epwalsh/pytorch-crf issues and pull requests

#94 - Tagging a sequence of sentences

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

#93 - fixes

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

#93 - fixes

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

#92 - close #77

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

#91 - Try integrated flair contextual string embeddings

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

#91 - Try integrated flair contextual string embeddings

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

#90 - close #87

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

#89 - Allow training on shorter sentences during earlier rounds of training

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

#89 - Allow training on shorter sentences during earlier rounds of training

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

#88 - Build jupyter notebook example for conll2003

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

#87 - Build jupyter notebook interface

Issue - State: closed - Opened by epwalsh over 6 years ago

#86 - close #85

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

#86 - close #85

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

#85 - For cyclic learning rate, update LR every iteration, instead of every epoch

Issue - State: closed - Opened by epwalsh over 6 years ago

#85 - For cyclic learning rate, update LR every iteration, instead of every epoch

Issue - State: closed - Opened by epwalsh over 6 years ago

#84 - Alternate method of loading training data

Issue - State: closed - Opened by epwalsh over 6 years ago
Labels: TYPE: feature

#83 - Build docker image for testing

Issue - State: closed - Opened by epwalsh over 6 years ago
Labels: TYPE: feature, PRIORITY: high

#83 - Build docker image for testing

Issue - State: closed - Opened by epwalsh over 6 years ago
Labels: TYPE: feature, PRIORITY: high

#82 - close #78

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

#81 - Integrate ELMo embeddings

Issue - State: open - Opened by epwalsh over 6 years ago
Labels: TYPE: feature

#81 - Integrate ELMo embeddings

Issue - State: open - Opened by epwalsh over 6 years ago
Labels: TYPE: feature

#80 - Log learning rate to TensorBoard

Issue - State: closed - Opened by epwalsh over 6 years ago

#80 - Log learning rate to TensorBoard

Issue - State: closed - Opened by epwalsh over 6 years ago

#79 - close #75

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

#79 - close #75

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

#78 - Implement differential learning rate across layers (pretrained embedding should have smaller learning rate)

Issue - State: closed - Opened by epwalsh over 6 years ago

#78 - Implement differential learning rate across layers (pretrained embedding should have smaller learning rate)

Issue - State: closed - Opened by epwalsh over 6 years ago

#77 - Implement LR auto-find

Issue - State: closed - Opened by epwalsh over 6 years ago
Labels: TYPE: feature, PRIORITY: low

#77 - Implement LR auto-find

Issue - State: closed - Opened by epwalsh over 6 years ago
Labels: TYPE: feature, PRIORITY: low

#76 - Implement snapshot ensembles

Issue - State: open - Opened by epwalsh over 6 years ago
Labels: TYPE: feature, PRIORITY: low

#75 - Implement cyclic learning rate schedule

Issue - State: closed - Opened by epwalsh over 6 years ago
Labels: TYPE: feature, PRIORITY: high

#75 - Implement cyclic learning rate schedule

Issue - State: closed - Opened by epwalsh over 6 years ago
Labels: TYPE: feature, PRIORITY: high

#74 - close #73

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

#74 - close #73

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

#73 - Allow training from pickled objects

Issue - State: closed - Opened by epwalsh over 6 years ago

#73 - Allow training from pickled objects

Issue - State: closed - Opened by epwalsh over 6 years ago

#72 - close #71

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

#72 - close #71

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

#71 - Allow sentence-level features

Issue - State: closed - Opened by epwalsh over 6 years ago
Labels: TYPE: feature, PRIORITY: high

#71 - Allow sentence-level features

Issue - State: closed - Opened by epwalsh over 6 years ago
Labels: TYPE: feature, PRIORITY: high

#70 - close #67

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

#69 - close #67, close #68

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

#68 - Get rid of duplicate code

Issue - State: closed - Opened by epwalsh over 6 years ago
Labels: TYPE: house cleaning

#67 - Implement CLI for labelling a test dataset

Issue - State: closed - Opened by epwalsh over 6 years ago
Labels: TYPE: feature, PRIORITY: high

#66 - close #65

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

#65 - Improvements

Issue - State: closed - Opened by epwalsh over 6 years ago

#65 - Improvements

Issue - State: closed - Opened by epwalsh over 6 years ago

#64 - close #58, close #63

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

#64 - close #58, close #63

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

#63 - Improve logging

Issue - State: closed - Opened by epwalsh over 6 years ago

#62 - close #61

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

#62 - close #61

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

#61 - Load pre-trained word embeddings directly from text file

Issue - State: closed - Opened by epwalsh over 6 years ago
Labels: TYPE: feature, PRIORITY: high

#61 - Load pre-trained word embeddings directly from text file

Issue - State: closed - Opened by epwalsh over 6 years ago
Labels: TYPE: feature, PRIORITY: high

#60 - Use sphinx to create API docs

Issue - State: open - Opened by epwalsh over 6 years ago
Labels: TYPE: documentation, PRIORITY: low

#60 - Use sphinx to create API docs

Issue - State: open - Opened by epwalsh over 6 years ago
Labels: TYPE: documentation, PRIORITY: low

#59 - close #52

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

#59 - close #52

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

#58 - Save best model after training completes

Issue - State: closed - Opened by epwalsh over 6 years ago

#57 - Document basics in README

Issue - State: closed - Opened by epwalsh over 6 years ago
Labels: PRIORITY: high, TYPE: documentation

#57 - Document basics in README

Issue - State: closed - Opened by epwalsh over 6 years ago
Labels: PRIORITY: high, TYPE: documentation

#56 - close #55

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

#55 - Organize TB scalars by dataset

Issue - State: closed - Opened by epwalsh over 6 years ago

#55 - Organize TB scalars by dataset

Issue - State: closed - Opened by epwalsh over 6 years ago

#54 - close #53

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

#53 - Use s3 for tensorboard logging

Issue - State: closed - Opened by epwalsh over 6 years ago

#52 - Save model checkpoints

Issue - State: closed - Opened by epwalsh over 6 years ago
Labels: TYPE: feature, PRIORITY: high

#51 - close #27

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

#50 - close #49

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

#49 - Integrate TensorBoard

Issue - State: closed - Opened by epwalsh over 6 years ago

#48 - close #46

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

#47 - CNN isn't learning well

Issue - State: closed - Opened by epwalsh over 6 years ago
Labels: PRIORITY: high, TYPE: bug

#46 - better decay schedule

Issue - State: closed - Opened by epwalsh over 6 years ago

#45 - Try BIOES labelling scheme

Issue - State: closed - Opened by epwalsh over 6 years ago

#45 - Try BIOES labelling scheme

Issue - State: closed - Opened by epwalsh over 6 years ago

#44 - Use micro-average F1 as target metric instead of macro-average

Issue - State: closed - Opened by epwalsh over 6 years ago
Labels: PRIORITY: critical

#43 - close #17, close #41

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

#43 - close #17, close #41

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

#42 - Wrap train logging statements into a logger class

Issue - State: closed - Opened by epwalsh over 6 years ago

#42 - Wrap train logging statements into a logger class

Issue - State: closed - Opened by epwalsh over 6 years ago

#41 - Apply dropout layers

Issue - State: closed - Opened by epwalsh over 6 years ago

#41 - Apply dropout layers

Issue - State: closed - Opened by epwalsh over 6 years ago

#40 - close #36

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

#40 - close #36

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

#39 - Implement automatic early stopping

Issue - State: closed - Opened by epwalsh over 6 years ago - 1 comment
Labels: TYPE: feature

#38 - close #34

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

#38 - close #34

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

#37 - Remove dependency on torchtext?

Issue - State: closed - Opened by epwalsh over 6 years ago
Labels: TYPE: house cleaning, PRIORITY: medium

#36 - Allow training from a configuration file, save results in that config file after training

Issue - State: closed - Opened by epwalsh over 6 years ago
Labels: TYPE: feature, PRIORITY: medium

#36 - Allow training from a configuration file, save results in that config file after training

Issue - State: closed - Opened by epwalsh over 6 years ago
Labels: TYPE: feature, PRIORITY: medium

#35 - close #32

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

#34 - Strip out the relevant parts from Allennlp and integrate directly into here, add license directly to each file

Issue - State: closed - Opened by epwalsh over 6 years ago

#33 - close #30

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

#32 - During training, keep track of best epoch measured by loss and validation macro avg. f1

Issue - State: closed - Opened by epwalsh over 6 years ago

#31 - close #29

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

#30 - Allow batch size greater than 1 (just not in parallel yet)

Issue - State: closed - Opened by epwalsh over 6 years ago
Labels: TYPE: feature

#30 - Allow batch size greater than 1 (just not in parallel yet)

Issue - State: closed - Opened by epwalsh over 6 years ago
Labels: TYPE: feature

#29 - Infer labels from dataset

Issue - State: closed - Opened by epwalsh over 6 years ago
Labels: PRIORITY: high, TYPE: enhancement

#29 - Infer labels from dataset

Issue - State: closed - Opened by epwalsh over 6 years ago
Labels: PRIORITY: high, TYPE: enhancement

#28 - close #18

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

GitHub / epwalsh/pytorch-crf issues and pull requests