diff options
| author | Georgios Pinitas <georgios.pinitas@arm.com> | 2019-07-30 15:09:10 +0100 |
|---|---|---|
| committer | Michele Di Giorgio <michele.digiorgio@arm.com> | 2019-07-31 14:39:12 +0000 |
| commit | 4d600c728a75792c5479b54114ec11c6d8fea61a (patch) | |
| tree | 8d0dc48d651c10f147037948da3c7e419871a3bf /src | |
| parent | a231ea6dcbee90c0079cad1645783f0311a02a6c (diff) | |
| download | ComputeLibrary-4d600c728a75792c5479b54114ec11c6d8fea61a.tar.gz | |
COMPMID-2500: Report error in Int8 Conv2d if multiplier > 1.
Change-Id: I7d0263eddfb4f9cf0145e94b35d5f9e18737cd2d
Signed-off-by: Georgios Pinitas <georgios.pinitas@arm.com>
Reviewed-on: https://review.mlplatform.org/c/1653
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Pablo Marquez <pablo.tello@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Diffstat (limited to 'src')
| -rw-r--r-- | src/runtime/NEON/functions/NEGEMMConvolutionLayer.cpp | 3 |
1 files changed, 2 insertions, 1 deletions
diff --git a/src/runtime/NEON/functions/NEGEMMConvolutionLayer.cpp b/src/runtime/NEON/functions/NEGEMMConvolutionLayer.cpp index bd46944f7a..e94c8933ae 100644 --- a/src/runtime/NEON/functions/NEGEMMConvolutionLayer.cpp +++ b/src/runtime/NEON/functions/NEGEMMConvolutionLayer.cpp @@ -187,7 +187,7 @@ Status NEGEMMConvolutionLayer::validate_mm(const ITensorInfo *input, const ITens float multiplier = iqinfo.scale * wqinfo.scale / oqinfo.scale; int output_multiplier; int output_shift; - quantization::calculate_quantized_multiplier_less_than_one(multiplier, &output_multiplier, &output_shift); + ARM_COMPUTE_RETURN_ON_ERROR(quantization::calculate_quantized_multiplier_less_than_one(multiplier, &output_multiplier, &output_shift)); // Merge activation with output stage int min_activation = 0; @@ -492,6 +492,7 @@ Status NEGEMMConvolutionLayer::validate(const ITensorInfo *input, const ITensorI // Output tensor auto inizialization if not yet initialized ARM_COMPUTE_RETURN_ON_ERROR(NEConvolutionLayerReshapeWeights::validate(weights, biases_to_use, nullptr)); weights_reshaped_info = TensorInfo(compute_weights_reshaped_shape(*weights, (append_bias && !skip_im2col)), 1, data_type); + weights_reshaped_info.set_quantization_info(weights->quantization_info()); weights_to_use = &weights_reshaped_info; if(!skip_im2col) |