bentrevett/pytorch-seq2seq issues and pull requests

#92 - Modifying training for captioning

Issue - State: closed - Opened by kushalj001 over 4 years ago - 1 comment

#91 - In the encoder. Why not use pack_padded_sequence (embedded, input_lengths)

Issue - State: closed - Opened by lightcome over 4 years ago - 3 comments
Labels: question

#90 - A few questions about Tutorial 4.

Issue - State: closed - Opened by labusi over 4 years ago - 3 comments

#89 - 6 - Attention is All You Need : nn.Embedding related Runtime error with positional encoding layer during training

Issue - State: closed - Opened by Ninja16180 over 4 years ago - 6 comments

#88 - A question about tutorial1.

Issue - State: closed - Opened by TomasAndersonFang over 4 years ago - 1 comment
Labels: question

#87 - Typo in notebook 5 ?

Issue - State: closed - Opened by garychl over 4 years ago - 1 comment

#86 - a mini spelling error

Issue - State: closed - Opened by zhiyiZeng over 4 years ago - 2 comments

#85 - Using 6th notebook for Video captioning

Issue - State: closed - Opened by kushalj001 over 4 years ago - 5 comments
Labels: waiting-reply

#84 - Why the prediction of evaluation function use gold target sequence in CNN seq2seq?

Issue - State: closed - Opened by fake-warrior8 over 4 years ago - 2 comments

#83 - What is `seq_len` in Transformers notebook?

Issue - State: closed - Opened by kushalj001 over 4 years ago - 1 comment

#82 - Tutorial - 4 How to apply pretrained embedding and initial hidden state

Issue - State: closed - Opened by rishab-32 over 4 years ago - 7 comments

#81 - Applying BERT pre-training

Issue - State: closed - Opened by aimanmutasem over 4 years ago - 2 comments

#80 - for convolutional seq 2 seq

Issue - State: closed - Opened by menglin0320 over 4 years ago - 7 comments

#79 - Packed Padded Sequences

Issue - State: closed - Opened by dariodematties over 4 years ago - 1 comment

#78 - Dockerize

Pull Request - State: closed - Opened by quinn-dougherty over 4 years ago

#77 - What is the best way to training convolutional seq2seq model with two datasets?

Issue - State: closed - Opened by aimanmutasem almost 5 years ago

#76 - Doubt regarding src_key_padding_mask

Issue - State: closed - Opened by thak123 almost 5 years ago - 2 comments

#75 - Simplified the attention class by refactoring v to a Linear layer

Pull Request - State: closed - Opened by M-e-r-c-u-r-y almost 5 years ago - 1 comment

#74 - Need of parameter v in Attention(Bahdanau) model?

Issue - State: closed - Opened by kushalj001 almost 5 years ago - 1 comment

#73 - A question about mask in 6th tutorial

Issue - State: closed - Opened by hmtbgc almost 5 years ago - 4 comments

#72 - Help to apply fasttext pre-trining embedding vector.

Issue - State: closed - Opened by aimanmutasem almost 5 years ago - 10 comments

#71 - Gramatical error in ConvS2S notebook

Issue - State: closed - Opened by M-e-r-c-u-r-y almost 5 years ago - 3 comments

#70 - Help needed for transformer encoder only function

Issue - State: closed - Opened by thak123 almost 5 years ago - 3 comments

#69 - Understanding Pos_embedding in Attention is all you need

Issue - State: closed - Opened by arushi-08 almost 5 years ago - 4 comments

#68 - BucketIterator splits function takes batch values seperately for train valid and test split

Issue - State: closed - Opened by thak123 almost 5 years ago - 2 comments

#67 - [Errno 2] No such file or directory: '.data\\multi30k\\train.de'

Issue - State: closed - Opened by s1530129650 almost 5 years ago - 12 comments

#67 - [Errno 2] No such file or directory: '.data\\multi30k\\train.de'

Issue - State: closed - Opened by s1530129650 almost 5 years ago - 12 comments

#66 - A problem with loss computation.

Issue - State: closed - Opened by yxdr almost 5 years ago - 9 comments

#65 - unk issue

Issue - State: closed - Opened by FrederickXZhang almost 5 years ago - 2 comments

#64 - packed padded sequences

Issue - State: closed - Opened by possible1402 almost 5 years ago - 2 comments

#63 - Not sure about final embeddings vectors

Issue - State: closed - Opened by MaxenceQueyrel almost 5 years ago - 4 comments

#62 - bleu score computation

Issue - State: closed - Opened by subhadarship almost 5 years ago - 1 comment

#61 - Questions about optimizer in 6th tutorial (Transformer)

Issue - State: closed - Opened by GorkaUrbizu almost 5 years ago - 2 comments

#60 - question about tutorial1 Sequence to Sequence Learning with Neural Networks

Issue - State: closed - Opened by possible1402 almost 5 years ago - 4 comments

#59 - Just want to say a big thank you!

Issue - State: closed - Opened by Hans0124SG almost 5 years ago - 2 comments

#58 - AttributeError: 'Tensor' object has no attribute 'bool'. 6th tuto (Transformer) with Pytorch < 1.2 and how to solve it.

Issue - State: closed - Opened by GorkaUrbizu almost 5 years ago - 2 comments

#57 - need help with 6th tutorial (transformer). Assert error at train when changing dataset.

Issue - State: closed - Opened by GorkaUrbizu almost 5 years ago - 6 comments

#56 - Regarding Teacher Forcing Ratio

Issue - State: closed - Opened by prakhar21 almost 5 years ago

#55 - A little question about pytorch-seq2seq/3 - Neural Machine Translation by Jointly Learning to Align and Translate.ipynb

Issue - State: closed - Opened by guoxiaojun1999 almost 5 years ago - 2 comments

#54 - RuntimeError: Number of dimensions of repeat dims can not be smaller than number of dimensions of tensor

Issue - State: closed - Opened by banunitte about 5 years ago - 3 comments

#53 - Why the attention is a function of encoder's output?

Issue - State: closed - Opened by ziadloo about 5 years ago - 2 comments

#52 - why training on the WMT'14 dataset is so slow?

Issue - State: closed - Opened by zhao1402072392 about 5 years ago - 11 comments

#51 - Question on Tutorial 3 (rnn output explanation)

Issue - State: closed - Opened by antgr about 5 years ago - 3 comments

#50 - tutorial 4 [typo]

Issue - State: closed - Opened by antgr about 5 years ago - 1 comment

#49 - Improved transformer model

Pull Request - State: closed - Opened by hawkeoni about 5 years ago - 2 comments

#48 - Inference in transformer seq2seq (ex. 6)

Issue - State: closed - Opened by GorkaUrbizu about 5 years ago - 2 comments

#47 - Why GRU instead of LSTM when adding attention?

Issue - State: closed - Opened by GorkaUrbizu about 5 years ago - 4 comments

#46 - lesson1，calculate the loss between trg and ouput when traning

Issue - State: closed - Opened by zhao1402072392 about 5 years ago - 6 comments

#45 - outputs[0] is the zeroes tensor initialized by torch.zeroes()

Issue - State: closed - Opened by wz337 about 5 years ago - 3 comments

#44 - About seq2seq tutorial 2

Issue - State: closed - Opened by deartonym about 5 years ago - 2 comments

#43 - Do we have to append <sos> and <eos> token to source sentence?

Issue - State: closed - Opened by Huffon about 5 years ago - 2 comments

#42 - found a small typo

Issue - State: closed - Opened by Huffon about 5 years ago - 1 comment

#41 - Convolutional Seq2Seq

Issue - State: closed - Opened by kuldeep7688 about 5 years ago - 2 comments

#40 - Why don't we use `RNNCell` instead of `RNN`?

Issue - State: closed - Opened by standbyme about 5 years ago - 5 comments

#39 - Predicting on One Sentence whose Target is Not Known using Convolutional Seq2Seq

Issue - State: closed - Opened by kuldeep7688 about 5 years ago - 4 comments

#38 - Convolutional Sequence to Sequence

Issue - State: closed - Opened by kuldeep7688 about 5 years ago - 2 comments

#37 - Multi Layer Decoder

Issue - State: closed - Opened by sheoranhimansh about 5 years ago - 2 comments

#36 - Small Typo in the title

Pull Request - State: closed - Opened by mahaloz about 5 years ago

#35 - notebook 6, type error

Issue - State: closed - Opened by sonalgarg2806 about 5 years ago - 1 comment

#34 - Lesson 3 (Attention Implementation)

Issue - State: closed - Opened by abhinavshaw1993 over 5 years ago - 3 comments

#33 - Update 4 - Packed Padded Sequences, Masking and Inference.ipynb

Pull Request - State: closed - Opened by yissachar over 5 years ago

#32 - notebook 5, pos_embeddings, magic number 100

Issue - State: closed - Opened by antgr over 5 years ago - 2 comments

#31 - Attention mechanism

Issue - State: closed - Opened by antgr over 5 years ago - 2 comments

#30 - notebook 5, padding in conv

Issue - State: closed - Opened by antgr over 5 years ago - 3 comments

#29 - multilayer decoder

Issue - State: closed - Opened by he727425209 over 5 years ago - 1 comment

#28 - All Hidden Layer of Encoder

Issue - State: closed - Opened by divyeshrajpura4114 over 5 years ago - 1 comment

#27 - tutorial 3, picture of bidirectional rnn

Issue - State: closed - Opened by antgr over 5 years ago - 1 comment

#26 - tutorial 1, equations for decoder

Issue - State: closed - Opened by antgr over 5 years ago - 5 comments

#25 - lesson 3 attention

Issue - State: closed - Opened by he727425209 over 5 years ago - 1 comment

#24 - On the output layer and use of LSTM in the decoder (Lesson 1)

Issue - State: closed - Opened by boyu-zhang-25 over 5 years ago - 1 comment

#23 - loss normalization

Issue - State: closed - Opened by jiangxiluning over 5 years ago - 1 comment

#22 - Used seq2seq as autoencoder, first token not predicted right

Issue - State: closed - Opened by exnx over 5 years ago - 5 comments

#21 - No tracking of other metrics than loss/perplexity

Issue - State: closed - Opened by serhiy-yevtushenko over 5 years ago - 7 comments

#20 - No way to generate output from models

Issue - State: closed - Opened by serhiy-yevtushenko over 5 years ago - 1 comment

#19 - lesson 2 decoder impl output to linear

Issue - State: closed - Opened by iyersathya over 5 years ago - 2 comments

#18 - How to do a single prediction?

Issue - State: closed - Opened by Jeblii over 5 years ago - 2 comments

#17 - Could you provide functions to convert the network outputs into sentences?

Issue - State: closed - Opened by wfchen-umich over 5 years ago - 2 comments

#16 - CUDA out of memory

Issue - State: closed - Opened by EthanChen911238 over 5 years ago - 3 comments

#15 - how to download de package

Issue - State: closed - Opened by jwc19890114 over 5 years ago - 4 comments

#14 - Decoding in conv seq2seq

Issue - State: closed - Opened by thinkwee over 5 years ago - 2 comments

#13 - Program running on cpu

Issue - State: closed - Opened by deekshaVarshney over 5 years ago

#12 - How to add some functionalities?

Issue - State: closed - Opened by liperrino over 5 years ago - 1 comment

#11 - RuntimeError: Invalid device, must be cuda device

Issue - State: closed - Opened by liperrino over 5 years ago - 10 comments

#10 - Notation in creating the context vector

Issue - State: closed - Opened by graviraja almost 6 years ago - 1 comment

#9 - Discussion about reversing the order of the source sentence

Issue - State: closed - Opened by littleflow3r almost 6 years ago - 2 comments

#8 - AttributeError: 'Field' object has no attribute 'vocab'

Issue - State: closed - Opened by Blaze404 almost 6 years ago - 1 comment

#7 - length of target in inference

Issue - State: closed - Opened by CuriousDeepLearner almost 6 years ago - 5 comments

#6 - Question about the attention

Issue - State: closed - Opened by Han8931 almost 6 years ago - 2 comments

#5 - How to generate new unseen translations

Issue - State: closed - Opened by fcarichon almost 6 years ago - 2 comments

#4 - arguments are located on different GPUs

Issue - State: closed - Opened by SoheilZabihi almost 6 years ago - 4 comments
Labels: bug

#3 - forward function in encoder

Issue - State: closed - Opened by Han8931 about 6 years ago - 11 comments

#2 - forward function in encoder

Issue - State: closed - Opened by Han8931 about 6 years ago

#1 - Something wrong with jupyter notebook to open?

Issue - State: closed - Opened by fabrahman about 6 years ago - 1 comment

GitHub / bentrevett/pytorch-seq2seq issues and pull requests