Kyubyong/transformer issues and pull requests

#50 - sinosuidal embeding vs positional what is the difference?

Issue - State: open - Opened by 4pal about 6 years ago

#50 - sinosuidal embeding vs positional what is the difference?

Issue - State: open - Opened by 4pal about 6 years ago

#49 - How does multihead_attention works？

Issue - State: open - Opened by chenwi about 6 years ago - 2 comments

#49 - How does multihead_attention works？

Issue - State: open - Opened by chenwi about 6 years ago - 2 comments

#48 - REPLACE / PREVENT UNK AT INFERENCE TIME

Issue - State: open - Opened by 4pal about 6 years ago - 1 comment

#48 - REPLACE / PREVENT UNK AT INFERENCE TIME

Issue - State: open - Opened by 4pal about 6 years ago - 1 comment

#47 - <S> is begin, </S> is end, </S> is used, but why don’t you use <S>?

Issue - State: open - Opened by azzhu about 6 years ago - 3 comments

#47 - <S> is begin, </S> is end, </S> is used, but why don’t you use <S>?

Issue - State: open - Opened by azzhu about 6 years ago - 3 comments

#46 - About Train Time

Issue - State: open - Opened by moondaiy about 6 years ago - 5 comments

#46 - About Train Time

Issue - State: open - Opened by moondaiy about 6 years ago - 5 comments

#45 - now ,my nltk，numpy ,regex and tensorflow all meet the requirements.i do not know why. I put the moduls of the "tf contrib. Linalg. LinearOperatorTriL " change to "tf.linalg.LinearOperatorLowerTriangular" ,the project can run.But I do not know if it is right or not.

Issue - State: open - Opened by jichangzhen about 6 years ago - 6 comments

#45 - now ,my nltk，numpy ,regex and tensorflow all meet the requirements.i do not know why. I put the moduls of the "tf contrib. Linalg. LinearOperatorTriL " change to "tf.linalg.LinearOperatorLowerTriangular" ,the project can run.But I do not know if it is right or not.

Issue - State: open - Opened by jichangzhen about 6 years ago - 6 comments

#44 - Why can't I run this file

Issue - State: open - Opened by jichangzhen about 6 years ago - 3 comments

#44 - Why can't I run this file

Issue - State: open - Opened by jichangzhen about 6 years ago - 3 comments

#43 - This model cannot handle extremely large dataset

Issue - State: open - Opened by RayXu14 over 6 years ago - 2 comments

#43 - This model cannot handle extremely large dataset

Issue - State: open - Opened by RayXu14 over 6 years ago - 2 comments

#42 - why normalization variables are trainable

Issue - State: closed - Opened by RayXu14 over 6 years ago - 2 comments

#42 - why normalization variables are trainable

Issue - State: closed - Opened by RayXu14 over 6 years ago - 2 comments

#41 - beam search implementation

Issue - State: open - Opened by 4pal over 6 years ago - 11 comments

#41 - beam search implementation

Issue - State: open - Opened by 4pal over 6 years ago - 11 comments

#40 - Wrong Batch Normalization

Issue - State: open - Opened by bryant03 over 6 years ago - 3 comments

#40 - Wrong Batch Normalization

Issue - State: open - Opened by bryant03 over 6 years ago - 3 comments

#39 - About the .png files in the directory "fig"

Issue - State: open - Opened by paintedveil5 over 6 years ago - 1 comment

#39 - About the .png files in the directory "fig"

Issue - State: open - Opened by paintedveil5 over 6 years ago - 1 comment

#38 - Wrong positional encoding

Issue - State: open - Opened by varshiths over 6 years ago - 2 comments

#38 - Wrong positional encoding

Issue - State: open - Opened by varshiths over 6 years ago - 2 comments

#37 - eval.py file number of item diff between test and result

Issue - State: open - Opened by behitek over 6 years ago

#37 - eval.py file number of item diff between test and result

Issue - State: open - Opened by behitek over 6 years ago

#36 - Traing Problem

Issue - State: open - Opened by guozhengpku over 6 years ago - 4 comments

#36 - Traing Problem

Issue - State: open - Opened by guozhengpku over 6 years ago - 4 comments

#35 - where are the data set from?

Issue - State: open - Opened by maozida880 over 6 years ago - 1 comment

#35 - where are the data set from?

Issue - State: open - Opened by maozida880 over 6 years ago - 1 comment

#34 - update tensorflow api to 1.8

Pull Request - State: closed - Opened by xu-song over 6 years ago - 2 comments

#34 - update tensorflow api to 1.8

Pull Request - State: closed - Opened by xu-song over 6 years ago - 2 comments

#33 - does the key masking work?

Issue - State: closed - Opened by liuwei1206 over 6 years ago - 6 comments

#33 - does the key masking work?

Issue - State: closed - Opened by liuwei1206 over 6 years ago - 6 comments

#32 - update tensorflow api

Pull Request - State: closed - Opened by xu-song over 6 years ago

#32 - update tensorflow api

Pull Request - State: closed - Opened by xu-song over 6 years ago

#31 - Understanding load_train_data.

Issue - State: open - Opened by srinivas-varadharajan over 6 years ago - 4 comments

#31 - Understanding load_train_data.

Issue - State: open - Opened by srinivas-varadharajan over 6 years ago - 4 comments

#30 - error running code

Issue - State: open - Opened by ashishkr23 over 6 years ago - 6 comments

#30 - error running code

Issue - State: open - Opened by ashishkr23 over 6 years ago - 6 comments

#29 - hello

Issue - State: open - Opened by sunnnnnnnny over 6 years ago - 1 comment

#29 - hello

Issue - State: open - Opened by sunnnnnnnny over 6 years ago - 1 comment

#28 - Add validation set

Issue - State: open - Opened by Crista23 over 6 years ago - 1 comment

#28 - Add validation set

Issue - State: open - Opened by Crista23 over 6 years ago - 1 comment

#27 - possible error in positional encoding computation

Issue - State: closed - Opened by greaber over 6 years ago - 2 comments

#27 - possible error in positional encoding computation

Issue - State: closed - Opened by greaber over 6 years ago - 2 comments

#26 - how to use the wmt dataset rather than the IWSLT 2016 de-en dataset

Issue - State: open - Opened by xumin2501 over 6 years ago - 1 comment

#26 - how to use the wmt dataset rather than the IWSLT 2016 de-en dataset

Issue - State: open - Opened by xumin2501 over 6 years ago - 1 comment

#25 - The way feed data in training

Issue - State: open - Opened by jiangbojian over 6 years ago - 2 comments

#25 - The way feed data in training

Issue - State: open - Opened by jiangbojian over 6 years ago - 2 comments

#24 - tqdm dependency

Issue - State: open - Opened by Djoels over 6 years ago - 2 comments

#24 - tqdm dependency

Issue - State: open - Opened by Djoels over 6 years ago - 2 comments

#23 - decoder_inputs

Issue - State: open - Opened by liujianliuku almost 7 years ago

#22 - Why your `feedforward` layer use `tf.layers.conv1d` instead of the two liear with relu proposed in paper?

Issue - State: open - Opened by FrankWork almost 7 years ago - 4 comments

#21 - Code of the original paper available now

Issue - State: open - Opened by TheWindRider almost 7 years ago

#20 - Fixed few bugs related to positional encodding

Pull Request - State: open - Opened by y12uc231 almost 7 years ago

#19 - Error for positional encoding

Issue - State: closed - Opened by y12uc231 almost 7 years ago - 2 comments

#18 - Embedding()

Issue - State: open - Opened by covarsha almost 7 years ago - 1 comment

#17 - multihead_attention

Pull Request - State: open - Opened by pengyuchen almost 7 years ago

#16 - Question on maskings

Issue - State: open - Opened by duyvuleo almost 7 years ago - 7 comments

#15 - scale fixed

Pull Request - State: open - Opened by tsingcoo about 7 years ago - 1 comment

#14 - Are the projection layers among multiple blocks shared?

Issue - State: closed - Opened by haoransh about 7 years ago - 3 comments

#13 - Update modules.py

Pull Request - State: closed - Opened by maximedb about 7 years ago

#12 - Evaluation

Issue - State: open - Opened by maximedb about 7 years ago - 1 comment

#11 - Dropbox link with pretrained model doesn't work

Issue - State: open - Opened by fogside about 7 years ago - 2 comments

#10 - position_encoding

Pull Request - State: closed - Opened by EternalFeather about 7 years ago

#9 - Shape mismatch error in eval

Issue - State: open - Opened by erickrf over 7 years ago - 1 comment

#8 - A problem about decoder input

Issue - State: closed - Opened by zz-jacob over 7 years ago

#7 - There is a input sequence, how to calculate probability for the special output sequence？

Issue - State: open - Opened by yick2232 over 7 years ago - 1 comment

#6 - Read the train sequences

Issue - State: closed - Opened by HuachunWang over 7 years ago - 1 comment

#5 - train.py error

Issue - State: open - Opened by EasonZhaoZ over 7 years ago - 5 comments

#4 - An easy question about linear

Issue - State: closed - Opened by guotong1988 over 7 years ago - 1 comment

#3 - Why doing Key and Query Masking ?

Issue - State: open - Opened by odie2630463 over 7 years ago - 9 comments

#2 - Add basic helpful steps.

Pull Request - State: closed - Opened by kimdwkimdw over 7 years ago - 1 comment

#1 - saver error occors

Issue - State: open - Opened by yxyuni over 7 years ago

GitHub / Kyubyong/transformer issues and pull requests