Commit History

  1. reduce padding and wrap

    acaabb6
  2. implement worker pool

    6305c92
  3. silly first implementation of workers

    dc2b31f

  1. move json files to datasets/

    67c1004

  1. silly sequal computations done

    85c26c6
  2. first step in sequential calculations

    ea54a38
  3. reorganise files and fix import bugs

    aec018b
  4. all logic exposed in frontend

    0386dad
  5. initial frontend work

    b500a16
  6. tokenize after removing word

    fa33ecb
  7. countLogLoss

    3bd66ee
  8. move most expected values into json files

    b297208
  9. move vocabulary into json file

    8b6c9f1

  1. tokenizeTextLM loop and cache tokenizeLM

    37fc247
  2. tokenizeTextLM

    efb68ae
  3. update jsconfig.js

    1224769
  4. tokenizeLM (no cache)

    8b43a5b
  5. also return vocabulary from BPE

    4ebb38a
  6. unigramLM

    19d4dca
  7. tokenizeBPE

    5a9f18c
  8. fix preprocess

    cdbb891
  9. BPE

    7e96b6d
  10. mostFrequentPair

    2c0ab92
  11. mergeBigrams

    dd9e999
  12. pairCount

    ff51fd5
  13. preprocess and initialVocabulary

    2191a1b
  14. some setup

    85163ad
  15. initial commit

    be53355