Migrated repository
Go to file
2018-12-03 21:54:22 +08:00
.github add logging in Trainer & Tester 2018-08-17 11:16:13 +08:00
docs * update README.md 2018-11-29 23:27:15 +08:00
fastNLP add magic iter in dataset 2018-12-03 21:54:22 +08:00
reproduction refine git commits 2018-11-27 22:43:29 +08:00
test add magic iter in dataset 2018-12-03 21:54:22 +08:00
.travis.yml - add progress bar for data set loading 2018-10-01 20:33:29 +08:00
LICENSE add LICENSE, setup.py & requirements.txt 2018-05-25 18:32:02 +08:00
README.md * update README.md 2018-11-29 23:27:15 +08:00
requirements.txt * update README.md 2018-11-29 23:27:15 +08:00
setup.py refine git commits 2018-11-27 22:43:29 +08:00

fastNLP

Build Status codecov PyPI version Hex.pm Documentation Status

FastNLP is a modular Natural Language Processing system based on PyTorch, built for fast development of NLP models.

A deep learning NLP model is the composition of three types of modules:

For example:

Requirements

  • numpy>=1.14.2
  • torch>=0.4.0
  • tensorboardX

Resources

Installation

Run the following commands to install fastNLP package.

pip install fastNLP

Project Structure

module type functionality example
encoder encode the input into some abstract representation embedding, RNN, CNN, transformer
aggregator aggregate and reduce information self-attention, max-pooling
decoder decode the representation into the output MLP, CRF
fastNLP an open-source NLP library
fastNLP.api APIs for end-to-end prediction
fastNLP.core data representation & train/test presedure
fastNLP.models a collection of NLP models
fastNLP.modules a collection of PyTorch sub-models/components/wheels
fastNLP.io readers & savers