Intel #5721

Open

kofzhoudong wants to merge 1,467 commits into master from intel

+194,904 −29,868

kofzhoudong commented Jun 29, 2017

No description provided.

sfraczek and others added some commits Feb 20, 2017

$@sfraczek$	sfraczek	`task-ssd-detection-output-speedup2: _OPENMP flag corrected Merge remo…` …te-tracking branch 'origin' into task-ssd-detection-output-speedup2 ifdef openmp added detection_output speedup by adding '#pragma omp parallel for' (scoring)	`527347f`
$@sfraczek$	sfraczek	`task-ssd-softmax-speedup2: cannot improve it further`	`7b7dc37`
	michalkuligowski	`Merge branch 'master' of repos.igk.intel.com:intelcaffe into task-lay…` …erDataCompareToolSimplified # Conflicts: # include/caffe/util/compareToolUtilities.h	`575c20a`
	michalkuligowski	`Merge branch 'master' of repos.igk.intel.com:intelcaffe into task-lay…` …erDataCompareToolSimplified	`a827d6e`
	michalkuligowski	`UT fix`	`1572fe3`
	tpatejko	`MKLDNN backward pooling enabled`	`2cb6929`
	jczaja	`- changes of MKLDNN LRN BW pass integrated` - Enalbed MKLDNN LRN bW pass tests	`dda1c4f`
	tpatejko	`Changes after review`	`1c904db`
	jczaja	`task-mkldnn-lrn-backward-enabled: - changes of MKLDNN LRN BW pass int…` …egrated	`e8f728d`
	tpatejko	`task-caffe-mkldnn-integration-pooling: MKLDNN backward pooling enabled`	`ee7a807`
	pawelszewc	`Update of Resnet-50 model`	`1dbc048`
	pawelszewc	`Merge branch 'master' of repos.igk.intel.com:intelcaffe into task-res…` …net50-topology-update	`c4dabb9`
$@sfraczek$	sfraczek	`Softmax was actually improved! Revert the revert`	`31fdd18`
	AdrShv	`fixed problem in SPPLayer`	`eb95af9`
	lwencel	`Multinode optimized models`	`5c0d016`
	Sand3r-	`task-perfmon-temp-strings: Temporary strings in caffe time 2 fixed`	`fefab72`
$@sfraczek$	sfraczek	`task-ssd-softmax-speedup3: Softmax was actually improved! Revert the …` …revert	`b25828f`
	AdrShv	`integrated backward pass for MKLDNN Concat Layer`	`187b912`
	lwencel	`task-multinode_optimized_models: Multinode optimized models`	`3cd14d1`
	AdrShv	`integrated MKLDNN BatchNorm backward`	`bbdc468`
	AdrShv	`fixed bug in LayerSetUp, hide statistics from solver`	`e04136c`
	pawelszewc	`task-resnet50-topology-update: Update of Resnet-50 model`	`213072b`
	AdrShv	`task-mkldnn_batchnorm_backward3: fixed bug in LayerSetUp, hide statis…` …tics from solver	`2d41403`
	AdrShv	`task-mkldnn-concat-backward: integrated backward pass for MKLDNN Conc…` …at Layer	`79cbf4c`
	michalkuligowski	`task-layerDataCompareToolSimplified: Upgraded compare tool Upgraded c…` …ompare tool	`0ead2cd`
	michalkuligowski	`Added more switches to compare tool`	`52845ef`
	michalkuligowski	`Added more switches to compare tool`	`09f8615`
	michalkuligowski	`Added more switches to compare tool`	`ac707f6`
	EC TF Bind account	`Initial empty repository`	`05373db`
	opencici2006	`Merge branch 'master from intel-caffe' into tf`	`452cc9b`
	michalkuligowski	`minor review changes`	`c42a5d1`
	tsocha	`Fix mkl convolution net reshape`	`bce926c`
$@sfraczek$	sfraczek	`mbox`	`f68b69f`
	AdrShv	`fixed memory leaks detected by LeakSanitizer`	`1af7124`
$@sfraczek$	sfraczek	`changed mbox perf measurement to static`	`9263663`
	michalkuligowski	`task-compareToolSwitchesAdded: Added more switches to compare tool`	`81e88e1`
	AdrShv	`task-fixed_memory_leaks_detected_by_LeakSanitizer: fixed memory leaks…` … detected by LeakSanitizer	`05a3521`
	tsocha	`task-mkl-reshape-fix: Fix mkl convolution net reshape`	`1d95f0b`
	jczaja	`- Template for MKLDNN split layer added` - Draft of FW split layer unit test added - Added working MKLDNNSplitLayerTest::TestSetup - Descriptors for MKLDNNSplit BW implemented - Added invoking of MKLDNN Sum primitve for purpose of Split layer - Fix to TestGradient of MKLDNNSplit - Lint fixes - Remvoed unneeded attribute	`56ce805`
	michalkuligowski	`Fix for Klocwork error`	`6d7f32f`
	hshen14	`Improve Caffe time 2.0`	`7fa7dc4`
	jczaja	`task-mkldnn-split-backward: - Template for MKLDNN split layer added`	`94a8936`
	michalkuligowski	`task-fixKlocwork1620: Fix for Klocwork error`	`902fa4b`
$@sfraczek$	sfraczek	`Layer factory enabled eltwise and split layer`	`855c86a`
$@sfraczek$	sfraczek	`ELtwise disabled for now`	`ebfafec`
$@sfraczek$	sfraczek	`changed === to == and fixed test`	`0da45fd`
	jgong5	`Merge repos.igk.intel.com:intelcaffe`	`5c63b12`
$@sfraczek$	sfraczek	`task-mkldnn-layer-factory2: changed === to == and fixed test`	`4f54ea3`
	lwencel	`MKLDNN version update`	`8f44a4b`
	hshen14	`task-caffe-time-2.0-0310: Improve Caffe time 2.0`	`e02de8c`
	tpatejko	`Optimization for annotated data layer with OpenMP with fixes: empty q…` …ueue of RNG, overflow in base convolution layer	`5b4a58a`
	tpatejko	`SSD annotated data layer: changes in data_transformer.cpp missing`	`e04a6eb`
	tpatejko	`SSD annotated data layer optimization: removing redundant code`	`43a553e`
	Yu, Chong	`Implement of MKLDNN-eltwise_layer and related unit tests.`	`83946ed`
	Yu, Chong	`Implement of MKLDNN-eltwise_layer and related unit tests(Changed acco…` …rding to Jacek's comments).	`328d5af`
	tpatejko	`task-ssd-annotated-data-layer-optimization3: SSD annotated data layer…` … optimization: removing redundant code	`f8d0261`
	Yu, Chong	`task-MKLDNN-eltwise_layer-Chong: Implement of MKLDNN-eltwise_layer an…` …d related unit tests.	`73aae23`
	jczaja	`- Fix to Split layer`	`b13f5b1`
	jczaja	`- Enabled MKLDNN Average pooling in layer factory`	`ae34178`
	jczaja	`task-mkldnn-googlenet-training-fixes: - Enabled MKLDNN Average poolin…` …g in layer factory	`991638c`
	Yu, Chong	`Check the setting ot coefficients in the Backward.`	`8033f1f`
$@sfraczek$	sfraczek	`task-performance-ssd: changed mbox perf measurement to static mbox`	`f5ee516`
	Yu, Chong	`task-MKLDNN-eltwise_layer-Backward_coefficient_Check-Chong: Check the…` … setting ot coefficients in the Backward.	`55116c5`
	Sand3r-	`CaffeTime2.0 support for nested events improved`	`ce8ca11`
	Sand3r-	`Merge branch 'master' of repos.igk.intel.com:intelcaffe into task-per…` …fmon-nested-fix	`fa535ed`
	tsocha	`Fix a custom layer engine from prototxt`	`bb2d4fc`
	tsocha	`task-custom-layer-engine: Fix a custom layer engine from prototxt`	`b34f8ab`
	lwencel	`task-mkldnn-version-update2: MKLDNN version update`	`7ac495e`
	tpatejko	`Fixing MKLDNN part in Makefile. Updating git hash to the newest one`	`e67a594`
	tpatejko	`Handle ccache because earlier versions of CMake have problems handlin…` …g it	`4841ff0`
	tpatejko	`Correcting issue with missing separator`	`4c00af6`
	tpatejko	`Plus added so MKLDNN_CMAKE_FLAGS could be set from command line`	`1d211ad`
	tpatejko	`Corrected C compiler passed to MKLDNN`	`b48a874`
	tpatejko	`Prevent building when there's mkldnn already built`	`a2a14f8`
	tpatejko	`Remove Debian-specific stuff from Makefile`	`63df400`
	tpatejko	`Downloading MKLDNN is not executed when test are run`	`f0cd66d`
	tpatejko	`Disabled Test1x1Convolution tests because of problems on IvyBridge`	`b9f5c48`
	tpatejko	`Enabling mkldnn eltwise layer for failing unit test`	`bf617d1`
	tpatejko	`Both cmake and makefile use the same mkldnn commit hash`	`ad2e525`
	tpatejko	`Disabling the rest of MKLDNN convolution tests because of issues with…` … fmadd	`2c52ca6`
	tpatejko	`task-caffe-mkldnn-selfcontained-with-makefile3: Selfcontained IntelCa…` …ffe and MKLDNN with Makefiles: post-review changes	`4f69180`
	Adrian.Shvay	`fix bug in loading trained model, for models which contain BatchNorm …` …layers	`c7a057f`
	Adrian.Shvay	`Merge branch 'master' into task-fix_bug_loading_trained_model_with_Ba…` …tchNorm	`749dbab`
	Adrian.Shvay	`task-fix_bug_loading_trained_model_with_BatchNorm: fix bug in loading…` … trained model, for models which contain BatchNorm layers	`bee02c4`
	jczaja	`- Bumped MKL-DNN version to get benefits of pooling optimizations`	`246209c`
	jczaja	`task-bump-mkl-dnn-version: - Bumped MKL-DNN version to get benefits o…` …f pooling optimizations	`61cca5c`
	Sand3r-	`task-perfmon-nested-fix: CaffeTime2.0 support for nested events improved`	`d26146a`
	AdrShv	`added deleting for split_top_vec elements`	`0d7833e`
	AdrShv	`Merge branch 'master' into prv-memory_leaks_fix`	`674cc53`
	AdrShv	`task-fix_memory_leask_in_test_spp_layer: fixed memory leaks in SPPLay…` …er Test	`8b2e76f`
	tsocha	`Fix a MKL reshape`	`8a73e9f`
	tsocha	`task-mkl-total-reshape-fix: Fix a MKL reshape`	`0dc44cc`
	hshen14	`Fix make regression`	`ceae29d`
	hshen14	`task-make-regression: Fix make regression`	`177d5cf`
	tsocha	`Fix a bug in relu and change solvers from GPU to CPU`	`c4cddd5`
	tsocha	`Merge branch 'master' into task-mkl-total-reshape-hotfix`	`137ef7f`
	tsocha	`task-mkl-total-reshape-hotfix: Fix a bug in relu and change solvers f…` …rom GPU to CPU	`b7b601b`
	Yu, Chong	`Fix the compiler warnings in the source code.`	`67d1232`
	Sand3r-	`Fixed released memory access at InnerProduct-TestForwardTranspose`	`3a08e63`
	Yu, Chong	`Fix the compiler warnings in the source code.`	`4c59615`

tpatejko and others added some commits Apr 20, 2017

tpatejko	`task-collect-compare-inplace-computations: Compare-collect tool does …` …not check/report data used in in-place computations		`614c605`
Sand3r-	`Fix memory leaks in hdf5, ip and spp tests`		`5ce744b`
Sand3r-	`Remove whitespace at eol at spp test`		`161b2fa`
xinanlin	`fix issue ICL-64`		`102bf13`
Yu, Chong	`Fix the threshod of failed unit test.`		`20299f2`
Yu, Chong	`Enable googlenet-v2, fix the bias in the innerproduct, fix the batchn…` …orm when the input blob with the size of 2.		`2f63616`
Yu, Chong	`Disable the average pooling unit test.`		`6d6ee48`
Yu, Chong	`Just update the mkldnn to the version to involve the fix of inner pro…` …duct layer, with the newest mkldnn, average polling unit test will fail.		`79eace3`
daisyden	`Add LICENSE file for resnet`		`0b043e4`
Sand3r-	`task-fix-test-mem-leaks: Fix memory leaks in hdf5, ip and spp tests`		`abac898`
Yu, Chong	`task-enable-googlenet-v2-with-innerproduct-and-batchnorm-fix: Just up…` …date the mkldnn to the version to involve the fix of inner product layer, with the newest mkldnn, average polling unit test will fail. Disable the average pooling unit test. Enable googlenet-v2, fix the bias in the innerproduct, fix the batchnorm when the input blob with the size of 2.		`34c4459`
Yu, Chong	`Add the unit tests for MKLDNN inner product layer to cover the cases …` …without bias.		`8ee9fad`
Yu, Chong	`Enable the ResNet50.`		`0ba59d4`
Yu, Chong	`Eanble ResNet-50 in shanghai local repository.`		`36ea405`
Yu, Chong	`Add a empty line to push the patch for trying to trigger the PRE-CI.`		`409699c`
hshen14	`Merge branch 'intelcaffe'`		`d6574e4`
Yu, Chong	`Fix the customer issue of Concat layer. Remove the unnecessary reshap…` …e in the batch norm backward process.		`5ff1e2f`
Yu, Chong	`Fix the customer issue of Concat layer. Remove the unnecessary reshap…` …e in the batch norm backward process.		`3e255f5`
Yu, Chong	`Merge changes I409699c3,I36ea405b` * changes: Add a empty line to push the patch for trying to trigger the PRE-CI. Eanble ResNet-50 in shanghai local repository.		`8b1c616`
Yu, Chong	`Merge "Fix the customer issue of Concat layer. Remove the unnecessary…` … reshape in the batch norm backward process."		`bcf15e7`
Yu, Chong	`task-add-unittests-mkldnn-innerproduct-layer-without-bias: Add the un…` …it tests for MKLDNN inner product layer to cover the cases without bias.		`dca65f1`
Yu, Chong	`task-enable-ResNet50: Enable the ResNet50.`		`8000c6a`
tpatejko	`Enabling mkldnn convolution tests after qb was updated`		`ff1c871`
tpatejko	`task-mkldnn-convolution-tests-enabled: Enabling mkldnn convolution te…` …sts after qb was updated		`9122ba9`
Yu, Chong	`task-fix-concat-layer-for-customer-issue: Fix the customer issue of C…` …oncat layer. Remove the unnecessary reshape in the batch norm backward process.		`219ebcd`
hshen14	`Merge DCG C1`		`364af6f`
hshen14	`Merge branch 'intelcaffe'`		`1c69b49`
Yu, Chong	`Fix the issue IntelCaffe builds MKL-DNN without MKL GEMM`		`db951c0`
Yu, Chong	`Fix Eltwise Layer: Incorrect top data memory format passed when creat…` …ing sum primitive.		`caf5c42`
hshen14	`Upgrade MKLDNN commit id`		`aa92506`
hshen14	`Upgrade MKL-DNN to version ecf1883a94239a19d442356d32e1076a15b88e7a`		`b0ef323`
Yu, Chong	`Merge "Fix Eltwise Layer: Incorrect top data memory format passed whe…` …n creating sum primitive."		`57c0277`
hshen14	`Upgrade MKL to 20170425`		`dc10f0b`
hshen14	`Merge branch 'master' into dcg_c1`		`5e87a13`
Yu, Chong	`Fix Extra and Incorrect Reorder Primitives executed in FWD/BWD Convol…` …ution. This is a major fix.		`9f6924b`
Yu, Chong	`Comment average pooling unit test for further debugging.`		`03101d3`
hshen14	`Merge branch 'master' into dcg_c1`		`badec0f`
hshen14	`Merge branch 'dcg_c1' of ssh://git-ccr-1.devtools.intel.com:29418/dl_…` …framework-intel_caffe into dcg_c1		`172ade8`
hshen14	`Merge branch 'master' into dcg_c1`		`0c52087`
xinanlin	`fix issue ICL-64`		`348359f`
Yu, Chong	`Fix the threshod of failed unit test.`		`4a0c837`
daisyden	`Add LICENSE file for resnet`		`a02fdc9`
Yu, Chong	`Change the wordings in the comments.`		`9697e63`
Yu, Chong	`Merge "Change the wordings in the comments."`		`0b9431c`
hshen14	`Use MKL 0425 link in Caffe github`		`24e64f4`
Yu, Chong	`Fix the issue that Bwd LRN desc is created with incorrect src memory …` …descriptor.		`122557c`
hshen14	`Warning removal`		`0ca7a6c`
linxinan	`test gerrit`		`283fae1`
linxinan	`Revert "test gerrit"` This reverts commit `283fae1`.		`61d99ed`
Yu, Chong	`Relu optimization related to ICL-84.`		`5bc1e99`
Yu, Chong	`Merge "Relu optimization related to ICL-84."`		`cef51e9`
Yu, Chong	`Fix Pooling bwd executed using non-optimized layout for diff data.`		`c7941f0`
Yu, Chong	`Further improve the Relu performance when prv top diff format is diff…` …erent from prv bottom data format.		`211915e`
Yu, Chong	`Optimize the LRN layer bwd using optimized layout for bottom diff data.`		`845985e`
Yu, Chong	`Fix Googlenet v3 Caffe time test aborted with MKLDNN engine.`		`e2e7332`
daisyden	`Merge branch 'release_1.0.0_rc2' of ssh://git-ccr-1.devtools.intel.co…` …m:29418/dl_framework-intel_caffe	All checks have failed 1 errored and 1 failing checks continuous-integration/travis-ci/push — The Travis CI build could not complete due to an error Details continuous-integration/appveyor/pr — AppVeyor was unable to build non-mergeable pull request	`b8cb4f5`
hshen14	`Align MKL-DNN external latest commit id: a7e17b753c622906f8bdc78f8510…` …e023fc10daaf (2017-06-08)		`9e6c1ba`
daisyden	`fix icl-40 Test phase at the end of training was removed. Remove goog…` …lenet_v3 topology from models		`93a9e4d`
jgong5	`fix a bug of mkldnn integration: no need to call prv descriptor when …` …the head is already on CPU or synced		`439a1c3`
Yu, Chong	`1. Merge the fix of a bug of mkldnn integration: no need to call prv …` …descriptor when the head is already on CPU or synced 2. Revert the workround in the dropout layer. 3. Fix the typo in the Debug info.		`d5b6220`
jgong5	`enable model parallelism and sync bn statistics and scale/shift`		`37c1261`
jgong5	`simplify ByPass logic`		`8bd0ca2`
linxinan	`Support customized batch size and engine for cpp_classification.` Change-Id: I60639cbc827bbea7a3fd17f89db3cd4e887cd402		`23cbe39`
Yu, Chong	`1. Update the mkldnn version to fix the ResNet regression and improve…` … the KNL performance. 2. Change to keep the backward compatibility of cpp_classification.		`8e32af1`
jgong5	`Merge "enable model parallelism and sync bn statistics and scale/shift"`		`ee1cfe5`
jgong5	`Merge "simplify ByPass logic"`		`487ee9e`
jgong5	`Merge "use MLSL_DEFAULT_COLOR to replace -1"`		`a09179f`
jgong5	`Merge "disable model parallelism for cross-channel LRN"`		`401413e`
Yu, Chong	`Merge "1. Update the mkldnn version to fix the ResNet regression and …` …improve the KNL performance. 2. Change to keep the backward compatibility of cpp_classification."		`f636ba6`
jgong5	`use MLSL_DEFAULT_COLOR to replace -1`		`98622e2`
jgong5	`disable model parallelism for cross-channel LRN`		`ade99a1`
Yu, Chong	`Add the shell script to support the automatic downloading and install…` …ation of MLSL.		`178f530`
Yu, Chong	`1. Update the prepare_mlsl script for return value. 2. Enable the mls…` …l download and installation for make and CMake.		`ad7ace9`
Yu, Chong	`Update the prepare_mlsl shell script to handle the case that MLSLROOT…` … is set with no proper version.		`35ddee9`
Yu, Chong	`Update the prepare_mlsl.sh for more complex situations of customized …` …MLSLROOT.		`70aaf4b`
Yu, Chong	`Fix the asum_data() logic of SYNCED_PRV, fix ICL-74.`		`77cddab`
hshen14	`Fix batch normalization global status checker; add optimization for b…` …atch size 1; correct data type for 3D convolution		`9ea6387`
Yu, Chong	`1. Change the tab format. 2. Change the MLSLROOT to MLSL_ROOT.`		`e58f238`
linxinan	`Add option for average pooling include/exclude padding, and default i…` …s include. Change-Id: Id69a653ea65e6282e8cdc83edcfcfd2f4d736b85		`6a61456`
Yu, Chong	`Fix the problem that VERSION_LINE may not be an integer if the MLSL_R…` …OOT is provided by User.		`31ef95e`
jgong5	`fix BN bug of multinode`		`5bb522e`
Yu, Chong	`Merge "Fix the asum_data() logic of SYNCED_PRV, fix ICL-74."`		`88b9626`
Lin, Xinan	`Merge changes I31ef95ee,Ie58f2388` * changes: Fix the problem that VERSION_LINE may not be an integer if the MLSL_ROOT is provided by User. 1. Change the tab format. 2. Change the MLSLROOT to MLSL_ROOT.		`07c6fc9`
hshen14	`Merge "Fix batch normalization global status checker; add optimizatio…` …n for batch size 1; correct data type for 3D convolution"		`3cd80b4`
jgong5	`Merge changes Ib9bd590d,I5bb522e4` * changes: Merge branch 'master' of ssh://git-ccr-1.devtools.intel.com:29418/dl_framework-intel_caffe fix BN bug of multinode		`498286a`
Yu, Chong	`Fix (ICL-128) GoogleNet-V2 inference resut with MKLDNN engine will be…` … -nan with two times running		`cfa6761`
hshen14	`Update Dockfile`		`385e770`
hshen14	`Add Winograd support`		`d4117e4`
jgong5	`Merge branch 'master' of ssh://git-ccr-1.devtools.intel.com:29418/dl_…` …framework-intel_caffe		`b9bd590`
linxinan	`Add mkl_deconvolution layer and corresponding unitest` Change-Id: I1c56fb598b6e517ef85b7177d6fe81d3f17b851b		`da63fb3`
linxinan	`Merge "Add mkl_deconvolution layer and corresponding unitest"`		`c33d743`
Jin, Ge	`Add RemoveBNScale feature` Signed-off-by: Jin, Ge <ge.jin@intel.com>		`fde50b1`
Jin, Ge	`Fix bug in RemoveBNScale feature` Signed-off-by: Jin, Ge <ge.jin@intel.com>		`7a01fb1`
hshen14	`Fix Winograd fallback to direct convolution`		`28cf6b8`
Jin, Ge	`Fix bug in RemoveBNScale feature with GoogleNetv2` Currently, we don't handle fc--bn--scale structure, so the bn layers should be added to kept_bn_layers list. Signed-off-by: Jin, Ge <ge.jin@intel.com>		`792b0d1`
Jin, Ge	`Fix bug in RemoveBNScale` To handle "conv----bn" structure Signed-off-by: Jin, Ge <ge.jin@intel.com>		`881163a`
Jin, Ge	`Fix bug in RemoveBNScale` Signed-off-by: Jin, Ge <ge.jin@intel.com>		`97cd241`
daisyden	`Merge branch 'release_1.0.1_rc1' of ssh://git-ccr-1.devtools.intel.co…` …m:29418/dl_framework-intel_caffe		`9f0108f`
Yu, Chong	`Update and refine the shell script to download the MLSL release and h…` …andle the old version of MLSL. (cherry picked from commit a0fdbf07bb15e6e357e80f63541257b3a4f646a3)		`b0d1651`
daisyden	`Merge branch 'release_1.0.1_rc1' of ssh://git-ccr-1.devtools.intel.co…` …m:29418/dl_framework-intel_caffe	All checks have failed 1 errored and 1 failing checks continuous-integration/travis-ci/push — The Travis CI build could not complete due to an error Details continuous-integration/appveyor/pr — AppVeyor was unable to build non-mergeable pull request	`5b13124`

BVLC/caffe

Join GitHub today

Intel #5721

Conversation

kofzhoudong commented Jun 29, 2017

sfraczek and others added some commits Feb 20, 2017

tpatejko and others added some commits Apr 20, 2017

All checks have failed

All checks have failed