Paddle

Commit Graph

Author	SHA1	Message	Date
Wu Yi	29c63d180f	[Feature] dist op role and lr op role, to support memory optimize with dist training (#13220 ) * wip * clean up * should fix running with memopt * add ut * mark lr schedule op role * hide lr_schedule_guard * use op_role_var instead of ufind * unify dist test name * wip for py3 support * fix var deref * fix python3 mem_opt order * remove comments	6 years ago
Yu Yang	e1913bc5d8	Fix MixedVector	7 years ago
sneaxiy	0a36ef3c18	enhance eager deletion	7 years ago
Yu Yang	6d2c6f96f1	Revert "Revert "Merge pull request #13431 from chengduoZH/refine_lod"" This reverts commit `a6c8d6b9a2`.	7 years ago
Yu Yang	a6c8d6b9a2	Revert "Merge pull request #13431 from chengduoZH/refine_lod" This reverts commit `bd79e04667`, reversing changes made to `6b4d290c18`.	7 years ago
chengduo	d402234ba8	Feature/op_fuse_pass (#12440 ) * Add Preface * Add demo code * Save file * Refine code * seems can work * use elementwise strategy * Use ElementwiseComputeEx * Add comments * extract functions from operator * Refine code * Follow comment * code refine * add op_fuse pass * add backward * code refine * use TopologySortOperations * follow comments * refine IsFusible * code enhance * fix op_fusion_pass * refine code * refine fuse_elemwise_act_op * adjust the input and output * refine logic * add intermediate_edge * disable inplace * follow comments * refine logic * follow comments * Remove the removable IntermediateOut * change strategy * code refine * enable fuse backward * code refine * code refine * rename unit test * follow comments	7 years ago
Xin Pan	23ba766252	Merge pull request #13475 from panyx0718/ir5 avoid creating dangling ir::Node.	7 years ago
Xin Pan	0bd7a67eaf	avoid creating dangling ir::Node. Node should be created by Graph::CreateXXX so that they are managed by graph.	7 years ago
chengduoZH	6534f8527a	Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into refine_lod	7 years ago
Xin Pan	a83a4fab5c	Merge pull request #13441 from panyx0718/ir2 simplify and hide bcast_params	7 years ago
Xin Pan	e5b322051b	clean	7 years ago
chengduoZH	def2a8b1b0	Async memcpy	7 years ago
sneaxiy	612e1a3155	modification	7 years ago
sneaxiy	d0b2453ecd	merge develop	7 years ago
sneaxiy	24ea39c4c6	feature/eager_delete_tensor	7 years ago
Yancey1989	757f9683ab	update comment text	7 years ago
Yancey1989	9ee1b7bc04	add some comments	7 years ago
Yancey1989	bad4ea192e	update by comment	7 years ago
Yancey1989	1664899b63	update	7 years ago
Yancey1989	1e1b6622fd	update by comment	7 years ago
Yancey1989	5ce1a960a5	move bcast op into pass	7 years ago
Xin Pan	5ae052bd14	remove some unused codes	7 years ago
dzhwinter	ab1097cd8e	Feature/template (#13093 ) * remove template operator * "fix compile" * "fix ci" * "fix ci"	7 years ago
Yu Yang	819af27d63	Merge pull request #13079 from reyoung/feature/fix_bug Fix bug when loss@GRAD is reused.	7 years ago
Wu Yi	0b8067c0dc	fix dist train reduce mode (#13068 ) * fix dist train reduce mode * fix previous fix	7 years ago
yuyang18	555f03d8f0	Fix bug when loss@GRAD is reused.	7 years ago
Wu Yi	0ee6fed05b	Refine dist rpc deps (#12899 ) * refine dist train RPC deps * clean up * clean up * fix ut * remove input for fetch_barrier * follow comments	7 years ago
Wu Yi	c6f212a339	Merge pull request #12847 from typhoonzero/add_scope_info_in_graphviz Add scope info in graphviz debug	7 years ago
Wu Yi	b8da70c370	Resovle multi gpu async deps (#12828 ) * dist transpiler add control dependency var between send and recv * fix async deps * follow comments and refine * fix deps connect for rpc ops	7 years ago
typhoonzero	dd7a79158b	add scope info in graphviz debug	7 years ago
yuyang18	b8029fd650	Follow comments	7 years ago
yuyang18	05cadf1b24	Add FastExecutor	7 years ago
yuyang18	c6eb7a89ff	Merge branch 'feature/better_exception_holder' into feature/fast_executor	7 years ago
yuyang18	aac80ef4cc	Polish API of exception holder	7 years ago
yuyang18	d49763a87d	Stash	7 years ago
Xin Pan	626abfc33a	code clean up and renaming Reduce one level of inheritence.	7 years ago
Yu Yang	c4f8afa258	Fix bug when memopt optimize loss.grad and use ParallelExecutor	7 years ago
chengduo	97a77512b4	Fix the order of sum (#12562 ) * fix the order of sum * add doc * check whether need to copy * follow comments	7 years ago
Xin Pan	caf10b474f	make profiler use thread_id from g_thread_id Add a few more RecordEvent. Cleanup	7 years ago
fengjiayi	bc1b7b96ec	Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into dev_exception_holder	7 years ago
qiaolongfei	c8c8c01a23	fix mac build of graph_executor	7 years ago
fengjiayi	3e4083ed1f	Make exception handling of threaded_ssa_graph_executor an independent class	7 years ago
Xin Pan	4b8ae523c4	Merge pull request #12367 from panyx0718/ir_pass Ir pass	7 years ago
Xin Pan	398cfb47b1	disable dist_se_resnext since it's not stable yet. fix fluid_benchmark.py	7 years ago
Xin Pan	ab72d28a5e	clean up and correctness check	7 years ago
Xin Pan	aa1085ddc5	all passes add doc	7 years ago
Xin Pan	e4d7d7ae8f	pass refactoring	7 years ago
Xin Pan	142e832d21	pass registration	7 years ago
qiaolongfei	64e7902530	Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into parallel-executor-support-prefetch	7 years ago
qiaolongfei	754e96a30c	distribute lookup table work with parallel executor	7 years ago
Xin Pan	5173a53c8a	fix reorder issue.	7 years ago
Xin Pan	21a45420f0	polish and test	7 years ago
Xin Pan	93355cc0d2	fix control deps	7 years ago
Xin Pan	f6d99d1f73	polish	7 years ago
Xin Pan	c3f6e0e8a2	add namespace to Graph	7 years ago
Xin Pan	0b3465d215	better	7 years ago
Xin Pan	dcaf183daa	builder SSA graph at the beginning.	7 years ago
Yu Yang	b06309381b	Merge pull request #12149 from reyoung/feature/combine_open_files_and_double_buffer Change and polish readers	7 years ago
chengduo	a3ac54b672	Fix Reduce functor (#12262 ) * Fix Reduce and Gather * Fix unit test	7 years ago
Tao Luo	6fd75edae5	Fix var_handle CMake dependence (#12261 )	7 years ago
qiaolongfei	0e30c9d6fb	fix mac build	7 years ago
yuyang18	7268760fbc	Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into feature/combine_open_files_and_double_buffer	7 years ago
Xin Pan	da5efa735a	fix	7 years ago
Xin Pan	5daad16218	polish	7 years ago
Xin Pan	ff5a7b67ed	polish	7 years ago
Xin Pan	9c9e28b57b	fix program to graph	7 years ago
Xin Pan	64eaa4c829	clean	7 years ago
Xin Pan	10786a243e	polish graph	7 years ago
Xin Pan	2fa8df1caf	separate graph building pass and graph-based pe builder	7 years ago
Xin Pan	37e514432b	op compose node and update nodes.	7 years ago
Xin Pan	9605fcd124	all graphs	7 years ago
Xin Pan	af79b19207	add a simple program to graph	7 years ago
Xin Pan	68aa500451	polish attrs	7 years ago
Xin Pan	9b9603306c	graph attrs	7 years ago
Xin Pan	2eeaa8d5cf	Graph in ParallelExecutor Builder	7 years ago
Xin Pan	7781297c70	variants	7 years ago
yuyang18	b8975d6842	Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into feature/combine_open_files_and_double_buffer	7 years ago
chengduo	86b0a72576	Refine multi thread cpu parallel exe (#11406 ) * refine multi-thread CPU Parallel exe * refine multi thread CPU Parallel exe * Refine CPU version for ParallelExecutor * add share_parameter_between_cards_ * Fix ParallelExecutor bug * Fix unit test * Fix parameter opt balance * Fix with opti (param->grad) * Add grad to op var * Remove shard_param_between_cards	7 years ago
yuyang18	8c2eba71e2	Refine demo	7 years ago
fengjiayi	335e4b9d52	fix a dead lock bug	7 years ago
yuyang18	c822d0309b	Refine code	7 years ago
yuyang18	0a445da631	Make scope_buffered_ssa_graph_executor Exception safe	7 years ago
yuyang18	8e86721fe7	Fix data balance on single GPU	7 years ago
fengjiayi	ce16b40b04	Merge pull request #11891 from JiayiFeng/dev_eof_exp Add EOFException to represent EOF in C++ reader	7 years ago
yuyang18	2d0e5592b5	Use std::map for Place <--> DeviceContext	7 years ago
fengjiayi	3fab4f65a4	Add EOFException to represent EOF in C++ reader	7 years ago
fengjiayi	ff4317cee9	follow comments	7 years ago
fengjiayi	4b950951d3	Add unittests and fix a few bugs	7 years ago
fengjiayi	077434c26c	fix CPU compile error	7 years ago
fengjiayi	47388020a2	fix bugs	7 years ago
fengjiayi	2e320079d3	fix bugs	7 years ago
fengjiayi	5b4f283069	Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into dev_data_balance	7 years ago
fengjiayi	b6dc3a59f1	Add DataBalanceOpHandle to MultiDeviceSSAGragh	7 years ago
yi.wu	88cb47bd86	Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into fix_paraexe_bcast	7 years ago
yi.wu	3d69a82b83	fix dist train broadcasting bug	7 years ago
chengduo	bb29800aaa	small refine (#11460 )	7 years ago
yi.wu	254154a9bf	fix sparse paraexe dist train	7 years ago
fengjiayi	e45a55523b	Merge pull request #11639 from JiayiFeng/fix_mac_compile fix mac compile	7 years ago
chengduo	da556ed6d4	enhance ParallelExecutor stable (#11637 )	7 years ago
fengjiayi	964f515e9a	fix mac compile	7 years ago
Yu Yang	9b3f48d7e6	Merge pull request #11616 from chengduoZH/fix_parallel_exe Enhance Parallel Executor stable	7 years ago
chengduoZH	c99fca5f90	Add No Mutex	7 years ago
chengduoZH	13de72388d	Fix broadcast	7 years ago
Xin Pan	0151e4eb3d	Merge pull request #11608 from panyx0718/doc small thread-safety fix and doc improvements.	7 years ago
chengduoZH	28a86aebc3	Fix Parallel Exe(VarHandel's version)	7 years ago
Xin Pan	df31926fcf	small thread-safety fix and doc improvements.	7 years ago
Yancey1989	3d875b693f	Merge branch 'develop' of github.com:PaddlePaddle/Paddle into overlap_memcpy_with_dist	7 years ago
Yancey1989	f52d78d189	update by comment	7 years ago
Yancey1989	6d752bafd8	use get_appropriate_dev to schedule rpc op	7 years ago
Yancey1989	4444e79e46	Merge branch 'develop' of github.com:PaddlePaddle/Paddle into overlap_memcpy_with_dist	7 years ago
qiaolongfei	16658f7b59	Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into refine-prefetch	7 years ago
qiaolongfei	83a577e8ce	fix build problem	7 years ago
chengduoZH	173d72b481	Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into enable_cpu_on_pe	7 years ago
chengduoZH	aadaadf735	replace use_event with use_cuda, because use_event means the program running with CUDA, so use_cuda maybe more intuitive.	7 years ago
gongweibao	627d7a64f8	Clean `sendop` `recv` operator. (#11309 )	7 years ago
chengduo	fa29ef0b0d	Merge pull request #11277 from chengduoZH/check_ssa_graph Check SSA Graph	7 years ago
chengduoZH	961fbce8e2	follow comments	7 years ago
chengduoZH	7b723839ef	Add cpu test for parallel_executor_crf executor_fetch_feed, and enable these tests	7 years ago
chengduoZH	d24e046c1e	fix allReduce bug	7 years ago
chengduoZH	a57e8a4338	add cpu test	7 years ago
chengduoZH	495368c243	ADD CPU_NUM	7 years ago
chengduoZH	27073c284d	nccl_all_reduce_op_handle => all_reduce_op_handle	7 years ago
chengduoZH	2d94697a82	code refine	7 years ago
chengduoZH	5a3c8bf813	fix in c++ side	7 years ago
fengjiayi	d745840a6e	fix a small compile error on Mac	7 years ago
chengduoZH	0c851cab22	add SSA graph checker	7 years ago
Yancey1989	d5a88b9340	Merge branch 'develop' of github.com:PaddlePaddle/Paddle into overlap_memcpy_with_dist	7 years ago
chengduoZH	1076e85135	refine logic	7 years ago
Yancey	0aa9546eed	fix dist train error (#11281 ) * fix dist train error * update by comment	7 years ago
chengduoZH	8291b916d6	replace graph_builder_factory with ssa_graph_builder_factory	7 years ago
chengduoZH	9ac785be39	check graph's validation	7 years ago
fengjiayi	a1e046bfc0	Merge pull request #11270 from JiayiFeng/fix_a_error_on_max fix a compile error on Mac	7 years ago
fengjiayi	2f5e310167	fix a compile error	7 years ago
yuyang18	8149b0a9aa	Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into add_fuse_var_op_handle	7 years ago
Yancey1989	23433def4b	Merge branch 'develop' of github.com:PaddlePaddle/Paddle into overlap_memcpy_with_dist	7 years ago
yuyang18	d9af153232	SSA Graph Builder Factory * Use Builder Chain to decorate new builders. It is easy to extend builders. * Make graphviz path as a build strategy, not a FLAGS.	7 years ago
chengduoZH	a584bc86dd	add fuse var op handle	7 years ago
Yancey1989	e533a4b4ab	Merge branch 'develop' of github.com:PaddlePaddle/Paddle into overlap_memcpy_with_dist	7 years ago
Yancey1989	82d741c4b9	fix op name typo	7 years ago
Yancey1989	6d69ae0c6e	code cleanup	7 years ago
Yancey1989	93401c98e1	overlap rpc op memcpy in distributed training	7 years ago
yuyang18	86a61c177f	Add ScopeBufferedSSAGraphExecutor	7 years ago
chengduoZH	e330cd032e	balance parameter update	7 years ago
Yancey1989	20c24c05aa	singleton rpc_client	7 years ago
Yancey1989	28596a3386	add gflag ssa_graph_path	7 years ago
Yancey1989	ad6c0142c4	clean up codes	7 years ago
Yancey1989	268e9dc1c6	polish code	7 years ago
Yancey1989	0aa6f9e934	Merge branch 'develop' of github.com:PaddlePaddle/Paddle into overlap_send_op	7 years ago
Yu Yang	d406c76a30	Merge pull request #10744 from reyoung/feature/refine_parallel_executor Disable and fix tests on multi devices.	7 years ago
Yancey1989	540b45350d	use req_count as atomic type	7 years ago

1 2 3 4 5 ...

423 Commits (484c24b756a93b40799c713c5d4f58b416c43542)