summaryrefslogtreecommitdiff
path: root/caffe2/python/layers
AgeCommit message (Expand)AuthorFilesLines
2019-02-28add dropout during eval (#17549)Huan Gui1-1/+3
2019-01-25Revert D13551909: [fbcode] logdevice for generic feature typeJunjie Bai1-3/+1
2019-01-24logdevice for generic feature type (#16191)Qin Huang1-1/+3
2018-11-27Resubmit: Set the correct engine name for position weighted pooling when fp16...Jiyan Yang1-11/+30
2018-11-24Add Recency Weighted into SparseLookup (#14291)Huan Gui1-2/+9
2018-11-11fc layer accept axis argument (#13822)Yan Zhu1-5/+13
2018-10-25Implement bucket-based attention pooling for IdScoreList features (#13004)Frank Jiang1-0/+68
2018-10-15Revert D10123465: Set the correct engine name for position weighted pooling w...Andrey Malevich1-34/+12
2018-10-12Set the correct engine name for position weighted pooling when fp16 is used f...Jiyan Yang1-12/+34
2018-10-11Allow use substitute ops for LayerNorm (#12177)Xiaolong Wang1-2/+46
2018-09-14Support FP16 sparse lookup (#11674)Jiyan Yang1-3/+30
2018-09-05refactor preproc, support dense in TumHistory layerYan Zhu1-0/+30
2018-08-16Remove unused code base for distributed training (#10282)Hassan Eslami2-2/+0
2018-08-15move matrix formation for dot products to precompute/request-only (#10531)Bangsheng Tang1-2/+2
2018-08-04support generic feature in DPER2 (#10197)Qin Huang1-125/+179
2018-07-28Fix bug that always uses the same blob when repeating poolingsHuayu Li1-1/+1
2018-07-25Co-disitillation with different archs and/or feature set (#9793)Yuan Xie1-2/+2
2018-06-27fbshipit-source-id: ba600fcd2b5cefc7621357bdeb05e24cea02e5afOrion Reblitz-Richardson1-36/+29
2018-06-26Update from Facebook (#8887)Orion Reblitz-Richardson4-60/+229
2018-06-13Update from facebook (#8384)sf-wind4-45/+257
2018-06-01Update from facebook 1ee4edd286a3 (#8040)Bram Wasti1-1/+20
2018-05-29Update from facebook (#7855)Sebastian Meßmer1-2/+2
2018-05-19Update from facebook (#7696)bddppq1-0/+6
2018-05-10Update from facebook (#7451)Paul Jesse Hellemn1-0/+2
2018-05-01[Caffe2][fbcode=>GH sync] Update from facebook 4323b18ce13c (#7116)Lu Fang2-1/+8
2018-04-17Update from Facebook (#6692)Orion Reblitz-Richardson2-0/+6
2018-04-10[Caffe2] Changes done inside Facebook (#6378)Yinghai Lu5-9/+45
2018-03-30Revert D7394363: [GanH]: Log D Trick for Cross Entropy with SigmoidAndrey Malevich1-4/+0
2018-03-30[GanH]: Log D Trick for Cross Entropy with SigmoidXiaolong Wang1-0/+4
2018-03-27[D2]: Improve loss weight by allowing omitted weightsXiaolong Wang1-1/+10
2018-03-27[GanH] Label Smooth: Add Layer and Integrate to SparseNNXiaolong Wang2-5/+102
2018-03-27[layer model] remove duplicated init opsXianjie Chen1-8/+17
2018-03-27Remove Apache headers from source.Orion Reblitz-Richardson37-555/+0
2018-03-20[GanH] Weighting Layers: Adaptive/Constant/HomotopyXiaolong Wang3-0/+288
2018-03-20[GanH]: homotopy JSDXiaolong Wang1-12/+72
2018-03-20Allow calculating average margin rank lossYan Shang1-2/+6
2018-03-12Update caffe2 from facebook 4f527ef46abf (#2234)sf-wind4-24/+103
2018-03-06[GanH] Model TestDmytro Dzhulgakov1-2/+7
2018-03-06[GanH]fuse jsd with lr loss/xentDmytro Dzhulgakov1-5/+29
2018-03-06[DPER] Fix bug in uint8 quantization shortcut.Dmytro Dzhulgakov1-1/+12
2018-02-26[c2] update SparseFeatureHash layerOrion Reblitz-Richardson1-20/+19
2018-02-15Use 8-bit quantization only in cases when it makes sense.Andrey Malevich1-0/+4
2018-02-15Add CPU_ONLY tag for sparse_feature_hash layerXianjie Chen1-0/+8
2018-02-13Allow custom component tagging in DeviceOptions.node_nameMatan Appelbaum1-6/+20
2018-02-08Make TypeInference work for HalfToFloat & FloatToHalf.Lin Yang1-0/+2
2018-02-02Add regularization for sparse featuresYan Shang1-3/+6
2018-01-31change ModOp to support output sign configurationsXue Feng1-1/+1
2018-01-26testPairwiseDotProductLin Yang2-12/+45
2018-01-25Integrate Fused8BitRowwise ops with DPERPeter Goldsborough1-10/+24
2018-01-23add error msg in fc input_recordXue Feng1-1/+2
generated by cgit v1.2.3 (git 2.25.1) at 2024-11-26 17:02:05 +0000