From 9801b17f9d0786ce107119af88b92a654e04396f Mon Sep 17 00:00:00 2001 From: Keith Mok Date: Sun, 20 Dec 2020 19:47:25 -0800 Subject: Add argmax deconv support for caffe parser armnn support argmax and deconv , but caffe parser does not. Add back this feature. Signed-off-by: Keith Mok Change-Id: I6b99cc4b58491204c41c6e1d11f583c65c628ee4 --- src/armnnCaffeParser/CaffeSupport.md | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) (limited to 'src/armnnCaffeParser/CaffeSupport.md') diff --git a/src/armnnCaffeParser/CaffeSupport.md b/src/armnnCaffeParser/CaffeSupport.md index ab0d3a4ab8..3501a78ae8 100644 --- a/src/armnnCaffeParser/CaffeSupport.md +++ b/src/armnnCaffeParser/CaffeSupport.md @@ -16,8 +16,10 @@ Although some other neural networks might work, Arm tests the Arm NN SDK with Ca The Arm NN SDK supports the following machine learning layers for Caffe networks: +- Argmax, excluding the top_k and out_max_val parameters. - BatchNorm, in inference mode. -- Convolution, excluding the Dilation Size, Weight Filler, Bias Filler, Engine, Force nd_im2col, and Axis parameters. +- Convolution, excluding Weight Filler, Bias Filler, Engine, Force nd_im2col, and Axis parameters. +- Deconvolution, excluding the Dilation Size, Weight Filler, Bias Filler, Engine, Force nd_im2col, and Axis parameters. Caffe doesn't support depthwise convolution, the equivalent layer is implemented through the notion of groups. ArmNN supports groups this way: - when group=1, it is a normal conv2d @@ -35,4 +37,4 @@ The Arm NN SDK supports the following machine learning layers for Caffe networks - Softmax, excluding the Axis and Engine parameters. - Split. -More machine learning layers will be supported in future releases. \ No newline at end of file +More machine learning layers will be supported in future releases. -- cgit v1.2.1