Please enable JavaScript.
Coggle requires JavaScript to display documents.
Pipeline (Build Data (get total vocab (get total word embedding (get…
Pipeline
Build Data
get total vocab
get total word embedding
get intersection
vocab & word embedding
write vocab to file
write trimmed w_embedding's vocab
write all word vocab
export embedding vocabulary
intersec all combination of a word to large word set
Large fasttext word set
pre process to token
map vowel char to base
genertate all combination vowel of a word
return vocab with
:star: key : base word
:star: value : all aternative words
Load vocab
If you have dumped necessary file
Create
make pair source-target data
initial pair train dev test
Save
dump train dev test
tokenize
words to ids
Labels to ids ~ [0,62]
Model
Dynamic
Sequence Labeling
Build
run_train(is_continue=False)