summaryrefslogtreecommitdiff
path: root/caffe2/share
AgeCommit message (Expand)AuthorFilesLines
2018-04-12[caffe2] Fix bug in NNPACK bindings for convolution in precomputed transform ...Marat Dukhan1-1/+5
2018-04-09[caffe2] Support fused Conv+Relu with NNPACK (#6375)Marat Dukhan2-4/+61
2018-03-27Remove Apache headers from source.Orion Reblitz-Richardson4-64/+0
2018-03-21Log NNPACK profile to std::cout instead of LOG(INFO)Marat Dukhan1-1/+4
2018-03-06Caffe2 module update: move observers as well as binaries. (#2145)Yangqing Jia12-649/+0
2018-03-02Explicitly require listing additional libraries if a binary needs things beyo...Yangqing Jia1-0/+2
2018-03-02make -DUSE_OBSERVERS=ON workKoan-Sin Tan1-5/+1
2018-03-01Change caffe_add_linker_flag to caffe2_interface_library (#2109)sf-wind1-1/+1
2018-02-27[wip] Cmake modernization (#2066)Yangqing Jia1-1/+0
2018-02-22Support convolution without bias in NNPACK bindingsMarat Dukhan1-6/+21
2018-02-20Allow more backends in caffe2_benchmark (#1979)sf-wind2-8/+10
2018-02-12If a blob is not specified in the init net, create the blobFei Sun1-0/+3
2018-02-02Give NetObserverReporter a virtual destructor for correct destructionMarco Zandonadi1-0/+2
2018-01-26Remove OpenGL code from benchmarkFei Sun1-24/+2
2018-01-11NNPACK: Use new bindings and custom thread poolMarat Dukhan4-21/+30
2018-01-09Move NNPACK integration to share/contrib/nnpackMarat Dukhan2-0/+775
2017-12-21Enable building operator QuantDecompZstdFei Sun3-0/+25
2017-12-14Update observer when attached to RNN opsMartin Schatz2-3/+5
2017-11-29Backed out changeset 4e1241fe65cd (revert a revert :) )Alexander Sidorov2-13/+9
2017-11-22Set a default input typeFei Sun1-1/+1
2017-11-22Add input type to caffe2_benchmarkFei Sun1-1/+13
2017-11-20Backed out changeset 119623addbbdAapo Kyrola2-9/+13
2017-11-17Improve observer framework overheadAlexander Sidorov2-13/+9
2017-11-15Move quant_decomp_zstd.* to share/contribHao Lu2-0/+194
2017-11-14Attach observers to operators inside step netQinqing Zheng2-0/+7
2017-11-13Refactor the observer code to use one function to report both net and operatorFei Sun4-31/+30
2017-11-13Caffe2_benchmark can benchmark multiple backend enginesFei Sun1-3/+41
2017-11-02Update the sample rate function call since the API is changedFei Sun1-2/+2
2017-11-01Update observer sample rateFei Sun4-23/+59
2017-10-18Disable logging by defaultFei Sun1-2/+2
2017-10-16expose observers to python, add multiple observers per observableBram Wasti2-3/+10
2017-10-14Revert D6035393: [caffe2] expose observers to python, add multiple observers ...Scott Yost2-9/+3
2017-10-14expose observers to python, add multiple observers per observableBram Wasti2-3/+9
2017-10-11Consolidate the observer implementationFei Sun2-1/+10
2017-10-11Self register the observer reporter when the file is included in the sourceFei Sun4-5/+18
2017-10-11Revert D6028262: [caffe2][fix] update observer api in perf_observerBram Wasti2-10/+3
2017-10-11update observer api in perf_observerBram Wasti2-3/+10
2017-10-10Rename library name to lowerYangqing Jia2-2/+6
2017-10-06Update Caffe2 benchmark file to write text outputFei Sun1-3/+52
2017-10-04Caffe2 Benchmarking FrameworkFei Sun13-0/+496
2017-09-11Revert D5803245: [caffe2][MPSCNN][segmentation] Make android segmentation net...Hao Lu2-662/+0
2017-09-11Revert D5803411: [caffe2][segmentation]Make iOS segmentation net run with OpenGLHao Lu2-736/+0
2017-09-11Remove the files added by PR 1203Fei Sun5-1900/+0
2017-09-11Make iOS segmentation net run with OpenGLHao Lu2-0/+736
2017-09-11Make android segmentation net run with MPSCNNHao Lu2-0/+662
2017-09-11Bugfix for concat frontendLuke Yeager1-0/+25
2017-09-11Enable use of GPUDirect through argument to Gloo AllreduceOpPieter Noordhuis4-0/+1875