bd66eed50a
Trainer save load params ( #10386 )
...
* Load/save the params from the params_path
* Switch to use load_persistables and save_persistables
* Instaed of setup the executor to run program and scope. Pass the program to the load_persistables
8 years ago
5812076e7d
Merge pull request #10316 from helinwang/dist
...
Fluid new API: dist train without modifying code
8 years ago
f9c680c43e
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into feature/update_sparse_parameter
8 years ago
3bb99c4f66
Added auto transform to beam_search_decode_op ( #10286 )
...
* Added auto transform to beam_search_decode_op
* Added some comment
* Added unittest for beam_search_decode_op on GPU
8 years ago
ddf6167213
Correct filename ( #10384 )
8 years ago
773bf29b70
Merge pull request #10377 from Yancey1989/fix_delete_ops
...
fix delete_ops
8 years ago
8cc91bc025
initial commit ( #10387 )
8 years ago
7a86069422
Add float16 demo code and put float16 work in contrib/float16 folder ( #10331 )
...
* add test float16 inference accuracy example
* complete the test
* clean code
* add argument parse and refine tests
* add shell script
* add float16 benchmark code
* refine code
* prepare for contrib/float16
* put things in contrib float16 folder
* update benchmark result
* further update benchmark report
* add float16 inference report
* update report
8 years ago
8ee23da846
Fluid new API: dist train without modifying code
...
Works with 1 trainer 1 pserver. 2 trainer 1 pserver will stuck at the
end of first step, still investigating.
The user only need to set envrionment variables to enable distributed
training.
run pserver:
PADDLE_TRAINING_ROLE=PSERVER PADDLE_PSERVER_IPS=127.0.0.1 PADDLE_TRAINERS=2 PADDLE_CURRENT_IP=127.0.0.1 python no_test_word2vec_new_api.py
run trainer:
PADDLE_TRAINING_ROLE=TRAINER PADDLE_PSERVER_IPS=127.0.0.1 PADDLE_TRAINERS=2 PADDLE_TRAINER_ID=0 python no_test_word2vec_new_api.py
8 years ago
f428e82d25
Prediction should be a part of inference_network in new API ( #10356 )
8 years ago
ea522dabc9
refine delete ops
8 years ago
4145258296
fix delete_ops
8 years ago
3000e9946f
Write the Understand Sentiment book example with stacked LSTM using new API ( #10355 )
...
* Add understand apiu with stacked lstm for new API
* Complete exam
8 years ago
be219ac0ea
Merge pull request #10354 from helinwang/scaffold
...
Improve trainer API
8 years ago
73650a8320
Merge pull request #10342 from Yancey1989/refine_distribute_transpiler_api
...
Refine distribute transpiler api
8 years ago
e9737d600f
add a private function to find adam opt pass
8 years ago
1ae086ed19
Merge pull request #10340 from wanghaoshuang/fix_future_div
...
Make Variable support for future.division.
8 years ago
88d79dfe95
Merge pull request #10292 from typhoonzero/fix_grpc_server_ready_condition
...
Fix grpc server ready condition
8 years ago
4a497b826d
MKLDNN implementation of batch normalization ( #9904 )
...
* Initial implementation of forward pass for MKLDNN batch norm
* Added attributes for MKLDNN batch norm
* MKLDNN batch norm forward pass passes unittest. Started working on backward
* Backward pass for MKLDNN batch norm added
* MKLDNN batch norm: scoring added to forward pass
* MKLDNN batch norm: bias as input added; handling AnyLayout when kernel is looked up
* MKLDNN batch norm: python unit tests added; mkldnn tests removed
* MKLDNN batch norm: changes required by cpplint
* MKLDNN batch norm: refactoring the operator
* MKLDNN batch norm: saved variance inversed in backward pass for correct execution of MKLDNN unit tests
* MKLDNN batch norm: refctoring, function for static/const cast to void* added
* MKLDNN batch norm: remove AnyLayout from batch norm
* MKLDNN batch norm: only NCHW format is supported. Unittests refactored
* MKDNN batch norm: use_mkldnn added to attributes
* MKLDNN batch norm: AnyLayout removed from unittest
* MKLDNN batch norm: added CUDNN defines to batch norm
* MKLDNN batch norm: undefined data_format variable corrected
* MKLDNN batch norm: use_cudnn added, use of setUp method for configuring attributes
* MKLDNN batch norm: added use_cudnn attribute to batch norm operator
* MKLDNN batch norm: correcting batch norm unit tests for MKLDNN
* MKLDNN batch norm: MKLDNN tests moved to another file; reverting changes for saved variance not being inverted
* Change default layout to NCHW
* MKLDNN batch norm: init_kernel_type method added to unit tests
* MKLDNN batch norm: style changes
* MKLDNN batch norm: unit tests refactored
* MKLDNN batch norm: added use_mkldnn attribute to batch norm python interface
8 years ago
0fca8a14ef
Fix fluid/__init__.py
8 years ago
a66052c6ff
improve trainer API
...
- The trainer and inferencer will load params from disk if param_path
argument is not None in their constructor.
- Remove params.py, we will expose core.Scope to the user if needed
(e.g., for GAN). Currently we will not expose it, unless we clearly
know doing so can support GAN.
- Add `save_params` to Trainer (a TODO item).
- rename "network" to "program"
8 years ago
6d5e582d20
Merge pull request #10343 from reyoung/feature/new_api_train_impl
...
A naive implement trainer.train by executor
8 years ago
c891189568
update sparse gradient parameter with reduce and broadcast
8 years ago
6422c0e4f6
update by comment
8 years ago
5ff1ef36ee
update sparse parameter
8 years ago
1bb579a3f5
A naive trainer implementation
8 years ago
faebadd938
Merge pull request #10228 from jacquesqiao/use-multi-thread-todo-update
...
Use multi thread to do update
8 years ago
da960ada49
redefine distribute transpiler api
8 years ago
ff99d94197
Merge pull request #10164 from Yancey1989/lookup_sparse_table_op
...
add lookup_sparse_table_op
8 years ago
3d846fc3f2
Make Variable support for future.division.
8 years ago
6a378251dc
Merge remote-tracking branch 'helin/word2vec' into feature/new_api_train_impl
8 years ago
9a8be9daca
Merge pull request #10223 from pkuyym/fix-10219
...
Change `customize_loss_grad` to `use_default_grad_scale`.
8 years ago
a785a837b9
update the example with the latest API
8 years ago
b5dd215d46
improve comments
8 years ago
3eef539a42
add word2vec test for the new API
8 years ago
392a9dd9a2
fix yapf stype check
8 years ago
10cb9424fd
remove enum reference from trainer
8 years ago
6d58d6dc48
add comments
8 years ago
a2ffbd5326
scaffolding for the new Fluid API
8 years ago
fb7ca48c06
Add image classification unit test using simplified fluid API ( #10306 )
8 years ago
95d2651bc2
Build: simplify travis CI script. ( #10245 )
...
* Build: simplify travis CI script.
* Add linkcheck and run gen_doc in host machine not in docker.
* Add LinkCheck python package.
8 years ago
eb6f9dd5de
Feature/cuda9 cudnn7 ( #10140 )
...
* "re-commit "
* "picked up"
* "fix ci"
* "fix pdb hang up issue in cuda 9"
8 years ago
f61dfeedcc
Merge pull request #10263 from chengduoZH/add_FLAGS_use_deterministic_algo
...
Add FLAGS_cudnn_algo_use_autotune
8 years ago
4434f8b4bd
Use the OrderedDict in the framework.py for the variable map. ( #10282 )
8 years ago
9fda5c92cd
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into add_FLAGS_use_deterministic_algo
8 years ago
f456cd8079
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into fix-10219
8 years ago
c5774e3282
add FLAGS_use_deterministic_algo
8 years ago
ac2911c399
fix code style
8 years ago
19a28b1611
[fix assign op]/2 using assign op instead of scale1.0
8 years ago
8806801eb8
[fix assign op] using assign op instead of scale1.0
8 years ago
1a93253f16
fix unittest
8 years ago
98a2179f80
fix ListenAndServ op
8 years ago
4a5bfa89c3
Modify RoI pooling op to use LoDTensor and expose it into Python API ( #10208 )
...
* modify roi pool with lod and expose ROI Pooling into Python API
* make lod code brief
* make doc more clearly
* make doc more clearly
8 years ago
87d86ee34e
Merge pull request #10247 from putcn/fix-transpiler
...
add missing grad_to_block_id
8 years ago
0ecc6fa8f3
Add float16 transpiler and image classification example ( #10109 )
...
* add float16 transpiler
* fix feed fetch target names mismatch
* fix cast op input change issue
* fix program desc flush error
* fix inconsistent var names in block desc bug
* code clean up
* add float16 infernce C++ example and fix prune bug
8 years ago
b133425915
add grad_to_block_id
8 years ago
08fbab9429
complete ListenAndServ layer in layers/io.py
8 years ago
c816121d11
optimized iou_similarity_op ( #10231 )
8 years ago
63bd38bd74
code optimize
8 years ago
bf824d854a
Refine doc.
8 years ago
fb1167c315
delete unused comment
8 years ago
dccd013bd3
refine distribute transpiler
8 years ago
82571deb89
Change `customize_loss_grad` to `use_default_grad_scale`.
8 years ago
e393c86c4a
Merge branch 'develop' of github.com:PaddlePaddle/Paddle into lookup_sparse_table_op
8 years ago
25d5e226ab
Fix typo in wmt16
8 years ago
63bf82ddea
Merge branch 'develop' into add-async-listen-and-serv-op
8 years ago
72ee737f3f
Merge pull request #9308 from wangyang59/bilinear
...
Bilinear interp op
8 years ago
2182ecfbbd
remove duplicated ShareLoD in gru_op and sequence_conv_op ( #10149 )
...
* remove share lod; it has already been performed in infershape
* slightly release test bound of test_image_classification
8 years ago
4ecc9b7bae
Merge pull request #10166 from reyoung/feature/train_and_test_recordio
...
Add demo for recordio train/test and parallel executor
8 years ago
bfbbe19fbb
Merge pull request #10150 from chengduoZH/fix_elementwise_gradient
...
Fix elementwise_gradient bug
8 years ago
0f5d5b1ffc
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into fix_elementwise_gradient
8 years ago
54ada9449e
Add demo for recordio train/test and parallel executor
8 years ago
44fa823841
Merge pull request #9949 from mozga-intel/mozga-intel/Mul_mkldnn
...
Initial implementation of multiplication operator for MKLDNN
8 years ago
e8d802159e
add lookup_sparse_table_op
8 years ago
d06c79c7a7
fix elementwise_grad op kernel and add unit test
8 years ago
d67b9ced05
Merge pull request #9820 from seiriosPlus/quick_help
...
fix quick start for fluid #9660
8 years ago
bf021f3432
resize test
8 years ago
469a349ae3
polishing after qingqing's comments
8 years ago
f67f0cae50
finished testing cpu bilinear_interp_op
8 years ago
0881d809cd
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into add-async-listen-and-serv-op
8 years ago
618db96ee7
fix a cpu bug in parallel_executor.py
8 years ago
a0ced3df82
async update can run
8 years ago
6e8510fa1d
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into quick_help
8 years ago
7a395881d4
Add customize_loss_grad option to PE
8 years ago
34f2818549
distribute transpiler support async config
8 years ago
63055a3e08
complete grad_to_id
8 years ago
39892feef2
Merge branch 'split-optimize-op-into-signle-blocks' of ssh://github.com/jacquesqiao/Paddle into add-async-listen-and-serv-op
8 years ago
1b5de9d4bd
Merge branch 'split-optimize-op-into-signle-blocks' of ssh://github.com/jacquesqiao/Paddle into add-async-listen-and-serv-op
8 years ago
ba1e68d5aa
clean code
8 years ago
18e0b7303c
fix _create_table_optimize_block
8 years ago
39f6274ecc
follow comment, optimize code
8 years ago
5d32008c3f
Merge branch 'split-optimize-op-into-signle-blocks' of ssh://github.com/jacquesqiao/Paddle into add-async-listen-and-serv-op
8 years ago
e05f4dfb49
add some comment
8 years ago
7907b6a959
split optimization ops on pserver to independenty blocks
8 years ago
d89a30680e
Fix a wrong variable name. ( #10090 )
8 years ago
c6937abdd1
tmp
8 years ago
ad91bfe5d4
fix a bug in test_batch_norm_op.py ( #10094 )
8 years ago
7c194acce4
Merge pull request #9929 from JiayiFeng/make_for_parallel_default_true
...
make for_parallel default True
8 years ago
f738691777
Merge pull request #9740 from dzhwinter/memory/activation
...
"polish activation"
8 years ago
8e00540778
Merge pull request #10058 from Xreki/core_fix_flush
...
Add flush of program desc to update the proto information.
8 years ago
92b5f49ee4
Eager destruct local scope
8 years ago
84ceffd02d
Fix api display errors in fluid ( #10051 )
8 years ago
7ffbcbcaf0
Add flush of program desc to update the proto information.
8 years ago
879b7c5601
Merge pull request #10049 from typhoonzero/fix_not_trainable_transpiler
...
Skip updating not trainable parameters in distribute transpiler
8 years ago
e6745be9ea
fix_not_trainable_transpiler
8 years ago
6db5309b56
`make install` does not remake whl package if only change .cc or .h ( #10030 )
...
`make install` does not remake whl package if only change .cc or .h
8 years ago
129b913cff
Fix the bug in executor.fetch_var
8 years ago
c6fc6a3603
fix a bugs
8 years ago
1fc9fe773d
merge develop
8 years ago
598035f985
Fix a bug in save_inference_model and prune when the program is initailized by load_inference_model ( #10011 )
...
* Fix bug in save_inference_model and prune when the program is initialized by load_inference_program.
* Save the transpiled program instead.
8 years ago
7a517dc93e
merge develop
8 years ago
1fd12840e7
Merge pull request #9827 from sharannarang/sentiment_analysis
...
Add cutoff parameter to word_dict for IMDB
8 years ago
90abd9dfeb
Merge pull request #9704 from lgone2000/fixunittest
...
fix pool and conv3d unittest when paddle is not build with gpu
8 years ago
d655417f90
Merge pull request #9956 from typhoonzero/split_byref_op
...
Split byref op
8 years ago
e69de096d0
remove redundent pserver block
8 years ago
0dff1e4140
Fix bugs
8 years ago
171471eada
Merge branch 'develop' into mozga-intel/Mul_mkldnn
8 years ago
35483a2094
Add neural transformer leanring rate decay function. ( #9951 )
...
Add neural transformer leanring rate decay function
8 years ago
ee9832a7a9
Add Top-k Python API. ( #9973 )
...
* Add topk Python API.
* Add unit test.
* Remove the repeated API.
8 years ago
c083ee703c
Merge pull request #9950 from JiayiFeng/add_parallel_executor_tests
...
Add parallel executor tests
8 years ago
425a1e76a2
Merge remote-tracking branch 'origin/develop' into memory/activation_memory
8 years ago
61f4baa13a
Merge pull request #9792 from luotao1/merge_bn
...
fuse batch norm for conv operator
8 years ago
2617d8bcda
Merge pull request #9993 from jczaja/prv-softmax-mkldnn-fix
...
- Added Epsilon (preventing softmax output from being too small) for softmax MKLDNN op
8 years ago
e2d56832e1
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into split_byref_op
8 years ago
e84d3a7fbf
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into add_parallel_executor_tests
8 years ago
8352f93809
Merge pull request #9933 from typhoonzero/add_dist_unit_test
...
Add dist unitest data compare, ensure that dist train have same behavior as local training
8 years ago
de8094f57e
Cosmetic fixes
8 years ago
acdf7cbd19
- Added EPS for softmax MKLDNN op
...
- EPS added to softmax mkldnn primitive outcome is limited to training
phase
Fixes after review
clang format fixes
clang format fixes
8 years ago
81c47b21ef
add type check and default scope
8 years ago
69188e5981
fix ut
8 years ago
c6d7c2bdff
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into split_byref_op
8 years ago
01b88f2177
Merge branch 'develop' into merge_bn
8 years ago
6e7b883bdd
Initial implementation of multiplication operator for MKLDNN
8 years ago
fee5b24cb4
Merge pull request #9985 from mozga-intel/mozga-intel/mkldnn_separate_tests
...
The mkldnn tests are skipped when the MKLDNN flag is OFF
8 years ago
ed89b7b7e6
dist train use split_by_ref
8 years ago
22f58382bc
run dist train test single process
8 years ago
fc6f0be276
Merge pull request #9942 from reyoung/feature/tuning_pe_trans
...
Feature/tuning pe trans
8 years ago
1de9edee51
Follow comments
8 years ago
93276fd146
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into add_parallel_executor_tests
8 years ago
0c6eef3e58
add split by ref test
8 years ago
ab86fb11c4
complete parallel accuracy test
8 years ago
5e13314dc1
Commit
8 years ago
7286954337
Add more comments
8 years ago
e05f3eaf69
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into quick_help
8 years ago
a1d910b144
Fix line witdth
8 years ago
b4aaa00a8a
Polish logic of ParallelExecutor
8 years ago
c0f4e0332c
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into feature/simplify_delay_logic
8 years ago
3fbe9c34f7
Fix typo in docstring ( #9829 )
8 years ago
d78cab1519
update
8 years ago
f086ccb7b6
Merge branch 'develop' of https://github.com/PaddlePaddle/paddle into add-python-label-smooth
8 years ago
415460b557
stash
8 years ago
e412b1aec7
Merge branch 'unify_executor_interface' into add_parallel_executor_tests
8 years ago
22df230ee4
rename 'feed_dict' in ParallelExecutor.run() to 'feed'
8 years ago
fbd5cf604f
stash
8 years ago
5fc8326732
Add parallel accuracy test
8 years ago
d04073050e
Add python wrapper for label smoothing
8 years ago
745aacfc38
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into add_dist_unit_test
8 years ago
c4de344ce2
update
8 years ago
2b05b010ff
fix chunk evaluator name in metrics
8 years ago
34cffe9169
add unitest file
8 years ago
8c1b257596
add dist unitest data compare
8 years ago
d0342f12aa
Simplify DelayOps Logic
8 years ago
e54f203c55
"move to a new PR"
8 years ago
1b5a6351af
make for_parallel default True
8 years ago
2f38264078
Merge pull request #9905 from panyx0718/mem-opt
...
Polish memory optimization transpiler
8 years ago
d4024a6ebd
Polish memory_optimizer code with mode comments and less identions
8 years ago
f22da580fa
fix compiler error in paddle:latest-dev image ( #9907 )
8 years ago
e4cfe477db
Merge pull request #9898 from reyoung/feature/mix_cpu_gpu_op
...
Feature/mix cpu gpu op
8 years ago
9bc0c23b17
Merge pull request #9901 from Yancey1989/fix_duplicate_lr_op
...
Fix duplicate lr op after distribute transpiler
8 years ago
ec512cdcce
add comment for branch network
8 years ago
c1bf06f9d9
Merge pull request #9289 from dzhwinter/speed/sequence_expand
...
Speed/sequence expand
8 years ago
ed2d7d7d26
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into feature/mix_cpu_gpu_op
8 years ago
253441b553
fix duplicate lr op after distribute transpiler
8 years ago
482314e3b1
Add CRF unittest
8 years ago
f45818e7f9
create new varible in scope
8 years ago
41a9146de3
Merge pull request #9787 from Yancey1989/random_selected_rows_value
...
Initialize large table value randomly
8 years ago
1bdc7261a5
Merge pull request #9578 from typhoonzero/threadpool_for_io
...
Multi stream thread pool
8 years ago
6e735e1e6e
Merge branch 'develop' into merge_bn
8 years ago
d24b5e060f
The fully connected: the operator is removed when the MKLDNN flag is OFF
8 years ago
ddf5783644
Merge pull request #9812 from jshower/develop
...
Changing network configuration, avoid nan
8 years ago
a006c54c11
fix code style #9660
8 years ago
e5113da39c
fix code style #9660
8 years ago
5ee67f15bf
fix code style #9660
8 years ago
2dfd8841a4
make quick start code more simpler #9660
8 years ago
4c55a6022a
Dist transpiler support prefetch ( #9714 )
...
* init
* add some check
* add dist transpile logic
* add insert op for block
* init change get_pserver_program
* optimize code
* fix a bug
* can run now
* start to do table split
* start to process table gradient
* complete pserver part
* can send_vars now
* revert cpplint
* fix a bug
* optimize code
* move dist test to models
* revert the interface of distribute_transpiler.transpile
* fix prefetch_block
* optimize trainspiler code
* add comment to sum_op
* add warning log
* fix comment
* fix test_send_recv
* fix test_send_recv
* fix train with no distributed table
* optimize GetDims
8 years ago
26cfc634b9
multi stream thread pool
8 years ago
7132bbe6b7
update by comment
8 years ago
ef169eb996
Merge pull request #9459 from wanghaoshuang/fix_avg
...
Make Average Model support for 'moving mean' and 'moving variance' of batch_normal op
8 years ago
648908004e
Merge pull request #9849 from qingqing01/parallel_exe_test
...
Improve test_parallel_executor.
8 years ago
a64edbf14e
delete backward.cc related code on the python side ( #9854 )
8 years ago
b26f505002
remove net op and cond_op ( #9663 )
...
* remove net op and cond_op
* fix cpplint
* fix dependency
* delete backward_test; fix compile
* disable batch_norm backward
* rm test_net.py
* make batchnorm test independent of backward.cc
* make test_layer_norm_op independent of backward.cc
* make test_layer_norm_op independent of backward.cc
* delete unused code
* clean up
8 years ago
09c1c692f3
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into sentiment_analysis
8 years ago
554cea6839
move most codes to uci_housing #9660
8 years ago
9fd76c4f3a
Improve test_parallel_executor.
8 years ago
652cf43002
Merge pull request #9746 from typhoonzero/multigpumultinode
...
[Feature] Enable multi gpu distributed training of fluid
8 years ago
2a3d49040c
Merge pull request #9786 from weixing02/initializer
...
Fix api docs display error for fluid Initializer
8 years ago
1aada35206
Merge branch 'develop' of github.com:PaddlePaddle/Paddle into random_selected_rows_value
8 years ago
d52fa26fda
Feature/metrics ( #9791 )
...
* "add metrics"
* "add fluid metrics"
* "add import guards"
* "show warnings"
* "add demo"
* "fix ci"
* "add some details"
* "fix cci"
* "add demo Python"
* "add metrics"
8 years ago
dfc60251bf
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into multigpumultinode
8 years ago
90084a25d2
Merge pull request #9743 from JiayiFeng/modify_readers_to_fit_parallel_executor
...
Modify readers to fit the parallel executor
8 years ago
7815cdffba
use clone method to flush in forth
8 years ago
996b3e15fa
Merge remote-tracking branch 'origin/develop' into speed/sequence_expand
8 years ago
8c1eb8693e
update unittest
8 years ago
0dacbbe1fe
update multi_pass_reader unittest
8 years ago
b1cc28dab3
update
8 years ago
273f4892b2
update recordio unittest
8 years ago
75a2e50581
Merge branch 'modify_dev' into modify_readers_to_fit_parallel_executor
8 years ago
72b5de05fe
update unittest
8 years ago
3f320c1345
Merge branch 'develop' into merge_bn
8 years ago
a7c6bf771c
Change do_model_average_for_mean_and_var to boolean in batch_normal.
8 years ago
0ad892a0d4
Merge pull request #9816 from luotao1/remove_op
...
add remove_op, remove_var in Python end
8 years ago
ec6976810d
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into multigpumultinode
8 years ago
fc346e3432
Add cutoff parameter to build_dict.
8 years ago
a84b81502c
Remove Readers' HasNext()
8 years ago
ce08dc8751
have stream removed error
8 years ago
5483258b7e
fuse batch norm for conv operator with bias
8 years ago
ad6ddf533c
for ci
8 years ago
0bf799a523
wip testing
8 years ago
284a213774
fix a name conflict
8 years ago
7c1434dd73
code style
8 years ago
5ecfb62de7
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into initializer
8 years ago
e0babe7c3e
Merge pull request #9780 from chengduoZH/feature/fix_batch_size_is_littler_than_gpu_count
...
Crash training, if the number of samples is less than the count of devices.
8 years ago
e7467d94a0
add remove_op, remove_var in Python end
8 years ago
56b04e5b2c
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into initializer
8 years ago
93940642ac
Adjust
8 years ago
cb7bbf426c
revert uniform_random_op
8 years ago
291aa23153
Merge branch 'random_selected_rows_value' of github.com:Yancey1989/Paddle into random_selected_rows_value
8 years ago
1cc09c7bc5
Merge branch 'develop' of github.com:PaddlePaddle/Paddle into random_selected_rows_value
8 years ago
3f6fc10b9f
new op that init table value randomly
8 years ago
ea0cf6f382
rewrite inference_transpiler in Python end
8 years ago
d9a5222385
code style
8 years ago
b44b6a4f00
Merge pull request #9798 from typhoonzero/fix_dist_transpiler_bug
...
Fix dist transpiler bug
8 years ago
ee178d5aeb
fix bugs
8 years ago
16e31343d8
Merge branch 'develop' into merge_bn
8 years ago
49d431d44c
Merge pull request #9807 from typhoonzero/fix_python_version_module
...
fix python package have no version.py
8 years ago
9fe938cb2a
Changing network configuration, avoid nan
8 years ago
71081a9bb9
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into initializer
...
Merge branch develop
8 years ago
0f38bb4593
add fp16 support to activation op ( #9769 )
8 years ago
86a32216eb
fix python package have no version.py
8 years ago
326b434ab5
Merge pull request #9699 from panyx0718/test
...
Avoid init_p2p all the times
8 years ago
92313a99c0
update
8 years ago
d02b17e597
fix dist transpiler bug
8 years ago
919aa2ea0a
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into initializer
...
Merge branch develop
8 years ago
cd41dca2ab
Adjust
8 years ago
17833d30e3
fuse batch norm for conv operator without bias
8 years ago
fe48dfcb6e
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into modify_dev
8 years ago
bbbc3592db
Merge pull request #9735 from panyx0718/dist
...
Add some comments for distribute_transpiler
8 years ago
759979998c
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into modify_dev
8 years ago
f909ff1a36
update unit test
8 years ago
0586b9e5c9
Add *Initializer to fluid api docs
8 years ago
5416bac5d8
Make shared decorated readers' creater be only in main_program
8 years ago
7e7611d067
when the number of samples of current batch is less than the count of devices, let it crash.
8 years ago
31464f3425
Merge pull request #9772 from reyoung/feature/add_gitignores
...
Add gitignores
8 years ago
a4bf635795
Add gitignores
8 years ago
3f90a583b4
update unittest
8 years ago
055662f9f4
add insert_op for block
8 years ago
2b7e5bd366
Support testing during training by ParallelExecutor. ( #9738 )
...
* Support testing during training by ParallelExecutor.
* Add unit test.
* Improve the interface.
* Follow comments.
8 years ago
75c9eb11c4
refine
8 years ago
9b3c964341
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into modify_readers_to_fit_parallel_executor
8 years ago
ec89ed69ef
Merge pull request #9655 from kbinias/kbinias/mkldnn-activations-dim2
...
Added support for dim equals 2 in activation functions
8 years ago
c1b6692fe5
Fix debuger bugs. ( #9705 )
...
Fix debuger bugs
8 years ago
5ad2486905
fix errors
8 years ago
4459fa6726
fix seq tag model dist transpile
8 years ago
fca9e8847d
Update Readers Python API
...
1. Combine 'open_files', 'multi_pass_reader' and 'threaded_reader'
together to make the new 'open_files' interface.
2. Add some docstring.
3. Simplify interface names of 'create_XXX_reader', e.g, rename
'create_double_buffer_reader' to 'double_buffer'.
8 years ago
6be51f10ec
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into modify_readers_to_fit_parallel_executor
8 years ago
bdb47cd933
Add some comments for distribute_transpiler
8 years ago
38ba7e5af4
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into update_reader
8 years ago
de6802d2bd
fix missing core.so on mac
8 years ago
6651301b29
Modify comments of `sync_with_cpp`
8 years ago
3682035f14
Merge pull request #9695 from emailweixu/round_robin
...
Fix a minor bug for distributed_spliter.round_robin
8 years ago
9d95ccc887
fix pool and conv3d unittest when paddle is not build with gpu
8 years ago
6ba262572c
fix test_conv2d_op when compile without cuda ( #9698 )
8 years ago
b138d29c38
Avoid init_p2p all the times
8 years ago
560d960b27
Fix a minor bug for distributed_spliter.round_robin
...
Also fixed typo and comments.
8 years ago
b2a1c9e8b7
Add float16 support to non-cudnn softmax op on GPU ( #9686 )
...
* initial commit
* fix error
* fix typo and order
8 years ago
3d3e5aeb9c
Merge pull request #9648 from chengduoZH/feature/expose_CUDAPinnedPlace_to_python
...
Expose CUDAPinnedPlace to Python
8 years ago
0935f2cf05
shrink test_parallel_executor size.
8 years ago
09b4a1a361
Build: generate all the build related files into one directory. ( #9512 )
8 years ago
ff4208e676
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into sgd-support-update-selected-rows
8 years ago
17842e33c3
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into feature/expose_CUDAPinnedPlace_to_python
8 years ago
043c23080c
Merge pull request #9637 from panyx0718/feed
...
Add feed support for ParallelExecutor
8 years ago
8e4e155c52
add PyCUDAPinnedTensorSetFromArray
8 years ago
92e92ceba1
Add type check
8 years ago
c4107748d0
Add support for dim equals 2 in activation functions
8 years ago
2e40660e7a
Fix some issues.
8 years ago
5eb9cecc4a
Merge pull request #9607 from luotao1/remove_var
...
add remove_var from c++ end
8 years ago
c00a5dec78
Merge pull request #9644 from typhoonzero/fix_condition_op_in_optimize
...
Fix transpiler condition op in optimize
8 years ago
a84a580e65
Add CUDA kernel for prior_box_op. ( #9553 )
8 years ago
a16a872783
update
8 years ago
6dcfd97a92
add docstring
8 years ago
e0b396e7ba
update by comment
8 years ago
af24290123
Add 'buffer_size' api for open_files op
8 years ago
39277e9282
fix transpiler condition op in optimize
8 years ago
4bbfa9eccb
Add feed to ParallelExecutor
8 years ago
a98a3fdc46
Merge pull request #9385 from mozga-intel/mozga/mkldnn-fc
...
Implementation of MKLDNN FC
8 years ago
d02a68c447
Merge pull request #5768 from NHZlX/add_upsample_layer
...
Add upsample layer
8 years ago
19db989ec2
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into fix_avg
8 years ago
09b53c086d
add remove_var from c++ end
8 years ago
32f8ac7d3b
Remove additional message
8 years ago
34a8084328
Added new fc files, register fc kernel
8 years ago
2811ea4440
Implementation of MKLDNN FC
8 years ago
cbfec1f7d6
diable test of sparse_parameter_sgd on GPU
8 years ago
957104564f
Merge pull request #9600 from luotao1/sync_with_cpp
...
refine sync_with_cpp when remove ops or remove vars
8 years ago
49313d4066
Merge pull request #9548 from panyx0718/group_nccl_all_reduce
...
Group nccl all reduce and improve performance (~14% for 4 device resnext)
8 years ago
3c370ee0d5
Merge branch 'develop' of https://github.com/PaddlePaddle/paddle into add-python-pad
8 years ago
103407aa60
refine sync_with_cpp when remove ops or remove vars
8 years ago
cf251eb8cf
shrink test size
8 years ago
af1d3f5bc0
remove print
8 years ago
a994327fb1
add TestSGDOpOptimizeSelectedRows
8 years ago
de2d82d6a9
fix a bug
8 years ago
b25a9c488d
Reduce test size to avoid GPU memory issue.
8 years ago
b123ce88a1
Add enable/disable for delayed ops
8 years ago
46f3a39e91
polish and add comments.
8 years ago
d0ac92531d
Improve ParallelExecutor performance
8 years ago
abb7deee39
optimize test_lookup_table_op.py
8 years ago
649ae2700e
fix bugs
8 years ago
c72450d24d
"seperate test"
8 years ago
44d5f42a7e
update reader
8 years ago
e64dda7e44
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into sgd-support-update-selected-rows
8 years ago
b154f311b9
init TestSGDOpOptimizeSelectedRows
8 years ago
37d9a72e9e
Merge pull request #9575 from jacquesqiao/lookup_table_support_SelectedRows_as_parameter
...
Lookup table support selected rows as parameter
8 years ago
13ecb5e50a
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into lookup_table_support_SelectedRows_as_parameter
8 years ago
f801f743d1
fix the __all__ in dataset package
8 years ago
f837eee724
add paddle.v2.reader,dataset back for backward compatibility
8 years ago
0b8534f2a4
Refine python wrapper for pad_op
8 years ago
f02968bb21
disable test_recv_op
8 years ago
2be10ebe8a
disable test_recv_op
8 years ago
7a6ffb6280
add TestLookupTableWIsSelectedRows
8 years ago
62ba8723ac
Merge remote-tracking branch 'origin/develop' into speed/sequence_expand
8 years ago
629c9210d1
Merge branch 'develop' of https://github.com/PaddlePaddle/paddle into add-python-pad
8 years ago
4c8eef5739
Add python wrapper for pad_op
8 years ago
b9d8bbe4f2
Merge pull request #9558 from typhoonzero/fix_dist_transpile_one_trainer
...
fix single pserver single trainer error
8 years ago
d908c3b2fa
Merge pull request #9008 from lcy-seso/enhance_reshape
...
Enhance reshape
8 years ago
772cdfe196
fix single pserver error
8 years ago
6cfc0c1497
"polish code" ( #9318 )
...
* "polish code"
* "fix ci"
* "fix ci"
* "done"
8 years ago
61f4706b1d
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into fix_avg
8 years ago
9708b21f19
Refine average model option
...
1. Add attr 'average' into ParamAttr.
2. Make 'params_grads' optional for AverageModel.
3. Add option 'average_mean' and 'average_variance' for batch_normal.
8 years ago
bcf7c36b0b
Make paddle.fluid no longer depends on paddle.v2
...
In this way we can build and test using WITH_FLUID_ONLY flag being set
to ON.
- move paddle.v2.dataset,reader to paddle.dataset,reader
- remove unused code (which depends on v2) in paddle.dataset,reader
8 years ago
95658767eb
Merge pull request #9428 from JiayiFeng/kernel_of_increment_op
...
kernels of IncrementOp
8 years ago
374f1ca3b7
Fix dist error with lr decay layer ( #9489 )
...
Fix dist error with lr decay layer
8 years ago
d1a7b47e04
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into fix_avg
8 years ago
fa21436d0d
Merge pull request #9080 from reyoung/cpp_parallel_executor
...
Cpp parallel executor
8 years ago
9bbd753425
change WITH_FLUID to WITH_FLUID_ONLY ( #9427 )
8 years ago
af230d9bef
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into cpp_parallel_executor
8 years ago
663f0358a6
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into add_upsample_layer
8 years ago
34a440fa64
Revert "make append activation in place by default ( #9417 )"
...
This reverts commit ce16400dae
8 years ago
a80bf702f3
Merge remote-tracking branch 'origin/develop' into speed/sequence_expand
8 years ago
b661fe1d76
"fix ci"
8 years ago
8425c2c859
Speed/sequence op1 ( #9217 )
...
* "add functors"
* "remove old code"
* "fix"
* "fix ci"
* "add details"
* "fix ci"
* "fix ci"
* "fix ci"
* "fix ci"
* "remove unused code"
8 years ago
5b8bb34470
Refine reshape_op by following comments.
8 years ago
24100e1fb8
Merge pull request #9449 from chengduoZH/feature/add_cos
...
Add cos and sin
8 years ago
869ef01d66
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into kernel_of_increment_op
8 years ago
52574733a6
Add KernelType switch for IncrementOp kernel
8 years ago
bdda08d9f2
add sin
8 years ago
01c5ca7364
fix bugs
8 years ago
ce16400dae
make append activation in place by default ( #9417 )
8 years ago
917b205c1c
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into kernel_of_increment_op
8 years ago
2e577379ca
add cos
8 years ago
764723d4f5
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into fix_avg
8 years ago
e1290c4fd7
Make Average Model support for 'moving mean' and 'moving variance' of batch_normal op
8 years ago
857a8997de
Merge pull request #9384 from luotao1/removeVar
...
remove vars when remove ops
8 years ago
4747b2c92d
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into add_upsample_layer
8 years ago
0be1e09f2c
"fix ci"
8 years ago
b0775588c0
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into cpp_parallel_executor
8 years ago
055fb215a1
remove unnecessary 'force_cpu'
8 years ago
f2d29be784
Disable transformer
8 years ago
09743b6117
Refine test_reshape_op
8 years ago
5408854090
Disable model evaluation in unittests
8 years ago
f34f2d4026
make bn inplace in img_conv_group by default
8 years ago
54a8c04fab
add inplace attr to bn
8 years ago
4bfbc59122
Merge branch 'develop' of https://github.com/PaddlePaddle/paddle into enhance-ReshapeOp
8 years ago
c078ed4608
Enhance reshape_op by adding Input(Shape)
8 years ago
7f4012247e
adjust remove rule for variables
8 years ago
c42c4a6718
Add performance tests
8 years ago
9af870854e
Use heap variables
8 years ago
5447046aee
merge develop branch
8 years ago
123cf165fb
Set stop_gradient=True for some variables in SSD API. ( #9396 )
8 years ago
fe142d92ff
Merge pull request #9359 from wanglun/develop
...
Fix typo of Softmax document
8 years ago
1b67bc022c
Merge pull request #9329 from tpatejko/tpatejko/mkldnn-lrn
...
Improvements for MKLDNN LRN
8 years ago
c7bf77d0e1
Add in is_copy attribute to SelectCase. ( #9393 )
...
This is a temporary solution to allowing for variables to be copied during a channel send operations. Also fixed issue with is_copy for "channel_send" method, and also updated unit tests.
8 years ago
65534c4762
Fluid channels should match the semantics of Go Channels ( #9265 )
...
* Fluid Channel should match Go Channel in Semantics
* Fix Python channel_send
* Address code rveiew feedback
* Fix open_files_op.cc
* Add description to Channel Asserts
8 years ago
ccfec1bcb1
remove vars when remove ops
8 years ago
f3dc3112cc
add split ids op ( #9370 )
...
* add split_ids_op
* add TestSplitIdsOp
* fix comment
* add test for empty tensor
* clean code
* rm unused code
8 years ago
9dd64d83f3
WMT Model
8 years ago
cb40c33137
Update unittest
8 years ago
f765e59c41
Merge pull request #9327 from Yancey1989/fix_test_recv_op
...
Fix test_recv_op
8 years ago
54a85b7bfd
Add lrn layer ( #9157 )
...
* add LRN layer for fluid
* add LRN layer for fluid
* add documentation for LRN layer
* add paper reference for LRN layer
* add seperate documentation for LRN layer
* rm lrn.py in doc/fluid/dev/src
* change code style in lrn
* fix style of comments in lrn
8 years ago
54bd17fe7b
Complete Flowers
8 years ago
50e7e25db3
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into cpp_parallel_executor
8 years ago
5c7a523326
Add Graphviz output
8 years ago
cb3bbbd5c6
Merge pull request #9081 from kbinias/kbinias/mkldnn-activations
...
MKLDNN Relu Tanh Sqrt Abs activations added
8 years ago
8ccc61f334
support empty tensor ( #9338 )
...
* support empty tensor
8 years ago
30f1bd6a64
add the recordio in creator.py and change the " to \' ( #9358 )
8 years ago
f96f2860f9
Fix typo of Softmax document
8 years ago
edfd741e3a
Add simple python wrapper for ParallelExecutor
8 years ago
6461e800a5
Inheritance added for MKLDNN tests
8 years ago
a64b312e3a
Correcting for PR comments
8 years ago
4466f0bec8
MKLDNN Relu Tanh Sqrt Abs activations added
8 years ago
9075049a29
add unit test
8 years ago
76ae540f8e
Move Select to concurrency.py; incorporate outputs ( #9136 )
...
* Move Select to concurrency.py; incorporate outputs
* CLang formatting for concurrency
* Remove extra bracket - formatting fix - 3
* Comment fix
8 years ago
9c35b0dc1b
Merge pull request #9287 from typhoonzero/pserver_prepare_before_run
...
Pserver prepare before run
8 years ago
14ba67c0ef
Function for running MKLDNN primitive added. Unittest added for is_test attribute
8 years ago
6cebbd7bcb
update
8 years ago
ba9f4c7873
fix test_recv_op
8 years ago
e027eb40d7
Merge pull request #9123 from tpatejko/tpatejko/mkldnn-lrn
...
Implementation of MKLDNN LRN
8 years ago
9367f11eb7
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into pserver_prepare_before_run
8 years ago
b7e83d2467
Merge branch 'develop' of https://github.com/PaddlePaddle/paddle into enhance-ReshapeOp
8 years ago
8f8728635a
Fix bug for backward tanspiler when using parallel_do operator. ( #9282 )
...
* Temporarily fix bug for backward tanspiler when using parallel_do operator.
* Fix bug for backward tanspiler when using parallel_do operator
8 years ago
972a102b92
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into pserver_prepare_before_run
8 years ago
5e6276edc1
fix transpiler bug
8 years ago
d4bb2ca71f
Follow comments and refine the python wrapper of reshape_op
8 years ago
dc07d52e31
Merge pull request #9314 from chengduoZH/v2_fix_doc/pool
...
Fix v2_pooling doc
8 years ago
990d6396fe
Reuduce memory copy when communication between trainer and pserver. ( #9271 )
8 years ago
8440046b7f
fix doc
8 years ago
b594251f89
Merge pull request #9082 from wanghaoshuang/average_model
...
Add model average optimizer for fluid
8 years ago
64c5c8f8b0
Merge pull request #9269 from kexinzhao/softmax_cudnn_fp16
...
Add float16 support to cudnn softmax kernel
8 years ago
b9e6364e3c
Merge pull request #9267 from kexinzhao/new_relu_fp16
...
Add float16 support to relu op
8 years ago
4eaa789730
resolve conflict
8 years ago
7c14724270
Add default value of keyword argument to DocString ( #9262 )
8 years ago
53c8c36a04
"debug the process"
8 years ago
7260e3a443
Merge pull request #9214 from jczaja/prv-softmax-mkldnn-operator-PR
...
Softmax MKLDNN FLUID operator
8 years ago
437f7a3279
Resolve conflict according to the latest code
8 years ago
3b95b55f07
- Softmax MKLDNN primitive integration
...
removed diagnostic
- Added Unit tests for Softmax MKLDNN Forward
Added fix for div by 0 to happen in cross_entropy backward
Conflicts:
paddle/fluid/operators/CMakeLists.txt
- Cosmetic fixes to SoftMax MKLDNN fluid operator
Added misssing softmax fluid operator file
Disabled MKLDNN softmax operator by default
Fix to softmax op unittest merge
clang_formater fixes
clang_formatter fixes
- Name changing of softmax mkldnn operator to maintin consistency
across codebase
- updated comment
fix to comment
8 years ago
ed2bc194c5
Merge pull request #9176 from kexinzhao/batch_norm_fp16
...
Add float16 support to batch norm operator
8 years ago
809530f418
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into dev_MultiEpochReader
8 years ago
7c041e48f4
Merge pull request #9182 from JiayiFeng/dev_MultipleReader
...
Multi-threaded reader in C++
8 years ago
edb4e29ab7
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into average_model
8 years ago
70e7122785
initial commit
8 years ago
018f3bda3d
small fix
8 years ago
d60180af39
inital commit
8 years ago
e0bc942ca1
Merge pull request #9224 from wangyang59/dilateconv
...
expose dilation to conv2d and bias/act to con2d_trans
8 years ago
c1e9b1e37e
Merge pull request #9231 from kexinzhao/elementwise_add_fp16
...
Add float16 support to Elementwise Add op
8 years ago
37a272e670
add executor.prepare ( #9022 )
...
optimize executor.run
8 years ago
a944d57181
refine code
8 years ago
a4f397fb68
add an unittest
8 years ago
30b70323b4
Expose RMSProp optimizer. ( #9247 )
...
* Add RMSProp optimizer warpper.
* Follow comments.
8 years ago
91b6d60003
Merge branch 'fix_bug_in_recordio' into dev_MultiEpochReader
8 years ago
bce08d19cc
Python wrapper for MultiPassReader
8 years ago
2532b922dc
Add more unittests and fix bugs
8 years ago
ad63722ed9
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into average_model
8 years ago
68c9f6ef11
Fix error while params_grads[1]==None
8 years ago
f863866471
Add an unitest
8 years ago
5008020d19
Merge pull request #9154 from typhoonzero/pserver_parallel
...
pserver runs in parallel
8 years ago
02b7d8bea5
Merge branch 'fix_bug_in_recordio' into dev_MultipleReader
8 years ago
d806fd4114
no print
8 years ago
7e526a6cb8
Merge pull request #9213 from wanghaoshuang/adadelta
...
Add python wrapper for Adadelta optimizer
8 years ago
381c6a026d
Merge pull request #9100 from pkuyym/fix-9049
...
Enhance sequence_expand operator
8 years ago
d307b5e4a6
Merge remote-tracking branch 'upstream/develop' into elementwise_add_fp16
8 years ago
5271c32d24
Merge pull request #9223 from kexinzhao/dropout_fp16
...
Add float16 support to dropout operator
8 years ago
d2cb3790e9
Wait all evernts
8 years ago
Jeff Wang
Helin Wang
chengduoZH
Qingsheng Li
Yi Wang
Yancey
Kexin Zhao
Abhinav Arora
Yancey1989
whs
Wu Yi
Tomasz Patejko
yi.wu
Yu Yang
Qiao Longfei
Yang yaming
Thuan Nguyen
Lei Wang
dzhwinter
chengduo
qingqing01
robot
dyning
Xi Chen
baiyf
Wojciech Uss
wangyang59
Yang Yang(Tony)
Tao Luo
tangwei12
fengjiayi
weixing
Liu Yiqun
gongweibao
Yibing Liu
dzhwinter
mozga-intel
Jacek Czaja
wanglun
guosheng
Xin Pan
jiaozhenyu
Sharan Narang
jshower
weixing02
liuguoyi
xuwei06
Krzysztof Binias
Zhaolong Xing
Guo Sheng
dragonwarrior
Burness Duan
wanglun
Varun Arora
Kexin Zhao
tensor-tang