diff options
| -rw-r--r-- | docs/00_introduction.dox | 8 |
1 files changed, 8 insertions, 0 deletions
diff --git a/docs/00_introduction.dox b/docs/00_introduction.dox index 90064399e7..c281cb21b7 100644 --- a/docs/00_introduction.dox +++ b/docs/00_introduction.dox @@ -270,6 +270,14 @@ v20.08 Public major release The default "axis" value for @ref NESoftmaxLayer, @ref NELogSoftmaxLayer is changed from 1 to 0. Only axis 0 is supported. - The support for quantized data types has been removed from @ref CLLogSoftmaxLayer due to implementation complexity. + - Removed padding requirement for the input (e.g. LHS of GEMM) and output in @ref CLGEMMMatrixMultiplyNativeKernel, @ref CLGEMMMatrixMultiplyReshapedKernel, @ref CLGEMMMatrixMultiplyReshapedOnlyRHSKernel and @ref CLIm2ColKernel (NHWC only) + - This change allows to use @ref CLGEMMConvolutionLayer without extra padding for the input and output. + - Only the weights/bias of @ref CLGEMMConvolutionLayer could require padding for the computation. + - Only on Arm Mali Midgard GPUs, @ref CLGEMMConvolutionLayer could require padding since @ref CLGEMMMatrixMultiplyKernel is called and currently requires padding. + - Added support for exporting the OpenCL buffer object to the OpenCL image object in @ref CLGEMMMatrixMultiplyReshapedKernel and @ref CLGEMMMatrixMultiplyReshapedOnlyRHSKernel. + - This support allows to export the OpenCL buffer used for the reshaped RHS matrix to the OpenCL image object. + - The padding requirement for the OpenCL image object is considered into the @ref CLGEMMReshapeRHSMatrixKernel. + - The reshaped RHS matrix stores the weights when GEMM is used to accelerate @ref CLGEMMConvolutionLayer. v20.05 Public major release - Various bug fixes. |