diff options
author | telsoa01 <telmo.soares@arm.com> | 2018-08-31 09:22:23 +0100 |
---|---|---|
committer | telsoa01 <telmo.soares@arm.com> | 2018-08-31 09:22:23 +0100 |
commit | c577f2c6a3b4ddb6ba87a882723c53a248afbeba (patch) | |
tree | bd7d4c148df27f8be6649d313efb24f536b7cf34 /src/armnnCaffeParser/CaffeSupport.md | |
parent | 4c7098bfeab1ffe1cdc77f6c15548d3e73274746 (diff) | |
download | armnn-c577f2c6a3b4ddb6ba87a882723c53a248afbeba.tar.gz |
Release 18.08
Diffstat (limited to 'src/armnnCaffeParser/CaffeSupport.md')
-rw-r--r-- | src/armnnCaffeParser/CaffeSupport.md | 5 |
1 files changed, 5 insertions, 0 deletions
diff --git a/src/armnnCaffeParser/CaffeSupport.md b/src/armnnCaffeParser/CaffeSupport.md index b5229ebf04..9e4f1fa993 100644 --- a/src/armnnCaffeParser/CaffeSupport.md +++ b/src/armnnCaffeParser/CaffeSupport.md @@ -17,6 +17,11 @@ The Arm NN SDK supports the following machine learning layers for Caffe networks - BatchNorm, in inference mode. - Convolution, 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 + - when group=#input_channels, we can replace it by a depthwise convolution + - when group>1 && group<#input_channels, we need to split the input into the given number of groups, apply a separate convolution and then merge the results - Concat, along the channel dimension only. - Dropout, in inference mode. - Eltwise, excluding the coeff parameter. |