- 15 10月, 2017 2 次提交
-
-
由 qiaolongfei 提交于
-
由 Qiao Longfei 提交于
* add target to Backward, generate var in block when call backward * modify backward_test * fix executor_test * set var desc default type to LOD_TENSOR * update backward_test * insert loss in the top level of backward * create grad vars for all blocks in current program * optimize code * update test_program.py * only create var for newly create blocks when backward
-
- 14 10月, 2017 2 次提交
- 11 10月, 2017 4 次提交
-
-
由 qiaolongfei 提交于
-
由 fengjiayi 提交于
-
由 qiaolongfei 提交于
-
由 fengjiayi 提交于
-
- 10 10月, 2017 3 次提交
-
-
由 Yu Yang 提交于
1. InferShape should be a method for `OpDescBind` not `Operator`, since `OpDescBind` is a compile-time concept. 2. Pre-create operators and store them into a map. Make InferShape faster
-
由 Yu Yang 提交于
It will significantly reduce binary size. It is useful for mobile deployment.
-
由 Yu Yang 提交于
-
- 28 9月, 2017 4 次提交
- 27 9月, 2017 2 次提交
- 26 9月, 2017 5 次提交
- 23 9月, 2017 11 次提交