Paddle

Commit Graph

Author	SHA1	Message	Date
liu zhengxi	ae2be49f40	Add cublas_handle() to expose cublas_handle to ops (#31157 ) * add get_cublas_handle() api * update format * add unittests * alter function name	4 years ago
Aurelius84	406f4a7513	[CustomOp] Support to specific extra_cflags and exctra_cuda_flags independently (#31059 ) * split cxx/nvcc compile flags * enhance input argument check * rename extra_cflags into extrac_cxx_flags * add name checking in setup * fix test_dispatch failed * fix word typo and rm usless import statement * refine import statement * fix unittest failed * fix cuda flags error	4 years ago
qingqing01	572cc8bd0f	Update doc for 2.0 API and some callback (#31180 ) test=document_fix	4 years ago
Pei Yang	00b09e86ac	[Paddle-TRT] support group_norm (#31040 ) * add group norm plugin * fix compile problems * move concat axis check to trt op teller * add nbDims for scale and bias nv dims * add group norm unit test * fix unittest * add trt version restriction for group norm op teller * fix unittest	4 years ago
Chen Weihang	c209751c8d	change test_multiprocess_reader_exception cmake (#31174 )	4 years ago
YUNSHEN XIE	153121457f	fix ut timeout (#31061 )	4 years ago
Chen Weihang	1ce96fa118	[CustomOp] Add new paddle custom op so (#31141 ) * add new custom op so * fix use new method error * fix test failed	4 years ago
tangwei12	ebbdf52557	fix entry (#31079 ) * fix entry * fix distributed lookup table fuse case * fix entry bug at first time * move entry from paddle.fluid -> paddle.distributed * fix ut with paddle.enable_static() Co-authored-by: malin10 <malin10@baidu.com>	4 years ago
Qi Li	ee76ea72de	[ROCM] update fluid collective op for rocm, test=develop (#31075 )	4 years ago
yaoxuefeng	d8fa65a3a8	fix heter compile (#30518 )	4 years ago
Aurelius84	dce2db4857	[CustomOp] Split build directory for each setup.py (#31124 ) * split build directory for each setup.py * fix template string	4 years ago
Zhou Wei	4b220550ef	[Custom OP]Fix problem of custom op unitests on Windows CI (#31114 ) * fix some problem of Windows custom op * fix some problem of Windows custom op * fix some problem of Windows custom op	4 years ago
chentianyu03	70131b475f	add warning message when dtypes of operator are not same (#31136 ) * add error msg when dtypes of operator are not same * add error msg when dtypes of operator are not same * change error msg to warning msg when dtypes of operator are not same * modify test case to fit for python2	4 years ago
Zhou Wei	be61c2d06b	support build whl and inference library nightly,test=windows3 (#30616 )	4 years ago
alncat	5d6a8c7b73	added support for fake_quantize_dequantize_abs_max op in quantization… (#30896 ) * added support for fake_quantize_dequantize_abs_max op in quantization inference pass * remove const_cast to pass ci * remove compare operator to pass ci-coverage * added detailed error message for unregistered tensorrt_subgrah_pass	4 years ago
Chen Weihang	e60fd1f6a8	[CustomOp] Split test and add inference test (#31078 ) * split test & add inference test * add timeout config * change to setup install * change to jit compile * add verbose for test * fix load setup name repeat * polish details * resolve conflict * fix code format error	4 years ago
Jacek Czaja	d3f09ad702	Update of onednn to 2.2 (#31067 )	4 years ago
Guanghua Yu	24ba5ee05c	merge develop conflict (#31122 )	4 years ago
xiemoyuan	edacb6293c	Optimization of Transformer API (#30957 ) * Support 'bool' and 'int' for attention mask. * Update docs. * Add unittest for Transformer. * fix bugs.	4 years ago
WeiXin	ee1801c1ad	Save load/save pickle protocol (#31044 ) * add default argument for paddle.save/static.save * edit documentation of * Add comments for special processing for protocol=2 and protocol=3. * Update python/paddle/fluid/io.py Co-authored-by: lanxianghit <47554610+lanxianghit@users.noreply.github.com> Co-authored-by: lanxianghit <47554610+lanxianghit@users.noreply.github.com>	4 years ago
Qi Li	cced930b61	[ROCM] update fluid operators for rocm (part1), test=develop (#31077 )	4 years ago
yukavio	99fd9815b6	fix flops api (#31081 ) * remove PrettyTable dependence from paddle.flops * fix bug in python2.7 * fix flops * fix flops * fix bug * fix bug	4 years ago
wangchaochaohu	364cfa2686	fix windows for optimization of elementwise_add Op (#31068 ) * fix windows for optimization of elementwise_add Op	4 years ago
joanna.wozna.intel	781df300d0	Unification of BF16 enablement process (#31034 ) * Unification of bfloat16 enablement process and refactor * Remove unnecessary function * Standardize the output name search	4 years ago
Zhong Hui	16fe11d71e	fix softmax cross entropy integer overflow (#30590 ) [BUG FIX] Fix softmax cross entropy overflow problem.	4 years ago
Zhou Wei	44ee251fde	fix UNIX cmake problem (#31113 )	4 years ago
Qi Li	a60d93fb77	[ROCM] update fluid framework for rocm (part2), test=develop (#31010 )	4 years ago
Thunderbrook	565354f676	support save multi sparse table in one path (#31108 ) * save multi table one path * format	4 years ago
Qi Li	50967135a5	[ROCM] update fluid framework for rocm (part3), test=develop (#31011 )	4 years ago
Huihuang Zheng	cf43a321a8	[Dy2stat] Refactoring tensor_shape_transformer.py to Fix Change after Assign Bug (#31082 ) Problem In our old shape transformer logic, if user write: ``` s = tensor.shape ... y = paddle.some_api(s) ``` Dy2stat will change it to ``` ... y = paddle.some_api(convert_var_shape(tensor)) ``` However it will cause fatal bug if user changes the shape of `x` after assign. For example: ``` s = tensor.shape ... tensor = paddle.some_change_shape_api(tensor) ... y = paddle.some_api(s) ``` Then the Dy2stat will get wrong result because the code is translated into: ``` tensor = paddle.some_change_shape_api(tensor) ... y = paddle.some_api(convert_var_shape(tensor)) # tensor shape has been changed, not origin `s` value ``` Solution Logic It can not be solved in the old logic, so I refactoring tensor_shape_transformer logic. Now we will use `s` to store shape attribute and generate a var `s__STATIC_CONVERT_VAR_SHAPE_SUFFIX` to store static shape API `shape(tensor)` ``` s = tensor.shape ... y = paddle.some_api(s) ``` Dy2stat will change it to ``` s = tensor.shape s__STATIC_CONVERT_VAR_SHAPE_SUFFIX = shape(tensor) ... y = paddle.some_api(choose_shape_attr_or_api(s, s__STATIC_CONVERT_VAR_SHAPE_SUFFIX )) ``` In this case, the code is consistent with origin dygraph meaning and it fixed the change after assign bug. Code Key Note To help reviewers, the key change of this PR is changing `self.name_to_var_shape` from "mapping name to shape node" to "mapping name to its STATIC_CONVERT_VAR_SHAPE_SUFFIX name", then if a variable name has the SUFFIX, we can choose to use attribute shape or shape api. Other changes go with the key change. Consideration The issue of this PR is that we store extra static `shape` API result, will it harms the speed of Dy2stat? In some cases it will, but we argue that the benefit would be greater than the cost. 1. The extra calling to static `shape` API will happen when coder assign among shape variables. Take the following dygraph code as an instance: ``` s1 = tensor.shape s2 = s1 s3 = s2 ... ``` Then we called extra static `shape` APIs again and again, however users seldom write code like this. 2. If the shape variable is used a lot, for example: ``` s = tensor.shape y1 = paddle.some_api1(s) y2 = paddle.some_api2(s) y3 = paddle.some_api3(s) ``` Our old logic will create 3 shape APIs but now just 1. This is more common user code pattern. In fact, if reviewers take a look at the current unit test in this PR, you could see the op numbers decrease after this PR. So we argue that this PR can also improve speed in this code pattern.	4 years ago
tangwei12	0e4b154298	fix dist fleet ctr ut (#31087 ) * fix dist fleet ctr ut Change-Id: I59bf5123c7bd47bd0e8f1ca2a26295257597c0f5 * fix dist fleet ctr ut Change-Id: Iafcdd172364be47fe67b753774ce09af050bcbce * Update CMakeLists.txt	4 years ago
Qi Li	8fe09faf14	[ROCM] update fluid framework for rocm (part1), test=develop (#31009 )	4 years ago
Qi Li	334296306c	[ROCM] update fluid platform for rocm39 (part4), test=develop (#30936 )	4 years ago
Shang Zhizhou	a5c56d83a1	update trt int8 calibrator to IEntropyCalibratorV2 (#31060 ) * update trt int8 calibrator to IEntropyCalibratorV2 * add delele opt_cache for trt_split_converter_test	4 years ago
Zhou Wei	adaec0073d	[2.0Custom OP]Support New Custom OP on Windows (#31063 ) * [2.0.1]Support New Custom OP on windows * fix CI * fix code style * fix CI * fix CI * fix coverage * fix CI * fix CI	4 years ago
Chen Weihang	2168f08ac8	add optional for param attr args, test=document_fix (#31105 )	4 years ago
Qi Li	1d996637e6	[ROCM] update fluid imperative for rocm (part1), test=develop (#31017 ) * [ROCM] update fluid imperative for rocm (part1), test=develop * [ROCM] update reducer.cc after merge, test=develop * update reducer cmake after merge, test=develop	4 years ago
JamesLim	b95eb38b8a	fix the bug in backward OP of index_sample. (#31026 )	4 years ago
Chengmo	6b3371e0c7	Remove PE special profiler (#30886 ) * remove pe special profiler * add profiler info	4 years ago
Chen Weihang	6beeafe797	[CustomOp] Add more dispatch marco for users (#31058 ) * add more dispatch marco * add more dispatch marco * add more tests * revert unneeded change * add timeout for test dispatch * add float and complex test * remove and marco	4 years ago
TTerror	d5323dab41	add squeeze_op/unsqueeze_op on kunlun;fix conv op and parallel executor;optimize lookup_table op (#31056 ) * add squeeze_op/unsqueeze_op on kunlun; fix conv op and parallel executor on kunlun; optimize lookup_table op on kunlun * update squeeze/unsqueeze op	4 years ago
123malin	16b4260b2f	test=develop, save/load, shrink (#30625 ) * test=develop, save/load, shrink Co-authored-by: seiriosPlus <tangwei12@baidu.com>	4 years ago
Shibo Tao	4424aac608	export paddle.static.normalize_program method. (#31072 ) * export paddle.static.normalize_program method. test=develop * fix ut coverage.test=develop	4 years ago
Jiabin Yang	628451af06	hide useless headers and add complex support (#31074 )	4 years ago
Wilber	463eae0383	update paddle_fluid.so to paddle_inference.so (#30850 ) * update paddle_fluid.so to paddle_inference.so	4 years ago
tangwei12	a2170a0866	change fleet reviewer (#31069 ) * change reviewer, test=document Change-Id: I7592ee5c93bd580300ce39df885b603597b09026 * Update check_file_diff_approvals.sh test=document_fix	4 years ago
liym27	5b367dab44	[static setitem] Support the index is Tensor; step>1; step<0 .(#30949 ) * [static setitem] support the index step > 1. tensor_a[::3] = value * [static setitem] support the index step < 0. Eg: tensor_a[::-3] = value * [static setitem] support the index is Tensor. eg: tensor_a[tensor_3:0:-1] = value * Add op version.	4 years ago
Qi Li	eb3050fa9a	[ROCM] update fluid inference for rocm (part1), test=develop (#31018 )	4 years ago
Jack Zhou	6df1ca54c8	add detail about states index in rnn result, test=document_fix (#31048 )	4 years ago
Huihuang Zheng	ef627ac5b9	Fix that convert_var_shape doesn't support slice like [0:], test=develop (#31051 ) As the title, when slice_node like 1:3 being passed to idx of convert_var_shape, it will cause syntax error because a function cannot take this as argument. This PR fixed it.	4 years ago

... 3 4 5 6 7 ...

30185 Commits (b48841ba2e7335eaa435a54436ed580d4aef001c) All Branches Search

30185 Commits (b48841ba2e7335eaa435a54436ed580d4aef001c)

All Branches