Commits

Branch selector

 master

User selector

All users

Datepicker

All time

Commit History

Commits on Jan 13, 2018

• Merge remote-tracking branch 'remotes/internal/release_1.1.0'

  daisyden

  committed
  f6d01ef

Commits on Jan 12, 2018

• Correct batch size pattern used on lmdb dataset benchmarking and make the error log more clear

  manofmountain

  committed
  4bb950c

• Merge branch 'release_1.1.0' of ssh://git-ccr-1.devtools.intel.com:29418/dl_framework-intel_caffe into release_1.1.0

  manofmountain

  committed
  114e2ec

• Fix batch size change issue

  manofmountain

  committed
  c29e6fa

• update models for int8

  xiaolil1

  committed
  17829dd

Commits on Jan 11, 2018

• Add benchmarking descriptions into resnet_50 solver prototxt file

  manofmountain

  committed
  187c913

• Enhance the accuracy calibration tool with below enhancement and fixing.

  Show description for fee2e02
  guomingz

  committed
  fee2e02

Commits on Jan 10, 2018

• update caffe version to 1.1.0

  ftian1

  committed
  69b63e7

Commits on Jan 8, 2018

• Fix for ICL-370, the root cause is we need to check the bottom blob is prv or cpu firstly.

  guomingz

  committed
  29eec8c

• change int8 model to new floating point model and add resnet-50 sparse model

  xiaolil1

  committed
  f9ae217

• Merge "update mklml link to 01.org"

  

  ftian1

  authored and

  Gerrit Code Review

  committed
  a43d6cf

Commits on Jan 7, 2018

• Update MKLDNN version to ae00102be506ed0fe2099c6557df2aa88ad57ec1.

  

  Yu, Chong

  committed
  2edff64

• Add int8 accuracy calibration script.

  guomingz

  committed
  22270ae

• Revert "Merge "Add the calibration tool for int8 accuracy adjustment.""

  Show description for cf53da8
  guomingz

  committed
  cf53da8

Commits on Jan 6, 2018

• fix multinode issue by removing link of MPI lib. originally MPI API is mapped to mpi library incorrectly, which should be mapped to API in MLSL library

  fzou1

  committed
  b448ca1

• Merge "Add the calibration tool for int8 accuracy adjustment."

  

  guomingz

  authored and

  Gerrit Code Review

  committed
  a95b335

Commits on Jan 5, 2018

• modify code to delete debug information

  xiaolil1

  committed
  b04e60f

• Add the calibration tool for int8 accuracy adjustment.

  guomingz

  committed
  c12407e

• update mklml link to 01.org

  ftian1

  committed
  89620ab

Commits on Jan 4, 2018

• correct inceptionv3_int8 topology

  xiaolil1

  committed
  3c5206c

• Add correct prototxt file of resnet_50 for benchmarking

  manofmountain

  committed
  74117c0

• Merge "Add the unit test case for conv/sum/relu fusion."

  

  guomingz

  authored and

  Gerrit Code Review

  committed
  53685f2

Commits on Jan 3, 2018

• correct license in quantization code

  xiaolil1

  committed
  7f0083e

• Merge "Add run_benchmark python script, support scaling test and provide config file for user to do flexible configure on benchmarking. Change-Id: I0eb7db07477f439899a4cec5c4be432123e4a6d7"

  

  manofmountain

  authored and

  Gerrit Code Review

  committed
  1c6a0bb

• Merge "sync LiuWei's SSD sampling bug fix 96175b to intel caffe"

  

  ftian1

  authored and

  Gerrit Code Review

  committed
  1043d90

• Add run_benchmark python script, support scaling test and provide config file for user to do flexible configure on benchmarking.

  Show description for c15df52
  manofmountain

  committed
  c15df52

Commits on Jan 2, 2018

• Merge int8 implementation from 'prv-inf' branch

  xiaolil1

  committed
  d2b4ec5

• Add the unit test case for conv/sum/relu fusion.

  guomingz

  committed
  5e6b199

• fix build issue of weight quantization + older MLSL version (<=2017.1.016)

  fzou1

  committed
  017bc66

• sync LiuWei's SSD sampling bug fix 96175b to intel caffe

  ftian1

  committed
  f2c8795

• move layer timing flag out of multinode flag in build script

  fzou1

  committed
  6df7362

Commits on Dec 27, 2017

• Fix label smoothing using right math utility functions

  hshen14

  committed
  fb26b33

• update SSD topologies to use pure MKLDNN engine.

  ftian1

  committed
  0e6b69c

• set CPU_ONLY to ON by default in CMAKE build

  ftian1

  committed
  6cc851d

Commits on Dec 26, 2017

• Fix error while loading shared libraries libmlsl.so.1 in make process.

  

  Yu, Chong

  committed
  6397ea3

Pagination

PreviousNext