Recent Releases of Nkululeko 1.0: A Python package to predict speaker characteristics with a high-level interface
Nkululeko 1.0: A Python package to predict speaker characteristics with a high-level interface - 1.0.0
After having published the JOSS paper and released a zenodo version, we have our first official stable release
- Python
Published by felixbur 4 months ago
Nkululeko 1.0: A Python package to predict speaker characteristics with a high-level interface - nearing version 1.0
we wait for the JOSS reviewers comments, but else we'd like to approach v1.0.0
- Python
Published by felixbur about 1 year ago
Nkululeko 1.0: A Python package to predict speaker characteristics with a high-level interface - many more models, segmentation, self-labeling
- Python
Published by felixbur over 2 years ago
Nkululeko 1.0: A Python package to predict speaker characteristics with a high-level interface - added test splits
- Python
Published by felixbur over 3 years ago
Nkululeko 1.0: A Python package to predict speaker characteristics with a high-level interface - added dropout to mlp
added dropout to mlp
- Python
Published by felixbur almost 4 years ago
Nkululeko 1.0: A Python package to predict speaker characteristics with a high-level interface - added filters
added filters for databases, e.g. force a specific feature to be existent or limit the number of samples per speaker
- Python
Published by felixbur almost 4 years ago
Nkululeko 1.0: A Python package to predict speaker characteristics with a high-level interface - dataroots, segmented file index
added internal segmented file index added external data folder root initialization file
- Python
Published by felixbur about 4 years ago
Nkululeko 1.0: A Python package to predict speaker characteristics with a high-level interface - support for databases without target label
support for test databases without target label (for semi supervised learning)
- Python
Published by felixbur about 4 years ago
Nkululeko 1.0: A Python package to predict speaker characteristics with a high-level interface - many bugfixes
many bugfixes
- Python
Published by felixbur about 4 years ago
Nkululeko 1.0: A Python package to predict speaker characteristics with a high-level interface - added wav2vec2
added wav2vec2 embeddings
- Python
Published by felixbur about 4 years ago
Nkululeko 1.0: A Python package to predict speaker characteristics with a high-level interface - bug fixes
categories can now be selected via "labels" for speaker and sex information now multiple data tables can be specified
- Python
Published by felixbur over 4 years ago
Nkululeko 1.0: A Python package to predict speaker characteristics with a high-level interface - fixed svm c value setting
Setting the c value in the ini file was not handled
- Python
Published by felixbur over 4 years ago
Nkululeko 1.0: A Python package to predict speaker characteristics with a high-level interface - bug fix
fixed bug that caused xgboost.core.XGBoostError: [16:53:22] ../src/data/data.cc:530: Check failed: labels.Size() == numrow_ (324 vs. 223) : Size of labels must equal to number of rows. when running the hello world example twice, caused by a missing new feature extraction when new test speakers were selected but the old features sets were reused. Can happen when splitstrategy = speakersplit and should be prevented automatically by setting ['FEATS']['needsfeatureextraction'] = 'True'
- Python
Published by felixbur over 4 years ago
Nkululeko 1.0: A Python package to predict speaker characteristics with a high-level interface - added openXBOW
Numerous bug fixes and support for openXBOW bag-of-audio-words features
- Python
Published by felixbur over 4 years ago
Nkululeko 1.0: A Python package to predict speaker characteristics with a high-level interface - tested with two usecases
First version hopefully consistent tested with classification and regression experiments combining 3 classifiers with 3 features sets
- Python
Published by felixbur over 4 years ago
Nkululeko 1.0: A Python package to predict speaker characteristics with a high-level interface - trill features and data statistic plots
- Python
Published by felixbur over 4 years ago
Nkululeko 1.0: A Python package to predict speaker characteristics with a high-level interface - Added plotting options
Added plotting options and other functionality
- Python
Published by felixbur over 4 years ago
Nkululeko 1.0: A Python package to predict speaker characteristics with a high-level interface - Continous for classification
Added support to bin continous data for classification
- Python
Published by felixbur over 4 years ago
Nkululeko 1.0: A Python package to predict speaker characteristics with a high-level interface - added mlp
added mlp and changed result reporting
- Python
Published by felixbur over 4 years ago
Nkululeko 1.0: A Python package to predict speaker characteristics with a high-level interface - bug fixes for 0.4.0
bug fixes for 0.4.0
- Python
Published by felixbur over 4 years ago
Nkululeko 1.0: A Python package to predict speaker characteristics with a high-level interface - global config
added a global configuration object
- Python
Published by felixbur over 4 years ago
Nkululeko 1.0: A Python package to predict speaker characteristics with a high-level interface - enable sex distinction
- Python
Published by felixbur over 4 years ago
Nkululeko 1.0: A Python package to predict speaker characteristics with a high-level interface - first regression version
- Python
Published by felixbur over 4 years ago
Nkululeko 1.0: A Python package to predict speaker characteristics with a high-level interface - first cross corpus version
supports now mainly cross corpus emotion classification experiments
- Python
Published by felixbur over 4 years ago