diff options
| author | Gian Marco Iodice <gianmarco.iodice@arm.com> | 2018-11-05 14:26:32 +0000 |
|---|---|---|
| committer | Gian Marco Iodice <gianmarco.iodice@arm.com> | 2018-11-08 13:31:33 +0000 |
| commit | 3139f03a74ede3b3bd7cfc6ff219e6c9bc556632 (patch) | |
| tree | cd67b8c9ec1688fcf739c576553b10e435e6938e /src/runtime/CL/functions/CLFullyConnectedLayer.cpp | |
| parent | 421405b6a21b124288a750e2da26dc01eb7391cb (diff) | |
| download | ComputeLibrary-3139f03a74ede3b3bd7cfc6ff219e6c9bc556632.tar.gz | |
COMPMID-1736: Fixed out-of-bound write in CLIm2Col
The issue was related to CLIm2Col when the number of input channels was less than
the number of elements processed by each thread.
The bug has been fixed in the validate_and_configure_window() function setting the correct number of elements accessed
in the output tensor.
Also fixed an issue GEMM3D when we have a single output channel
Change-Id: I094292d0c7662599c4a4c3916ec5f5821df5faef
Diffstat (limited to 'src/runtime/CL/functions/CLFullyConnectedLayer.cpp')
| -rw-r--r-- | src/runtime/CL/functions/CLFullyConnectedLayer.cpp | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/src/runtime/CL/functions/CLFullyConnectedLayer.cpp b/src/runtime/CL/functions/CLFullyConnectedLayer.cpp index c5637dba26..6a2aac6457 100644 --- a/src/runtime/CL/functions/CLFullyConnectedLayer.cpp +++ b/src/runtime/CL/functions/CLFullyConnectedLayer.cpp @@ -101,7 +101,7 @@ void CLFullyConnectedLayer::configure_mm(const ICLTensor *input, const ICLTensor else { // Configure matrix multiply kernel - _mm_gemm.configure(input, weights, nullptr, output, 1.f, 0.0f, GEMMInfo(false, false, true /* Reshape weights only for the first run */, 1, false, retain_internal_weights)); + _mm_gemm.configure(input, weights, nullptr, output, 1.f, 0.0f, GEMMInfo(false, false, true /* Reshape weights only for the first run */, 0, false, retain_internal_weights)); } } |