Ecosyste.ms: Issues

An open API service for providing issue and pull request metadata for open source projects.

GitHub / codertimo/BERT-pytorch issues and pull requests

#109 - Docs: Typo Fix

Pull Request - State: open - Opened by PalashDhabale 4 months ago

#106 - Why Segment Embedding number only 3?

Issue - State: open - Opened by UTimeStrange about 1 year ago - 1 comment

#105 - fix(sec): upgrade torch to 1.13.1

Pull Request - State: open - Opened by realize096 over 1 year ago

#104 - Fix require_grad typo

Pull Request - State: open - Opened by kit1980 over 1 year ago

#102 - GELU is available in PyTorch

Issue - State: open - Opened by hubin-keio almost 2 years ago - 1 comment

#101 - debug BERT-pytorch\bert_pytorch\model\embedding\position.py

Pull Request - State: open - Opened by wanghesong2019 about 2 years ago

#100 - how to fine tune model with trained weight

Issue - State: open - Opened by lishuaiCN about 2 years ago

#99 - What dataset did you use to train model?

Issue - State: open - Opened by Coding-Child over 2 years ago - 2 comments

#98 - why specify `ignore_index=0` in the NLLLoss function in BERTTrainer?

Issue - State: open - Opened by Jasmine969 over 2 years ago - 1 comment

#97 - Added a Google Colab Notebook that contains all the code in this project.

Pull Request - State: open - Opened by ginward over 2 years ago - 1 comment

#96 - bert-vocab?

Issue - State: open - Opened by nlpCSir almost 3 years ago - 1 comment

#95 - Pooler layer?

Issue - State: open - Opened by lihaoxin2020 about 3 years ago

#93 - dataset / dataset.py have one erro?

Issue - State: open - Opened by ndn-love over 3 years ago - 1 comment

#92 - Why not use torch.no_grad when evaluating test data?

Issue - State: open - Opened by EvanZ over 3 years ago - 1 comment

#91 - why language_model.py has different vectors

Issue - State: open - Opened by zysNLP over 3 years ago - 1 comment

#90 - Minor comment fix

Pull Request - State: closed - Opened by maengsanha almost 4 years ago

#88 - IndexError

Issue - State: open - Opened by LemonQC almost 4 years ago - 6 comments

#85 - .

Issue - State: closed - Opened by ghost over 4 years ago

#84 - residual connection and layernorm according to paper

Pull Request - State: open - Opened by dshwei over 4 years ago

#82 - how to do Ner

Issue - State: open - Opened by DeShuiYu over 4 years ago

#81 - Default model sizes are much smaller than BERT base

Issue - State: open - Opened by bertmaher over 4 years ago

#79 - How to use this bert model to load and finetune ?

Issue - State: open - Opened by dongrixinyu over 4 years ago

#78 - embedding/position.py with RuntimeError if d_model is odd

Issue - State: open - Opened by bookong22 over 4 years ago - 2 comments

#77 - Update retrain.py

Pull Request - State: open - Opened by watseob over 4 years ago

#76 - How to Output Embedded Sentence Vector?

Issue - State: open - Opened by rhypowang almost 5 years ago - 2 comments

#74 - How to use BERT model to fine-tune a cloze-style task?

Issue - State: open - Opened by OrchidXu almost 5 years ago

#73 - Can you share me a trained data?

Issue - State: open - Opened by vcbeaut about 5 years ago

#72 - predict next sentence ony? it also can do SQA assignments?

Issue - State: open - Opened by cqray1990 over 5 years ago

#69 - Vocab's load_vectors seems to be an old method?

Issue - State: open - Opened by JamesOwers over 5 years ago - 1 comment

#68 - How to Output Embedded Word Vector

Issue - State: open - Opened by enze5088 over 5 years ago - 6 comments

#67 - ONNX conversion: TransformerBlock problem

Issue - State: open - Opened by jbmaxwell over 5 years ago

#66 - Fix Bug in getting random line

Pull Request - State: open - Opened by Zenglinxiao over 5 years ago

#65 - Erroneous Code

Issue - State: open - Opened by LeoLai930603 over 5 years ago

#63 - Add: tqdm total

Pull Request - State: open - Opened by YongWookHa over 5 years ago

#62 - Minor comment typo

Pull Request - State: open - Opened by BoPengGit almost 6 years ago

#61 - How can I finetune with this project

Issue - State: open - Opened by Gpwner almost 6 years ago - 5 comments

#60 - self.d_k = d_model // h gives 64 dimension ?

Issue - State: open - Opened by BerenLuthien almost 6 years ago - 1 comment

#59 - would you provide some sample datasets for demo the pre-training

Issue - State: closed - Opened by SeekPoint almost 6 years ago - 9 comments

#58 - what does \t be used for here

Issue - State: closed - Opened by junchen14 almost 6 years ago - 1 comment

#57 - How to test your model after training on my own dataset?

Issue - State: open - Opened by riktimmondal almost 6 years ago - 1 comment

#56 - chooses 15% of token

Issue - State: open - Opened by makcedward almost 6 years ago - 1 comment
Labels: good first issue

#54 - Fix non-matching tensor with odd embedding size.

Pull Request - State: open - Opened by zmy about 6 years ago

#53 - PositionalEmbedding

Issue - State: open - Opened by fgaoyang about 6 years ago - 3 comments
Labels: good first issue

#52 - IndexError: list index out of range

Issue - State: open - Opened by ghost about 6 years ago - 13 comments

#51 - [BERT] Cannot import bert

Issue - State: closed - Opened by logicmd about 6 years ago - 2 comments

#50 - made a script to generate bert pre-train data

Issue - State: open - Opened by fucusy about 6 years ago

#49 - Question about the loss of Masked LM

Issue - State: open - Opened by zhezhaoa about 6 years ago - 5 comments
Labels: good first issue

#47 - ner

Issue - State: closed - Opened by sz128 about 6 years ago

#46 - Attention maybe changed

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

#45 - Maybe Bugs?

Issue - State: closed - Opened by sq6ra over 6 years ago - 3 comments

#44 - Very low GPU usage when training on 8 GPU in a single machine

Issue - State: closed - Opened by mapingshuo over 6 years ago - 13 comments
Labels: help wanted

#43 - Making Book Corpus

Issue - State: open - Opened by codertimo over 6 years ago - 5 comments
Labels: help wanted

#42 - Making Wikipedia Corpus

Issue - State: open - Opened by codertimo over 6 years ago - 1 comment
Labels: help wanted

#41 - the format of input

Issue - State: open - Opened by eveliao over 6 years ago - 4 comments
Labels: question

#40 - Tie the input and output embedding?

Issue - State: open - Opened by jiqiujia over 6 years ago - 5 comments
Labels: enhancement

#39 - Why doesn't the counter in data_iter increase?

Issue - State: closed - Opened by nateraw over 6 years ago - 6 comments
Labels: invalid, question

#38 - how to test the model?

Issue - State: open - Opened by eveliao over 6 years ago - 2 comments
Labels: enhancement, question

#37 - shape match for mul？

Issue - State: closed - Opened by guotong1988 over 6 years ago - 1 comment

#36 - Why output_label=0 in datasets generation for Masked LM

Issue - State: closed - Opened by leon-cas over 6 years ago - 5 comments
Labels: question

#35 - Tensor transform question in pretrain.py

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

#34 - Pretrained model transfer to pytorch

Issue - State: open - Opened by codertimo over 6 years ago - 8 comments
Labels: enhancement, help wanted

#33 - Vocab Replace \t to blank issue

Issue - State: open - Opened by NiHaoUCAS over 6 years ago - 2 comments
Labels: bug

#32 - pred_loss decrease fast while avg_acc stay at 50%

Issue - State: open - Opened by jiqiujia over 6 years ago - 53 comments
Labels: help wanted, invalid, question

#31 - Single Sentence Input support

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

#30 - The LayerNorm implementation

Issue - State: closed - Opened by egg-west over 6 years ago - 4 comments
Labels: invalid, question

#29 - when training the masked LM, the unmasked words (have label 0) were trained together with masked words?

Issue - State: open - Opened by coddinglxf over 6 years ago - 6 comments
Labels: enhancement, question

#28 - Example of Input Data

Issue - State: closed - Opened by nateraw over 6 years ago - 8 comments
Labels: question

#27 - How does Next-Sentence-Prediction benefit to both QA and NLI?

Issue - State: closed - Opened by guotong1988 over 6 years ago - 1 comment

#26 - Bidirectional Encoder = Transformer (self-attention), Is it true？

Issue - State: closed - Opened by guotong1988 over 6 years ago - 2 comments
Labels: question

#25 - how to get dataset.small from corpus.small

Issue - State: closed - Opened by jiangyangbo over 6 years ago - 1 comment

#24 - fixed a bug in position.py

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

#23 - update dataset.py

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

#22 - fix conditions for #13

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

#21 - Erroneous code

Issue - State: closed - Opened by winitzki over 6 years ago - 1 comment

#20 - readme has wrong commands

Issue - State: closed - Opened by daiwk over 6 years ago - 1 comment
Labels: bug

#19 - How to embedding segment lable

Issue - State: closed - Opened by waallf over 6 years ago - 3 comments
Labels: question

#18 - model/embedding/position.py

Issue - State: closed - Opened by zhupengjia over 6 years ago - 5 comments

#17 - The question about the implement of learning_rate

Issue - State: closed - Opened by wenhaozheng-nju over 6 years ago - 3 comments
Labels: enhancement

#16 - segmentation fault

Issue - State: closed - Opened by yangze01 over 6 years ago - 2 comments

#15 - EP_train:0: 0%|| 0/15636 [00:00<?, ?it/s]Segmentation fault (core dumped)

Issue - State: closed - Opened by yangze01 over 6 years ago - 1 comment

#14 - alpha-0.0.1a4 version released

Pull Request - State: closed - Opened by codertimo over 6 years ago - 1 comment

#13 - Question about random sampling.

Issue - State: closed - Opened by SongRb over 6 years ago - 3 comments

#12 - alpha0.0.1a3 version update

Pull Request - State: closed - Opened by codertimo over 6 years ago - 1 comment

#11 - Mask language model loss

Issue - State: closed - Opened by MarkWuNLP over 6 years ago - 3 comments

#10 - imbalance GPU memory usage

Issue - State: closed - Opened by WencongXiao over 6 years ago - 3 comments
Labels: help wanted, question

#9 - Circle-ci & Pypi auto building system

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