INT8 conv operator implementation with NCHWc data layout for Intel machines#1680

Merged

yzhliu merged 23 commits intoapache:masterfrom

anijain2305:master

Sep 25, 2018

Commits on Aug 24, 2018

Int8 implementation for convolution operator on Intel Skylake
Animesh Jain
committed

Commits on Aug 27, 2018

Int8 implementation for convolution operator on Intel Skylake
Animesh Jain
committed
Merge branch 'master' of https://github.com/dmlc/tvm
Animesh Jain
committed

Commits on Aug 31, 2018

Merge remote-tracking branch 'upstream/master'
Animesh Jain
committed
Merge branch 'master' of https://github.com/dmlc/tvm
Animesh Jain
committed
PR changes
Animesh Jain
committed
PR changes
Animesh Jain
committed
PR changes
Animesh Jain
committed

Commits on Sep 1, 2018

Fixing an error
Animesh Jain
committed

Commits on Sep 4, 2018

Fixing an error

Animesh Jain
authored and
Ubuntu
committed
Merge branch 'master' of https://github.com/anijain2305/tvm
Ubuntu
committed
Minor typos fix
Ubuntu
committed

Commits on Sep 5, 2018

Minor typos fix
Ubuntu
committed
Merge branch 'master' of https://github.com/anijain2305/tvm
Ubuntu
committed
Removing the broadcast16 CPP code. Using astype feature instead
Ubuntu
committed
Replacing constant by variable name num_elements_intel
Ubuntu
committed

Commits on Sep 6, 2018

Name fixes and tensorize update rule updated
Ubuntu
committed

Commits on Sep 11, 2018

Fixing the bug about checking skylake
Ubuntu
committed

Commits on Sep 13, 2018

Replacing bitcast with reinterpret
Ubuntu
committed
Isolating INT8 and FP32 schedules to ease out future AutoTVM PR merge
Ubuntu
committed

Commits on Sep 14, 2018

Putting check_skylake function in the x86 directory
Ubuntu
committed

Commits on Sep 22, 2018

Added documentation and organizing files to better locations
Ubuntu
committed

Commits on Sep 24, 2018

Tensor intrin renaming. Avoid code duplication for intrin by kernel reshaping
Ubuntu
committed