diff options
| author | Jongsoo Park <jongsoo@fb.com> | 2019-01-12 06:33:40 +0300 |
|---|---|---|
| committer | Facebook Github Bot <facebook-github-bot@users.noreply.github.com> | 2019-01-12 06:35:56 +0300 |
| commit | 36309fc56728e32b5d78c3be85b48a93f00ed0bf (patch) | |
| tree | 5172c281585bc1855073a92ed11dbe3096dc4c31 /bench | |
| parent | f6e4f991351d16738ebc92ee681cb4eac83d3941 (diff) | |
3x3x3 depthwise convolution with per channel quantization (#15775)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/15775
Pull Request resolved: https://github.com/pytorch/FBGEMM/pull/55
fbgemm didn't have per-channel quantization for 3x3x3 depth-wise convolution
Reviewed By: jianyuh
Differential Revision: D13587438
fbshipit-source-id: 91c36fae7a0e8386e3bc49808e18918b01681dd1
Diffstat (limited to 'bench')
| -rw-r--r-- | bench/Depthwise3DBenchmark.cc | 2 | ||||
| -rw-r--r-- | bench/DepthwiseBenchmark.cc | 1 |
2 files changed, 2 insertions, 1 deletions
diff --git a/bench/Depthwise3DBenchmark.cc b/bench/Depthwise3DBenchmark.cc index 63efc9d..a9b4e4b 100644 --- a/bench/Depthwise3DBenchmark.cc +++ b/bench/Depthwise3DBenchmark.cc @@ -215,7 +215,7 @@ int main() { C_uint8.data(), col_offsets.data(), bias.data(), - false /* fuse_relu */, + false, /* fuse_relu */ tid, num_threads); } diff --git a/bench/DepthwiseBenchmark.cc b/bench/DepthwiseBenchmark.cc index 13be6f6..0fe1b5c 100644 --- a/bench/DepthwiseBenchmark.cc +++ b/bench/DepthwiseBenchmark.cc @@ -292,6 +292,7 @@ int main() { C_uint8.data(), col_offsets.data(), bias.data(), + false, /* fuse_relu */ tid, num_threads); } |