pervrosen / fastPunct

Fast and accurate punctuation restoration on unsegmented text using sequence to sequence model consisting of bidirectional recurrent neural network model ( LSTM ) and attention mechanism.

fastPunct : Fast and accurate punctuation restoration with sequence to sequence networks.

Installation:

pip install --upgrade fastpunct

Supported languages:

en - english

Usage:

from fastpunct import FastPunct
# The default language is 'en'
fastpunct = FastPunct('en')
fastpunct.punct(["oh i thought you were here", "in theory everyone knows what a comma is", "hey how are you doing", "my name is sheela i am in love with hrithik"], batch_size=32)
# ['Oh! I thought you were here.', 'In theory, everyone knows what a comma is.', 'Hey! How are you doing?', 'My name is Sheela. I am in love with Hrithik.']

Note:

maximum length of input currently supported - 400