nlpyang/BertSum issues and pull requests

#132 - Training in Colab (CNN/DM)

Issue - State: closed - Opened by Holmes2002 over 1 year ago

#131 - default batch_size is 3000, I don't quite understand, why so huge?

Issue - State: closed - Opened by yongzhuo over 2 years ago - 1 comment

#130 - expected mask dtype to be Bool but got Long

Issue - State: open - Opened by haidequanbu over 2 years ago

#129 - Order inconsistency of output candidate file with original test.json when testing bertSum Extractive

Issue - State: open - Opened by cece00 over 2 years ago - 1 comment

#128 - I also got slightly Lower Rouge score for the same code

Issue - State: open - Opened by milktea0917 over 2 years ago

#127 - Extractive Setting?

Issue - State: open - Opened by acc-galenicum over 2 years ago - 2 comments

#126 - FileNotFoundError

Issue - State: open - Opened by xiaomengxin123 over 2 years ago - 1 comment

#125 - [CLS] similar context vector on Evaluation

Issue - State: open - Opened by samish-dev almost 3 years ago

#124 - Questions about imbalance sentence distribution of label 1 and label 0 when training

Issue - State: open - Opened by boy-be-ambitious about 3 years ago

#123 - utf -8 관련 내용 추가

Pull Request - State: open - Opened by eeHeaven about 3 years ago - 1 comment

#122 - Help needed when processing my own dataset for testing.

Issue - State: open - Opened by namln2k about 3 years ago

#121 - Format of the .story files

Issue - State: closed - Opened by namln2k about 3 years ago

#120 - Can someone please explain the validation metrics for me?

Issue - State: closed - Opened by namln2k about 3 years ago

#119 - I got xent of 3-4 after 30000 training steps, is this normal?

Issue - State: open - Opened by Anothernewcomer about 3 years ago - 1 comment

#118 - How to continue training from previous checkpoints?

Issue - State: closed - Opened by namln2k about 3 years ago - 4 comments

#117 - Hi i wonder if i want to do a multi-classification task, what should i change?

Issue - State: closed - Opened by Anothernewcomer about 3 years ago

#116 - Package Requirements - versions?

Issue - State: open - Opened by BeaBea1234 over 3 years ago - 1 comment

#115 - How to train TransformerExt baseline？

Issue - State: open - Opened by GongShuai8210 over 3 years ago

#114 - Question for models/trainer.py#L325 ?

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

#113 - How to enter a sentence for prediction

Issue - State: open - Opened by heyanxian over 3 years ago

#112 - How to implement Bert + CNN

Issue - State: open - Opened by nguyenvd-its almost 4 years ago

#111 - train.py file not found in git repository

Issue - State: closed - Opened by bm2515 almost 4 years ago - 1 comment

#110 - Added requirements.txt with torch 1.1.0; Updated README with referenc…

Pull Request - State: open - Opened by Kthom1 almost 4 years ago

#109 - sorry, I can't know how to set Step 2. Download Stanford CoreNLP We will need Stanford CoreNLP to tokenize the data. Download it here and unzip it. Then add the following command to your bash_profile: export CLASSPATH=/path/to/stanford-corenlp-full-2017-06-09/stanford-corenlp-3.8.0.jar

Issue - State: open - Opened by Yazooliu almost 4 years ago

#108 - created requirements.txt for easier setup, and resolved pytorch installation confusion.

Pull Request - State: open - Opened by shehab-as about 4 years ago

#107 - How to do inference in new data without summaries？

Issue - State: open - Opened by sunsunxunxun about 4 years ago

#106 - Is there a list of stop words?

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

#105 - Summary files not found

Issue - State: open - Opened by KarlDavidson over 4 years ago - 1 comment

#104 - [yalecyu]fix get src and tgt bug

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

#103 - question/problem about the training and the evaluation results

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

#102 - How do you insert your own data set to summarize?

Issue - State: open - Opened by noabenefraim over 4 years ago - 1 comment

#101 - RuntimeError: CUDA out of memory.

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

#100 - I get slightly Lower Rouge score for the same code

Issue - State: open - Opened by aadithyamd over 4 years ago - 5 comments

#99 - Help needed

Issue - State: closed - Opened by Pradhy729 over 4 years ago - 1 comment

#98 - Error in Step 4 of processing

Issue - State: open - Opened by MughilM over 4 years ago - 1 comment

#97 - Problems with my own dataset and with the format_to_bert function.

Issue - State: closed - Opened by DarlineFiedler over 4 years ago - 15 comments

#96 - "Step 4. Format to Simpler Json Files" no output

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

#95 - How long does it take with just one gpu?

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

#94 - KeyError: 'paper_id'

Issue - State: closed - Opened by sayeedk06 almost 5 years ago - 1 comment

#93 - no such file:home/luohm/.pyrouge/settings.ini

Issue - State: open - Opened by luohm111 almost 5 years ago - 2 comments

#92 - Segmentation fault

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

#91 - Abstractive and Extractive options

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

#90 - Pre-processing new data set for BertSum

Issue - State: closed - Opened by nimahassanpour almost 5 years ago - 8 comments

#89 - Running the code for the first time

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

#88 - No such file or directory: '../logs/cnndm.log'

Issue - State: closed - Opened by siddas27 about 5 years ago - 5 comments

#87 - How to report the averaged results on the test set?

Issue - State: open - Opened by liuwei941213 about 5 years ago

#86 - index out of range at /opt/conda/conda-bld/pytorch_1556653114079/work/aten/src/TH/generic/THTensorEvenMoreMath.cpp:193

Issue - State: closed - Opened by riteshkumarumassedu about 5 years ago

#85 - About the highlights in CNN/DM datasets

Issue - State: open - Opened by TTWen about 5 years ago - 2 comments

#84 - returned non-zero exit status 2

Issue - State: open - Opened by vishaljoshi-066 about 5 years ago - 3 comments

#83 - Testing results in 10-15% less summaries than input. What are some possible causes for this?

Issue - State: open - Opened by Santosh-Gupta about 5 years ago

#82 - "Step 4. Format to Simpler Json Files" not working

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

#81 - Can I get the processed NYT data from you?

Issue - State: open - Opened by xxandfwy about 5 years ago - 2 comments

#80 - What version of stanford Core NLP was used originally ?

Issue - State: open - Opened by ASH1998 about 5 years ago - 1 comment

#79 - Is there a pretrained model provided?

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

#78 - What do you mean by label and clss?

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

#77 - Are there any explicit limits on summary length besides number sentences being 3?

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

#76 - TOKENIZED_PATH for command

Pull Request - State: closed - Opened by Guillem96 over 5 years ago

#75 - BERTSUM for German dataset

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

#74 - fix instruction in step 3 of readme

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

#73 - Subtraction, the `-` operator, with a bool tensor is not supported.

Issue - State: open - Opened by daiwei12 over 5 years ago - 2 comments

#72 - Is it possible to save before '-save_checkpoint_steps' is reached?

Issue - State: open - Opened by Santosh-Gupta over 5 years ago

#71 - summaries

Issue - State: open - Opened by shang624cui over 5 years ago - 1 comment

#70 - stories

Issue - State: open - Opened by shang624cui over 5 years ago - 2 comments

#69 - hotfix for encoding error

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

#68 - about Chinese

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

#67 - Format to pytorch gives empty output

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

#66 - Is it possible to set `load_pretrained_bert=False` in `model = Summarizer(args, device, load_pretrained_bert=True)` if we use a checkpoint?

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

#65 - the model filefolder is empty

Issue - State: closed - Opened by martin6336 over 5 years ago

#64 - Question about test_rouge function

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

#63 - Error when running testing 'FileNotFoundError: [Errno 2] No such file or directory: '/root/.pyrouge/settings.ini''

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

#62 - What to put for 'urls' when preprocessing new data with format_to_lines

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

#61 - Sample tokenized datafile

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

#60 - Testing on unseen/new data

Issue - State: open - Opened by tanmaypandey7 over 5 years ago - 8 comments

#59 - token len more than 512

Issue - State: closed - Opened by daiwei12 over 5 years ago - 6 comments

#58 - Add a setup.py

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

#57 - Extractive gold label creation process

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

#56 - Model downloading speed is too slow and failing.

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

#55 - other types of summarization layer?

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

#54 - Use of <q> in gold / pred

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

#53 - Question about Learning rate

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

#52 - Processed data lacks files

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

#51 - requirements for bert-large?

Issue - State: open - Opened by rush86999 over 5 years ago - 1 comment

#50 - Truncated article without oracle

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

#49 - about load_pretrained_bert=True?

Issue - State: closed - Opened by rush86999 over 5 years ago - 1 comment

#48 - about positional encoding

Issue - State: closed - Opened by rush86999 over 5 years ago - 1 comment

#47 - how long does it take with 3 gpus?

Issue - State: closed - Opened by rush86999 over 5 years ago - 2 comments

#46 - question about candidate

Issue - State: closed - Opened by rush86999 over 5 years ago - 1 comment

#45 - Infinite loop on loading training dataset

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

#44 - run error due to dsataset

Issue - State: closed - Opened by angeluau over 5 years ago - 16 comments

#43 - about preprocessing step

Issue - State: closed - Opened by rush86999 over 5 years ago - 4 comments

#42 - About preprocessing for CNN/DailyMail

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

#41 - Doubt about processed data

Issue - State: closed - Opened by Shashi456 over 5 years ago - 9 comments

#40 - Adding inference api functions for quick forward passes during deployment

Pull Request - State: open - Opened by enzoampil over 5 years ago - 1 comment

#39 - About rouge score

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

#38 - Bert Fine-tuned problem

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

#37 - Tow confused parts

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

#36 - Format JSON file to Pytorch File

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

GitHub / nlpyang/BertSum issues and pull requests