New feature

  • Visualization of the download process of language models
  • Can set error level from python
  • Logistic regression is default for classification in textTrain.
  • Megatron language model functionality

Bug Fixes

  • When GPU is not found, CPU is used.

New feature

Bug Fixes

New Features

Bug Fixes

  • In textCentrality(), words to be plotted are selected with word_data1_all$extremes_all_x >= 1 (rather than ==1).
  • prompt option added to textrpp_initiate()
  • tokenization is made with NLTK from python.
  • Code has been cleaned up and prepared for CRAN

New Features

  • New functions being tested: textWordPredictions() (which has a trial period/not fully developed and might be removed in future versions); p-values are not yet implemented.
  • Possibility to use textPlot() for objects from both textProjection() and textWordPredictions()

Minor changes

  • Changed wordembeddigs to word_embeddings through out the code/package.

Bug Fixes

  • Warnings about seed when using multi-cores on Mac is addressed.

New Features

  • textrpp_initiate() runs automatically in library(text) when default environment exits
  • Python warnings a captured in embedding comments
  • Option to print python options to console
  • Updated the permutation test for plotting and textSimilarityTest().

Minor changes

  • Changed from stringr to stringi (and removed tokenizer) as imported package

New Features

New Features

New Features

  • In textProjetion() and textProjetionPlot() it now possible to add points of the aggregated word embeddings in the plot
  • In textProjetion() it now possible to manually add words to the plot in order to explore them in the word embedding space.
  • In textProjetion() it is possible to add color or remove words that are more frequent on the opposite “side” of its dot product projection.
  • In textProjection() with split == quartile, the comparison distribution is now based on the quartile data (rather than the data for mean)

Bug Fixes

  • If any of the tokens to remove is “[CLS]”, subtract 1 on token_id so that it works with layer_aggregation_helper. 0.9.11
  • Can now submit one word to textEmbed() with decontexts=TRUE.
  • textSimilarityTest() is not giving error when using method = unpaired, with unequal number of participants in each group.

New Features

  • textPredictTest() function to significance test correlations of different models. 0.9.11

Bug Fixes

  • If any of the tokens to remove is “[CLS]”, subtract 1 on token_id so that it works with layer_aggregation_helper. 0.9.11

This version is now on CRAN. ### New Features - Adding option to deselect the step_centre and step_scale in training. - Cross-validation method in textTrainRegression() and textTrainRandomForrest() have two options cv_folds and validation_split. (0.9.02) - Better handling of NA in step_naomit in training. - DistilBert model works (0.9.03)

Major changes

Bug Fixes

  • textProjectionPlot() plots words extreme in more than just one feature (i.e., words are now plotted that satisfy, for example, both plot_n_word_extreme and plot_n_word_frequency). (0.9.01)
  • textTrainRegression() and textTrainRandomForest() also have function that select the max evaluation measure results (before only minimum was selected all the time, which, e.g., was correct for rmse but not for r) (0.9.02)
  • removed id_nr in training and predict by using workflows (0.9.02).

Minor changes