- 26 2月, 2019 7 次提交
-
-
由 Yiqun Liu 提交于
Optimize the CUDA implementation of sequence_expand op by reduce the times of copying lod data from CPU to GPU. (#15493) * Optimize the CUDA implementation of sequence_expand op by reduce the times of copying lod data from CPU to GPU. test=develop * Refine the op benchmark to support setting lod in config. test=develop
-
由 Tao Luo 提交于
Add Conv Residual Connection UT for Projection
-
由 guomingz 提交于
* This PR improve performance of prior_box op about 1.25x faster on CPU. * Test Env:SKX 8180 with fake data on 28 threads(bs=1). * The below table shows the ~25% improvement which generated by [eval_tp_fake_data.py](https://github.com/PaddlePaddle/Paddle/issues/15618#issuecomment-464613976). | Type |Event | Calls | Total | Min. | Max. | Ave. | Ratio.| | ---------------- | ------------------ | ---- | ------- | -------- | -------- | ------------ | -------- | | w/ optimization | thread0::prior_box | 6000 | 921.201 | 0.110572 | 0.383402 | **0.153533** | 0.084585 | | w/o optimization | thread0::prior_box | 6000 | 1151.85 | 0.102276 | 0.426702 | **0.191976** | 0.103337 | test=develop * Fix the style issue. test=develop
-
由 Tao Luo 提交于
Refactor MKL-DNN Sum to use reference version on fallback
-
由 chengduo 提交于
* add alloc_continuous_space_op test=develop * Polish code test=develop * follow comment test=develop
-
由 dzhwinter 提交于
fix nightly build
-
由 wopeizl 提交于
fix build issue for cudaEvent_t
-
- 25 2月, 2019 19 次提交
-
-
由 Michal Gallus 提交于
test=develop
-
由 Michal Gallus 提交于
test=develop
-
由 dzhwinter 提交于
accelerate memory optimize process
-
由 peizhilin 提交于
-
由 peizhilin 提交于
test=develop
-
由 Qiyang Min 提交于
Imperative training network to the end
-
由 Xin Pan 提交于
Pass graph to parallel executor instead of program
-
由 wopeizl 提交于
fix build issue on windows for sample prop op
-
由 chengduo 提交于
* refile profiler test=develop * follow comment test=develop
-
由 qingqing01 提交于
* Refine doc of uniform_random and fix dtype * Update defaule value in the arguments
-
由 Xin Pan 提交于
add per kernel config and remove const_cast.
-
由 dzhwinter 提交于
-
由 Jacek Czaja 提交于
* - Implemented draft of primitive desc keeping in Tensor test=develop - TransposeMKLDNNHandler::AcquireSrcMemory was reimplemented - Added nchw and nc formats setting for sake of compatiblity Fixed unit tests - Worakaround to problem with 5D data in conv - Added 3D and 1D MKL-DNN formats for name handles for tensor test=develop - Fix to UTs test=develop - Conv fp32 op was updated Cosmetic fixes test=develop - tensor mkldnn cosmetics test=develop - Moved most of mkl-dnn specific code from Tensor to mkl-dnn utils * - Lint fixes test=develop * - setting prim dec in Tensor , sets also layout to kMKLDNN test=develop * - Moved creation of prim desc totally out of Tensor test=develop * - Cosmetic fixes adter review test=develop
-
由 Kaipeng Deng 提交于
fix pool3d doc. test=develop
-
由 Xin Pan 提交于
test=develop
-
由 minqiyang 提交于
test=develop
-
由 Xin Pan 提交于
test=develop
-
由 peizhilin 提交于
test=develop
-
由 dzhwinter 提交于
-
- 24 2月, 2019 7 次提交
-
-
由 dengkaipeng 提交于
-
由 dengkaipeng 提交于
-
由 dengkaipeng 提交于
-
由 dengkaipeng 提交于
-
由 dengkaipeng 提交于
-
由 Dun 提交于
-
由 xuezhong 提交于
fix bug for sampled softmax
-
- 23 2月, 2019 7 次提交
-
-
由 乔龙飞 Qiao Longfei 提交于
fix cpu broadcast
-
由 minqiyang 提交于
test=develop
-
由 Tao Luo 提交于
Change *(smart_ptr.get()) -> *smart_ptr
-
由 xuezhong 提交于
test=develop
-
由 Cheerego 提交于
* [Don't merge now]update_readme_to_1.3 * fix sth test=develop * update reademe_cn test=develop * fix en test=develop
-
由 xuezhong 提交于
add sample_logits and sampled_softmax_with_cross_entropy op
-
由 Qiao Longfei 提交于
-