projects / platform / upstream / pytorch.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 5e06fa0) | patch
Switch Int8Sigmoid to QNNPACK (#14883)
authorMarat Dukhan <marat@fb.com>
Sat, 8 Dec 2018 10:45:41 +0000 (02:45 -0800)
committerFacebook Github Bot <facebook-github-bot@users.noreply.github.com>
Sat, 8 Dec 2018 10:47:29 +0000 (02:47 -0800)
commitb0909ea6a0db0dd35ba33216f2e8d46008aadeb0
tree3616fb65464b9cf86f583ebef0c55a82edfb9f9etree | snapshot
parent5e06fa0bafb53c1c1a1203c74ddc08209ad87f84commit | diff
Switch Int8Sigmoid to QNNPACK (#14883)

Summary:
50x-100x speedup compared to current version.
Also, fixes a bug in the current version when batch size exceeds 1 (current version processes only the first image in this case).
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14883

Differential Revision: D13390655

Pulled By: Maratyszcza

fbshipit-source-id: 1b33a97bf2d0866d38faa2b42e64fd2859017898
caffe2/operators/quantized/int8_sigmoid_op.h diff | blob | history
third_party/QNNPACK diff | blob | history
Domain: Machine Learning / ML Framework;