| Age | Commit message (Expand) | Author | Files | Lines |
|---|
| 2019-04-03 | Upgrade mkldnn-bridge for dnnlowp support (#16308) | Gu, Jinghui | 1 | -18/+19 |
| 2019-03-29 | support pre-convert filter format for mkldnn training mode and change 'Optimi... | Cheng,Penghui | 1 | -12/+151 |
| 2019-01-11 | Remove StopGradient op when it is inplace in inference (#12152) | wuhuikx | 1 | -0/+30 |
| 2019-01-07 | Enable conv+add fusion, same as conv+sum (#15268) | Gu, Jinghui | 1 | -2/+13 |
| 2018-10-29 | seperate mkl, mklml, and mkldnn (#12170) | Gu, Jinghui | 1 | -3/+3 |
| 2018-09-05 | caffe2::DeviceType -> at::DeviceType (#11254) | Jerry Zhang | 1 | -1/+1 |
| 2018-08-13 | Fold AffineChannel to Conv, the same way as BN (for Detectron models) (#10293) | jgong5 | 1 | -34/+62 |
| 2018-08-10 | Optimize max_pooling for inference for MKL-DNN/IDEEP device (#10156) | jgong5 | 1 | -0/+31 |
| 2018-07-16 | Enable Conv fusion optimizations in optimizeForIdeep (#9255) | Gu, Jinghui | 1 | -29/+365 |
| 2018-07-02 | Store OperatorDef by copy (#9108) | Bram Wasti | 1 | -3/+3 |
| 2018-06-08 | [ideep] Fuse Conv-Relu after IDEEP graph rewrite, skip group conv (#8233) | Viswanath Sivakumar | 1 | -0/+8 |
| 2018-05-10 | [caffe2][nomnigraph] Make conv relu fusion more generic (#7437) | Bram Wasti | 1 | -2/+2 |
| 2018-05-08 | [caffe2][nomnigraph] Change the standard transform API to take in NNModule ra... | Bram Wasti | 1 | -6/+3 |
| 2018-05-08 | [Caffe2] Add conv+relu fusion for MKLDNN ops (IDEEP) (#7385) | Yinghai Lu | 1 | -0/+46 |
