Commit Graph

217 Commits (8002b2beb4a787d7f21a272289f4b1a36953c371)

Author SHA1 Message Date
Zhou Wei 8002b2beb4
Avoid logging.info be printed many times in dygraph_mode,test=develop (#23932)
5 years ago
Zhou Wei 66dc8e30f0
move the initialize position of grad_clip to optimizer(__init__),and speed up clip (#23782)
5 years ago
mapingshuo f0e743f136
fix AMP and recompute (#23551)
5 years ago
Zhou Wei 629b6c7896
add the prompt message of repeated settings of regularization,test=develop (#23355)
5 years ago
qingqing01 6162cf2f2e
Make optimizer consistent in dygraph and static-graph and remove some LOG-INFO. (#23426)
5 years ago
Zhou Wei e8efaee92d
update gradient clip english doc for new gradient clipping strategy
5 years ago
Leo Chen a62599a888
[feature] prune program by feed and fetch_list automatically (#22474)
5 years ago
Zhou Wei 7fda333ac1
add new method of gradient_clip, better to use,test=develop (#23224)
5 years ago
Leo Chen 488b2387e2
Feature/expand params in auto-generated pybind functions for dygraph operators (#23181)
5 years ago
Zhang Ting eec10aaba2
set op_device for loss_op_desc (#23027)
5 years ago
WangXi f2265d9ffd
Fix problem use recompute and dgc same time (#23010)
5 years ago
mapingshuo 08a772cb46
fix API param bug of recompute.backward() (#22582)
5 years ago
WangXi 62fd3209e1
Fix dgc param regularizer, test=develop (#22888)
5 years ago
Zhang Ting 4e8bc02461
add fluid.device_guard to specify the device type for Op (#22254)
5 years ago
tianshuo78520a 433cef03e5
fix typo word (#22784)
5 years ago
zhaoyuchen2018 72dde4abde
Refine adam op to improve performance, test=develop (#22346)
5 years ago
tianshuo78520a d2ba91aad1
fix typo words (#22653)
5 years ago
WangXi d69df9bf26
Add wrong info when use DGC in cpu (#22515)
5 years ago
Aurelius84 50af6b5d79
polish no_grad_set of gradient and append_backward (#22440)
5 years ago
hong 00c0139e6e add learning rate api of optimizer (#22080)
5 years ago
zhongpu b1c081f4c7 polish Optimizer's API description, test=develop (#22314)
5 years ago
Aurelius84 60a6d68fb9
remove _optimized_guard in dygrahpe_mode (#22143)
5 years ago
Leo Chen d4bdbf8cf0
Polish nn code, test=develop (#22237)
5 years ago
zhongpu d0f0a2520c test Optimizer in dygraph (#21949)
6 years ago
zhongpu 7d10edc5ee add clear_gradients for Optimizer and add clear_gradients api description (#21948)
6 years ago
Leo Chen 51cb918a05 update layers used in transformer dygraph model, test=develop (#22051)
6 years ago
Yibing Liu 434f7b9c0f
Fix the global_step & continuous applying error in EMA (#22090)
6 years ago
liym27 7d8d45991a
control flow: support optimizer called (#21851)
6 years ago
Chen Weihang 9a2204ee1f
Uniform append_backward & gradients parameter_list type to Variable (#21938)
6 years ago
Aurelius84 6b4c33ee30
fix decay param in DecayAdagrad test=develop (#22026)
6 years ago
Leo Chen 33f1306753
update layers used in mnist dygraph model, test=develop (#21947)
6 years ago
zhongpu dca075839b remove params in Tracer object (in dygraph) (#20815)
6 years ago
Aurelius84 51a86d2b6b Optimize adam speed (#21777)
6 years ago
Leo Chen 310edc0d0c
Update layers used in ptb model to use auto-generated op functions in dygraph mode (#21724)
6 years ago
tangwei12 934d99867a
add selected rows supported in framework (#21808)
6 years ago
hutuxian c5aec2fe68
Paddlebox Related to Framework (#21586)
6 years ago
Leo Chen cdd46d7e02
Split VarBase from Python Variable for Dygraph (#21359)
6 years ago
WangXi 768f9242e9 Fix dgc clip & rampup step, test=develop (#21491)
6 years ago
zhongpu 6ebf0f47b8 support SelectedRows in dygraph, test=develop (#21078)
6 years ago
Aurelius84 4bf115b42d
Fix AdamOptimizer and Scale sample code Bug (#21478)
6 years ago
Zeng Jinle 87ab93af01
fix adam fp64, test=develop (#21423)
6 years ago
Kaipeng Deng ebfb720a63
add Adam beta1/beta2 support Variable (#21234)
6 years ago
WangXi 8ac7687e36 Fix dgc accuracy by mv regularization to local (#21278)
6 years ago
WangXi de5d3ff688 Fix dgc buffer illegal & reuse velocity (#21012)
6 years ago
WangXi 250e72d254 Fix DGC algorithm flow to make it the same as paper (#20758)
6 years ago
WangXi 507afa8a8a Fix dgc nan by stripping nccl from sparseReduce. (#20630)
6 years ago
wopeizl 6b612a2807
optimize the english description for APIs test=develop test=document_fix (#20286)
6 years ago
Yibing Liu 899ab30df0
Fix several api docs (#20282)
6 years ago
Zeng Jinle 7fb958b2e1
replace data layers, test=develop, test=document_fix (#20368)
6 years ago
hong fa43e80e19 New save load interface (#20148)
6 years ago