Paddle

Commit Graph

Author	SHA1	Message	Date
liym27	39f41cb47f	Performance optimization for dynamic setitem: Call op set_value to speed up because the original call to TensorToPyArray will introduce unnecessary data copy. (#30817 )	4 years ago
joanna.wozna.intel	73cdea01d4	Add bf16 fast performance verification (#30551 ) * Update Xbyak and add bf16 fast performance verification * Fix formating * Change LOG message * Trigger an update of a new tag	4 years ago
WangXi	6e3856d3fb	fix xpu dygraph place (#30868 )	4 years ago
石晓伟	2ac4143b6c	support xpu with analysis predictor, test=develop (#30832 ) * support xpu inference with analysis predictor, test=develop * merge the cmake of the xpu toolchain, test=develop * add c-apis, test=develop * fix a bug in extern_xpu, test=develop	4 years ago
WangXi	b1026f64af	【kunlun】dygraph supports multi xpu card training (#30671 )	4 years ago
Thunderbrook	cb66c53c2d	dump to cpu (#30750 ) * dump to cpu * format * format * format	4 years ago
ShenLiang	3858f458ea	rm Singleton of reducer (#30775 )	4 years ago
Shang Zhizhou	ae0f88a988	add DLA support：C++&&Python api (#30165 ) * add dla * add dla done * add python api Co-authored-by: shangzhizhou <root@szth-rp-fanyi-opera49.szth.baidu.com>	4 years ago
Thunderbrook	1bebc09253	solve build gpu task core (#30626 ) * build gpu task core * format	4 years ago
wanghuancoder	90773473a0	use nvtx push pop in timeline (#30567 ) * delete empty line of pybing.cc, test=develop * use nvtx push pop in timeline, test=develop * change year, test=develop * add #ifdef PADDLE_WITH_CUDA, test=develop * add #ifndef WIN32, test=develop * is_pushed to is_pushed_, test=develop	4 years ago
wanghuancoder	d1b25ed9d7	add some RecordEvent, for dygraph timeline (#30299 ) * add some RecordEvent, for dygraph timeline, test=develop * change GpuMemcpySync to memory::Copy, test=develop * fix compile problem, test=develop * fix compile problem, test=develop * fix, test=develop * fix, test=develop	4 years ago
Leo Chen	7043b8cfc6	support layer_norm fp16 in dygraph amp (#30430 ) * support layer_norm fp16 in dygraph amp * add ut * refine code	4 years ago
wanghuancoder	59ad6ff3e3	delete empty line of pybing.cc, test=develop (#30529 )	4 years ago
hutuxian	e207fe6385	Ascend Framework Part2: pybind files (#30410 )	4 years ago
wanghuancoder	bd97192274	if pybind.cc changed, generate total report, test=develop (#30514 )	4 years ago
guofei	11e78ebaa3	Modify the calculation logic of LambOptimizer (#29313 ) * Modify the calculation logic of LambOptimizer	4 years ago
pangyoki	13d757362c	Add Inplace strategy (Output reuse Input Varbase) in dygraph (#30103 ) * add view strategy on squeeze,unsqueeze,reshape,flatten * add squeeze unittest * add unittests * use View strategy as name rather than Reuse Allacation * fix view api doc * fix format * use core.ops when input of reshape2 is Tensor * fix test_cross_entropy_loss error because of reshape2 * fix test_cross_entropy_loss error because of reshape2 * add inplace strategy * add elementwise_add sub * let backward op not use inplace * grad op do not use inplace * fix memory increase error and add leaf error message * delete selected_rows * change op_function * little change * solve HandleViewBetweenInputAndOutput * add unittest and leaf error message * merge view error * optimize op_function_generator format and support sum inplace op * fix format of basic_engine * fix format for framework * little change of variable wrapper * add reshape, squeeze, unsqueeze, scatter api * add relu elu tanh softmax inplace api * fix test_squeeze_op unittest * fix test_relu_op unittest * fix comment problems * delete sample code of inplace api * add reference of grad_pending_nodes in basic_engine * fix unittest name * add inplace apis into wlist * fix error message * add PADDLE_ENFORCE for set grad op twice * fix head file error	4 years ago
yaoxuefeng	6e0da01c61	Heter ps new (#30198 )	4 years ago
Leo Chen	3d015f1cf5	Set expected place in child thread for dataloader to avoid costing cuda memory on other card (#30338 ) * set expected place in child thread for dataloader * set device id when set tensor from numpy * revert tensor_py change * add compile guard * fix ci * fix bug	4 years ago
ShenLiang	a60f17b89d	Support unused parameters in dynamic graph distributed (#30224 )	4 years ago
tangwei12	25f80fd304	Fix/distributed proto (#29981 ) * rename sendrecv.proto to namespace paddle.distributed * split ps with distributed	4 years ago
Chengmo	d479ae1725	【Paddle.Fleet】Support local save sparse param (#30175 ) * add save tensor support Co-authored-by: seiriosPlus <tangwei12@baidu.com>	4 years ago
AshburnLee	924aac2216	Add tf32 switch for cuDNN (#29192 )	4 years ago
pangyoki	da16b33f2e	add View(reuse allocation) strategy on squeeze, unsqueeze, reshape, flatten op (#29913 ) * add view strategy on squeeze,unsqueeze,reshape,flatten * add squeeze unittest * add unittests * use View strategy as name rather than Reuse Allacation * fix view api doc * fix format * use core.ops when input of reshape2 is Tensor * fix test_cross_entropy_loss error because of reshape2 * delete selected_rows * change op_function * little change * solve HandleViewBetweenInputAndOutput	4 years ago
Leo Chen	8696335f86	Fix dtype of ungenerated grad var (#28511 ) * fix dtype of ungenerated grad var * update ut * refine code * set default dtype * fix could_use_cudnn bug * remove debug code * re-implement * fix bug	4 years ago
Leo Chen	1f97d61c68	Add callback after TensorCopy (#30123 ) * change to tensor copy sync * change to tensor copy sync * make copy_to safe when use TensorCopy * refine code * add ut * add cudapinned garbagecollector * add testcase: cpu place -> cuda pinned place	4 years ago
Chengmo	528e03fc08	【Paddle.Fleet】Fix tensor table (#30075 ) * add tensor table	4 years ago
123malin	198fbdfb60	Add Lookahead and ModelAverage Optimizer (#30004 ) * test=develop, add model_average and lookahead	4 years ago
Leo Chen	adac38c506	add dispenable input for core.ops.reshape2/expand/slice (#30072 ) * add dispenable input 'shape' for core.ops.reshape2 * add dispenable inputs for core.ops.reshape2/expand/slice * add ut	4 years ago
liym27	9922bd4125	Fix bug: In dynamic mode, if start or end is negetive, __getitem__ return wrong result(#30003 ) 1. when slice_item is a slice: 1) the start of __getitem__ should be std::max(start, 0) if slice 2) the start of __getitem__ should be std::min(end, dim) 2. when slice_item is an integer, it should be in [-dim_len, dim_len) 3. Fix error message to use accurate data	4 years ago
Thunderbrook	0b8e1fadc5	add topo-aware in heter-ps (#30087 ) * add topo aware * resource.h * topo aware * format	4 years ago
cc	68398abce9	[Inference] zero_copy_tensor supports int8_t (#30053 ) * zero_copy_tensor supports int8_t	4 years ago
liym27	9602a182b2	[Dynamic Inplace] Support ShareInplaceVersionCounterWith for C++ Tensor (#29842 ) * Revert "[inplace] Add ShareHolderWith for class Variable and SharePlaceholderWith in VarBase.detach() to share the same Tensor/SelectedRows (#29267)" This reverts commit `b10ecd9d3a`. * Support ShareInplaceVersionCounterWith to share the same inplace version counter for VarBase	5 years ago
liuyuhui	4427df37cf	[Kunlun] PR2: Support MultiDevicePass and BKCL in parallel executor (#29574 )	5 years ago
tangwei12	032414ca2a	[Feature] one ps (3/4) (#29604 ) * oneps (3/4) Co-authored-by: MrChengmo <cmchengmo@163.com> Co-authored-by: malin10 <malin10@baidu.com> Co-authored-by: chengmo <chengmo@baidu.com>	5 years ago
Thunderbrook	09b6e71928	heter box (#29734 ) * 　add heter box * add trainer, worker, wrapper... * format * for ci * format * remove boost get * boost & copyright * rename * 　rename * format * format * format Co-authored-by: yaoxuefeng6 <yaoxuefeng@baidu.com>	5 years ago
ShenLiang	01e2874a0e	Support multi-stream communication for dynamic graph distributed (#29525 ) * fix fleet for multi-stream * fix memcpy for ncclid * use sync to solve move operation	5 years ago
liuyuhui	f13c3a9cd7	[Kunlun] PR1:Support one Kunlun card training in parallel executor (#29337 )	5 years ago
Y_Xuan	76738504ad	添加rocm平台支持代码 (#29342 ) * 添加rocm平台支持代码 * 修改一些问题 * 修改一些歧义并添加备注 * 修改代码格式 * 解决冲突后的代码修改 * 修改operators.cmake * 修改格式 * 修正错误 * 统一接口 * 修改日期	5 years ago
AshburnLee	efea540ca9	Add tf32 support for A100 tensor core acceleration for cuBLAS (#28732 )	5 years ago
Wilber	78dad78610	fix none-contiguous bug for python api. (#29615 )	5 years ago
Zhou Wei	e74e1a226c	support deepcopy for Layer/Tensor/Paramerbase (#29387 ) * support deepcopy for Layer/Tensor/Paramerbase * fix some code	5 years ago
ShenLiang	2ef9e0e23c	Rebuild group automatically in dynamic graph distributed (#29255 ) * add tensor_indices in AssignGroupBySize * add rebuild group in reducer	5 years ago
yongqiangma	7c508d8668	update unbind norm add CUDAPlace api doc information (#29322 ) * enhance array_to_lod_tensor_op lod_tensor_to_array_op errors information. test=develop * fix format. test=develop * format fix. test=develop * add lod_rank_table. test=develop * fix format. test=develop * fix doc info. test=develop * fix np error * add unbind dygraph api. test=develop * fix unbind doc.test=develop	5 years ago
liym27	b10ecd9d3a	[inplace] Add ShareHolderWith for class Variable and SharePlaceholderWith in VarBase.detach() to share the same Tensor/SelectedRows (#29267 )	5 years ago
Chen Weihang	9ad800ebb2	Support type promote for basic math ops (quantum required) (#29265 ) * basic impl of type promote * add comment & another testcase * fix complex bugs & support python op promote type * fix failed unittests & polish code * add unittest for coverage * change to only promote complex type * polish code details * polish several comments	5 years ago
Zhen Wang	be3777a50a	Add pure fp16 training with master weights. (#27712 ) * add the weight decay func for the momentum op * Add the multi_precision function in Momentum Optimizer. * Make sure that the initial value of master weights are same with the fp16 weights. * add static loss scaling. * add the rescale_grad function in the pure fp16 training. * use the original momentum updating method. * Polish some codes, such as variable names. * add docstring for apis. * update the var creation details of _create_master_weight. * not modify codes about imperative momentum updating. * Fix the error of test_dist_sparse_tensor_load_momentum UT. * add unit test for multi precision fp16 training. * add more unit tests for CI. * Use lower threshold values for allclose comparing in test_multi_precision_fp16_train UT. * For CI Coverage Checking.	5 years ago
chentianyu03	8f45d14263	add complex64 and complex128 type; add +-/@ and slice opreator for c… (#29199 ) add complex64 and complex128 type; add +-/@ and slice opreator for complex types add test cases for complex elementwise, matmul and getitem unittest * add test cases for complex types * add test cases for complex matmul unittest	5 years ago
Zhou Wei	c0a991c874	accumulate gradient for leaf tensor with previous graph and expose leaf tensor concept (#28429 ) * The leaf tensor concept is exposed and the gradient accumulation of leaf tensor * The leaf tensor concept is exposed and the gradient accumulation of leaf tensor * fix coverage * fix api doc * fix CI unittest * fix CI unittest * fix unitest * empty tensor does’t need inner_var_ * fix some error message	5 years ago
liym27	865a45984f	Check whether there is any inplace operation affecting gradient calculation. (#27901 ) * Add a class TensorInplaceVersion to count the inplace version and put it in framework::Tensor instead of Allocation or Variable. * Add a new attribute `_inplace_version` for VarBase. * Raise exception if an inplace operation can result in incorrect gradient computation. * Add a new interface _bump_inplace_version() for VarBase to bump the version whenever the Tensor is modified through an inplace operation. * For api assign, call _bump_inplace_version() when it's an inplace operation inn dynamic mode. * Use original var_wrapper if the inplace_version is not changed. * Replace SnapshotVarWrapperList with SnapshotVarWrapper to optimize performane.	5 years ago

1 2 3 4 5 ...

879 Commits (010f2caa23d92337bf055a54e00ad088ed360f01)