Skip to content

Changelog.md

Latest commit

 

History

History
32 lines (28 loc) · 1.03 KB

Changelog.md

File metadata and controls

32 lines (28 loc) · 1.03 KB
Raw

Changelog

0.1.0

  • Added the initial project configuration
  • Introduced the scraping, processing and training pipelines
  • Added scripts:
    • scrape_google_reviews
    • process_dataset
    • perform_eda
    • train_model
    • summarize_experiment
  • Added v1.0 version of the paper

0.2.0

  • Reworked the data scraping pipeline
    • modified the input configuration format
    • scraped a large raw dataset
  • Introduced DVC
  • Updated and enhanced EDA
  • Reworked the data processing pipeline
    • introduced the train/test data processing separation
    • refactored modules responsible for fitting/processing the data
    • similified the extracted linguistic and numerical features
  • Redesigned the model's architecture and training process
    • introduced the embedding sequence processing block
    • added only a single main training objective (regression or classification)
    • introduced the secondary translation prediction task
  • Introduced more advanced evaluation utilities
    • enhanced experiments summary
    • added slice-based offline testing