Commit History

  1. Move common logic to function

    d8a44ef
  2. Update docstrings and comments

    d4ef46b
  3. Swap test dataset

    183f8cd

  1. Add amazonreviews model

    baf0dee
  2. Parallelize text cleaning

    1414454
  3. Fix wrong sentiment mapping on the test dataset

    3a96048
  4. Add min-df option

    8b10b79

  1. Improved models

    7f29122
  2. Change df

    cc21abf
  3. Add new model trained on sentiment140

    419453c
  4. Cache label data along with tokenized text data

    af84d9b
  5. Fix broken tokenization

    447f97e
  6. Add slang map

    e1645d7

  1. Add progress bar to serialize

    632adc4
  2. Add more vectorizers, classifiers and CLI options

    b0ade1a
  3. Refactor typing and update tokenization rules

    228859a
  4. Add document frequency threshold

    c5ed75e
  5. Ignore amazonreviews test

    d09d1f6
  6. Chunked serialization

    afaacd1

  1. Update options, force GC, tweak parameters and add flags

    18cc46a
  2. Ability to change number of parallel jobs for search

    8471e78
  3. Create model in train_model

    3854a1f
  4. Tokenization rework

    2c1f9dd
  5. Add dataset for testing

    e50b20c

  1. Slight optimizations

    0ca5366
  2. Change HF entry point and add examples

    b42b884
  3. Handle missing spacy model

    7ce074d
  4. Import spacy model as module

    7b9e59d
  5. Add evaluate command

    cdf1241
  6. Use spacy instead of nltk and move data functions to separate module

    a092d54
  7. Remove unused constanst

    9a96b6b

  1. Add cross validation

    5a2db0a
  2. Use stopwords from NLTK and download NLTK data

    204391c
  3. Fix merge

    0993d5e

  1. Merge branch 'master' of https://github.com/Tymec/projekt-psi

    b43b167
  2. Completely change the structure of the project

    85ac990
  3. Model

    391bd16
  4. Restructure project into package structure

    667fe9d