized network virtual INetworkPtr ExportNetwork() = 0; |