2f8010f805
!10975 Increase performance of fasttext
...
From: @zhouyaqiang0
Reviewed-by:
Signed-off-by:
4 years ago
2c4f7d92b7
update model_zoo/official/nlp/gnmt_v2/src/gnmt_model/beam_search.py.
4 years ago
84303c539f
!11003 upload chinese readme
...
From: @yoonlee666
Reviewed-by:
Signed-off-by:
4 years ago
2d1e8beec7
add chinese readme
4 years ago
8a0bca7d6b
Increase performance of fasttext
4 years ago
33277fe64b
fix bert init
4 years ago
f0b08e8bff
all reduce after each step in gradients accumulation mode for bert
4 years ago
811a785555
!10886 GPU update bert scripts
...
From: @VectorSL
Reviewed-by: @gaoxiong1,@c_34,@gaoxiong1
Signed-off-by: @c_34
4 years ago
a531628dd5
!10325 modify MetaTensor and Tensor
...
From: @Somnus2020
Reviewed-by: @kingxian
Signed-off-by: @kingxian
4 years ago
0c97835662
update control flow int adamweightdecay for bert
4 years ago
b858097ae4
modify MetaTensor and Tensor
4 years ago
048e3f3779
LSTM README
4 years ago
ad6db39c2d
fix LSTM D MS version in README
4 years ago
02c107b0e9
!10874 修改GPU设置device_id的问题
...
From: @yuzhenhua666
Reviewed-by: @c_34,@linqingke
Signed-off-by: @c_34
4 years ago
6b858480c8
fix GPU device_id bug
4 years ago
812b4b0eab
extract embedding table from unified interface
4 years ago
a152c510c5
fix textcnn export on master branch
4 years ago
92049c4012
!10684 add README_CN.md for bert_thor
...
From: @sl_wang
Reviewed-by: @wang_zi_dong
Signed-off-by:
4 years ago
1566022045
add READEME_cn.md for BERT_thor
4 years ago
a910086b50
!10653 bert on gpu for pre training script supports loss scale
...
From: @hanhuifeng2020
Reviewed-by: @c_34,@gaoxiong1
Signed-off-by: @c_34
4 years ago
2f2a624281
gnmt add hub_config
4 years ago
65bfefd92b
bert on gpu for pre training script supports loss scale
4 years ago
528f4ddd33
!10596 Add export.py to textcnn (PR to master)
...
From: @penny369
Reviewed-by: @guoqi1024,@pandoublefeng
Signed-off-by: @guoqi1024
4 years ago
0903777d3f
!10507 Modify dataset preprocess of textcnn
...
From: @zhouyaqiang0
Reviewed-by: @linqingke,@oacjiewen
Signed-off-by: @linqingke
4 years ago
a17c4352ec
!10586 add data sink note in transformer readme
...
From: @yuchaojie
Reviewed-by: @c_34,@oacjiewen,@linqingke
Signed-off-by: @c_34,@linqingke
4 years ago
1148c942e3
add export fun to textcnn
4 years ago
e3cdd89a2e
add data sink note in transformer readme
4 years ago
f79e3bdb39
!10533 lstm script fix
...
From: @ttudu
Reviewed-by: @liangchenghui,@c_34
Signed-off-by: @c_34
4 years ago
2cbd6d3c97
Modify data preprocess of textcnn
4 years ago
b3577286a2
lstm script fix
4 years ago
d3e561c5a7
add export file
4 years ago
216aa8535b
!10412 Delete the redundant inputs of GNMTv2 model
...
From: @gaojing22
Reviewed-by: @c_34,@oacjiewen
Signed-off-by: @c_34
4 years ago
1f3d9ddff8
delete schema_json
4 years ago
31fed1a2f6
change code to import APIs from mindspore.dataset rather than mindspore.dataset.engine
4 years ago
18ca7eaeb0
!10222 lstm D
...
From: @ttudu
Reviewed-by:
Signed-off-by:
4 years ago
c4eb32b43a
move textrcnn from official to research, and raise acc when using lstm as RNN arch
4 years ago
9103ff5d75
lstm D network
4 years ago
ca182f0131
!10341 new add mass and unet chinese readme.
...
From: @linqingke
Reviewed-by: @oacjiewen,@c_34
Signed-off-by: @c_34
4 years ago
1150ae3376
!10290 add fasttext model to model zoo
...
From: @zhaojichen
Reviewed-by:
Signed-off-by:
4 years ago
edc48b48b8
add fasttext to model_zoo
4 years ago
338c82d57e
new add mass and unet chinese readme.
4 years ago
00611b1c41
!10175 Add tensor.ndim and rename tensor.size() to tensor.size
...
From: @yanglf1121
Reviewed-by: @kingxian,@zhunaipan
Signed-off-by: @kingxian
4 years ago
d5d3c7b349
modify export script file to support mindr and GPU
4 years ago
66741fb545
!10225 fix ncf export bug, modify export script for bert, mass, transformer to support mindir and GPU
...
From: @yuzhenhua666
Reviewed-by: @liangchenghui,@c_34
Signed-off-by: @c_34
4 years ago
4128fb1155
modify export to support mindir
4 years ago
95ac8549db
fix textcnn accuracy problem
4 years ago
918679daa3
add tensor.ndim and rename tensor.size() to tensor.size
4 years ago
1631765922
!10165 Performance optimization of Bert on GPU by the graph_kernel
...
From: @hanhuifeng2020
Reviewed-by: @gaoxiong1,@ryanww
Signed-off-by: @ryanww
4 years ago
3988376b67
Performance optimization of Bert on GPU by the graph_kernel
4 years ago
08e2424fb8
textrcnn performance standard modified
4 years ago
4ce11a930b
!10118 raise textrcnn precision when using DynamicRNN
...
From: @chenmai1102
Reviewed-by: @guoqi1024,@oacjiewen
Signed-off-by: @guoqi1024
4 years ago
1df083b8d7
fixed the bad links
4 years ago
d572b47821
!10104 Improve accuracy of TextCNN a litte bit by adjusting initializer approaches
...
From: @penny369
Reviewed-by: @guoqi1024,@oacjiewen
Signed-off-by: @guoqi1024
4 years ago
d36abebe54
raise textrcnn precision on DynamicRNN
4 years ago
a15152141d
add transformer chinese readme
4 years ago
f76758cbe8
fix textcnn problem
4 years ago
0f610e11de
set HCCL_CONNECT_TIMEOUT=600 for transformer distribute training
4 years ago
f26003e723
fix mass lr import.
4 years ago
71f4bcc7b6
!9820 bert readme update
...
From: @wilfchen
Reviewed-by: @limingqi107,@cristoval
Signed-off-by: @cristoval
4 years ago
f708e4b108
fix xception and textrcnn
4 years ago
62652cc29d
bert readme update
4 years ago
84acebf19d
lstm add export.py
4 years ago
4b436d61a8
!9377 add model textrcnninto modelzoo
...
From: @zymaa
Reviewed-by:
Signed-off-by:
4 years ago
16c5013f6e
add textrcnn into model zoo
...
Signed-off-by: zymaa <317958662@qq.com>
TextRCNN, a model for text classification, which is proposed by the Chinese Academy of Sciences in 2015.
TextRCNN actually combines RNN and CNN, first uses bidirectional RNN to obtain upper semantic and grammatical information of the input text,
and then uses maximum pooling to automatically filter out the most important feature.
Signed-off-by: zymaa <317958662@qq.com>
4 years ago
d0e7ae66a4
!9689 fix model zoo typos
...
From: @alex-yuyue
Reviewed-by: @nsyca,@mikef
Signed-off-by: @mikef
4 years ago
59ca2ac708
!9693 add multi machine instruction for bert
...
From: @yoonlee666
Reviewed-by:
Signed-off-by:
4 years ago
91fb5867d3
!9235 textcnn
...
From: @xinyunfan
Reviewed-by:
Signed-off-by:
4 years ago
fe9443bfc6
add multi-machine
4 years ago
17d6f1c2f9
add option for graph kernel and mixed precision
4 years ago
5250b327ae
Fix some non-minddata typos
...
Signed-off-by: alex-yuyue <yue.yu1@huawei.com>
4 years ago
7bfe1a5d34
!9663 remove unused parameters in config
...
From: @sl_wang
Reviewed-by: @wang_zi_dong,@kisnwang
Signed-off-by: @wang_zi_dong
4 years ago
56c8c346dd
remove unused optimizer inf config file
4 years ago
213185569e
fix doc bug
4 years ago
de7e8850da
fix doc & code bugs for gnmt
4 years ago
7131707359
!9366 fix bert export bug
...
From: @yuzhenhua666
Reviewed-by: @c_34
Signed-off-by: @c_34
4 years ago
c3f3f6e67a
upload textcnn
4 years ago
0d1f209d69
fix bert export bug
4 years ago
0c7ba7a7fa
!9354 mode_mindir_suffix
...
From: @bai-yangfan
Reviewed-by: @c_34,@guoqi1024
Signed-off-by: @c_34
4 years ago
c46c4dffe4
mindir_suffix
4 years ago
b1189cd118
compatible name
4 years ago
510ed65300
!9264 export script for resnet50_quant and densenet121，fix centerface and transformaer bug
...
From: @yuzhenhua666
Reviewed-by: @linqingke,@yingjy
Signed-off-by: @linqingke
4 years ago
989744c61a
!9258 add squad for bert
...
From: @yoonlee666
Reviewed-by: @c_34,@guoqi1024
Signed-off-by: @c_34
4 years ago
ca66aef549
!8775 remove "is_grad" in SoftmaxCrossEntropyWithLogits and correct its comments
...
From: @wanyiming
Reviewed-by: @kingxian,@zh_qh
Signed-off-by: @kingxian
4 years ago
ceaca1708c
add export file for resnet50_quant and densenet121, fix centerface and transformer bug
4 years ago
a744ef9113
add squad
4 years ago
bebc37b389
!9192 reset the default activation function for bert
...
From: @shibeiji
Reviewed-by: @c_34,@liangchenghui
Signed-off-by: @c_34
4 years ago
bde36af04c
Modify the default activation function for bert
4 years ago
8bbd84eb62
remove network useless parameters
4 years ago
0aa63f21c5
!9039 register FastGelu for activation
...
From: @shibeiji
Reviewed-by:
Signed-off-by:
4 years ago
9130660a2f
!9055 add yolov3 export file and fix bert exprt bug
...
From: @yuzhenhua666
Reviewed-by: @liangchenghui
Signed-off-by:
4 years ago
a1db640211
fix bert bug and add export file for yolov3 and add test file for maskrcnn
4 years ago
cd850bd210
register activation operator fast_gelu
4 years ago
a0e4db6aae
!9016 delete input_mask_from_dataset option in transformer's config
...
From: @yuchaojie
Reviewed-by: @liangchenghui,@linqingke
Signed-off-by: @linqingke
4 years ago
f13388ee6b
delete input_mask_from_dataset option in transformer's config
4 years ago
48e688c166
add dynamic shape support
4 years ago
7689062c7d
!8727 Clean code for gnmt_v2 network
...
From: @gaojing22
Reviewed-by: @yingjy,@guoqi1024
Signed-off-by: @yingjy
4 years ago
a23ddaf409
clean_code
4 years ago
dad47cf2dd
mod_comments
4 years ago
641888f603
add definition of scale for GPT
4 years ago
2353756515
BugFix for GPT
4 years ago
mindspore-ci-bot
casgj
yoonlee666
zhouyaqiang
VectorSL
shibeiji
lilei
root
yuzhenhua
Yanjun Peng
mwang
hanhuifeng2020
yuchaojie
zhaojichen
Xiao Tianci
陈劢
linqingke
yanglf1121
caojiewen
wilfChen
CaoJian
zymaa
tronzhang
alex-yuyue
xinyun Fan
bai-yangfan
caozhou
liyong
wanyiming
alouhahaha