Commit message (Expand) | Author | Age | |
---|---|---|---|
* | Add custom call with layout constraints. | Mark Heffernan | 2018-10-08 |
* | [XLA] Modify the function that determines whether an instruction can change | Bixia Zheng | 2018-10-02 |
* | Further simplify the cuDNN wrappers. Instead of passing around | Tim Shen | 2018-09-24 |
* | Add heuristic on picking NHWC layout for (V100, fp16) convolutions. | A. Unique TensorFlower | 2018-05-24 |
* | Do not force default layout when there is no need to. | A. Unique TensorFlower | 2018-05-17 |
* | [TF:XLA] Separate on-host and on-device shape and layout in HloModule. | A. Unique TensorFlower | 2018-05-01 |
* | [XLA:GPU] Implement BatchNormThunk as a call into cudnn. | Justin Lebar | 2017-12-21 |
* | [XLA:CPU] Rename cpu/layout_assignment to cpu/cpu_layout_assignment | Sanjoy Das | 2017-12-07 |