issues
search
GateNLP
/
gateplugin-LearningFramework
A plugin for the GATE language technology framework for training and using machine learning models. Currently supports Mallet (MaxEnt, NaiveBayes, CRF and others), LibSVM, Scikit-Learn, Weka, and DNNs through Pytorch and Keras.
https://gatenlp.github.io/gateplugin-LearningFramework/
GNU Lesser General Public License v2.1
26
stars
6
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
Make dense json keras back-end work again
#73
johann-petrak
closed
6 years ago
1
On-demand caching/reusing of the training set for algorithm/hyperparameter exploration
#72
johann-petrak
opened
6 years ago
0
Add support for topic models by wrapping gensim
#71
johann-petrak
opened
6 years ago
1
NPE when exporting
#70
tasoslytos
closed
5 years ago
3
Exporter PR under Windows populates data directory field by default
#69
GenevieveGorrell
closed
6 years ago
2
Properly implement classification confidence scores
#68
johann-petrak
opened
6 years ago
3
Refactor creating class annotation code, move from ModelApplication to SeqEncoder
#67
johann-petrak
opened
6 years ago
0
Implement confidenceThreshold parameter
#66
johann-petrak
closed
6 years ago
0
Rethink API for sequence encoder and implement a few more
#65
johann-petrak
opened
6 years ago
1
BIO encoder seems to generate targets for overlaps where there are none
#64
johann-petrak
closed
6 years ago
3
Remove sequence encoder BIEO until implemented
#63
johann-petrak
closed
6 years ago
0
Export for chunking task fails
#62
johann-petrak
closed
6 years ago
0
NPE on Mallet Winnow algorithm
#61
GenevieveGorrell
closed
6 years ago
0
Refactor so that gatelib-interaction is used again
#60
johann-petrak
closed
5 years ago
1
(Temporarily) Merge in some code for mavenization, then factor out again
#59
johann-petrak
opened
6 years ago
2
Move towards mavenized build for Gate version 8.5
#58
johann-petrak
closed
6 years ago
2
Dense JSON Corpus Representation: double check escaping of new line characters
#57
johann-petrak
closed
6 years ago
2
Dense JSON Corpus Representation: consider using gzip compression
#56
johann-petrak
opened
6 years ago
0
Consider making it work with online learning algorithms
#55
johann-petrak
opened
6 years ago
0
Remove support start/stop features based on the WITHIN declaration
#54
johann-petrak
opened
6 years ago
0
Refactor JSON exporter for sequences
#53
johann-petrak
opened
7 years ago
1
Try and make all reads able to read from URLs instead of just files.
#52
johann-petrak
closed
7 years ago
2
Investigate serial classification that depends on previous classes
#51
johann-petrak
closed
7 years ago
2
Add support for N-grams for ATTRIBUTELIST
#50
johann-petrak
opened
7 years ago
0
Add START and STOP symbol representations to sequence learning problems.
#49
johann-petrak
opened
7 years ago
7
Add feature functions that can depend on the target
#48
johann-petrak
opened
7 years ago
0
Add ordinal and maybe isotonic regression algorithms
#47
johann-petrak
closed
6 years ago
1
Check odd confidence measures in the tests example for LibSVM
#46
johann-petrak
closed
6 years ago
1
Remove the wrapper home directory from the wrapper invocation argument list
#45
johann-petrak
closed
7 years ago
1
Support out-of-core exporting and training and alternate corpus representations
#44
johann-petrak
opened
7 years ago
7
Make sure that we have a free lunch
#43
domrout
closed
7 years ago
1
Add liblinear
#42
johann-petrak
opened
7 years ago
4
Implement our own options parsing class and use in Parms
#41
johann-petrak
opened
7 years ago
4
Allow multi-class sequence tagging
#40
johann-petrak
opened
7 years ago
1
Check desired vs. actual behaviour if both target feature and annotation type parms are missing
#39
johann-petrak
opened
7 years ago
0
Consider seq2seq learning
#38
johann-petrak
opened
8 years ago
2
Implement engines for HTTP servers
#37
johann-petrak
closed
8 years ago
1
Add support for other interesting learning frameworks and libraries
#36
johann-petrak
opened
8 years ago
1
Add support for tensorflow
#35
johann-petrak
opened
8 years ago
1
Think about how to add per-instance cost to classification.
#34
johann-petrak
closed
8 years ago
1
Add SnakeYaml as a Maven dependency again.
#33
johann-petrak
closed
7 years ago
1
Better error handling when there is no target feature
#32
johann-petrak
closed
6 years ago
0
Make feature file case insensitive
#31
johann-petrak
closed
8 years ago
0
Make sure that SnakeYaml uses the correct classloader to find our own classes
#30
johann-petrak
closed
8 years ago
1
Clean up the mess of when models, info files, engines etc are saved
#29
johann-petrak
opened
8 years ago
0
Add proper Missing value treatment to exporting
#28
johann-petrak
closed
8 years ago
1
Find out why we get a missing location when exporting abalone
#27
johann-petrak
closed
8 years ago
1
Investigate drop in Seq Learning performance
#26
johann-petrak
closed
8 years ago
1
Do not include instance annotation type in the feature names.
#25
johann-petrak
closed
8 years ago
1
Add Evaluation PR for regression problems
#24
johann-petrak
closed
8 years ago
1
Previous
Next