Neural Sequence Learning Using TensorFlow
The Neural Monkey package provides a higher level abstraction for sequential neural network models, most prominently in Natural Language Processing (NLP). It is built on TensorFlow. It can be used for fast prototyping of sequential models in NLP which can be used e.g. for neural machine translation or sentence classification.
The higher-level API brings together a collection of standard building blocks (RNN encoder and decoder, multi-layer perceptron) and a simple way of adding new building blocks implemented directly in TensorFlow.
neuralmonkey-train <EXPERIMENT_INI>
neuralmonkey-run <EXPERIMENT_INI> <DATASETS_INI>
neuralmonkey-server <EXPERIMENT_INI> [OPTION] ...
neuralmonkey-logbook --logdir <EXPERIMENTS_DIR> [OPTION] ...-
You need Python 3.6 (or higher) to run Neural Monkey.
-
When using virtual environment, execute these commands to install the Python dependencies:
$ source path/to/virtualenv/bin/activate # For GPU-enabled version (virtualenv)$ pip install --upgrade -r requirements-gpu.txt # For CPU-only version (virtualenv)$ pip install --upgrade -r requirements.txt
-
If you are using the GPU version, make sure that the
LD_LIBRARY_PATHenvironment variable points tolibandlib64directories of your CUDA and CuDNN installations. Similarly, yourPATHvariable should point to thebinsubdirectory of the CUDA installation directory. -
If the training crashes on an unknown dependency, just install it with pip. Remember to keep your virtual environment up-to-date with the package requirements file, which may be changed over time. To update the dependencies, re-run the
pip installcommand from above (pay attention to the distinction between GPU and non-GPU versions).
There is a tutorial that you can follow, which gives you the overwiev of how to design your experiments with Neural Monkey.
-
bin: Directory with neuralmonkey executables -
examples: Example configuration files for ready-made experiments -
lib: Third party software -
neuralmonkey: Python package files -
scripts: Directory with tools that may come in handy. Note dependencies for these tools may not be listed in the project requirements. -
tests: Test files
You can find the API documentation of this package here. The documentation files are generated from docstrings using autodoc and Napoleon extensions to the Python documentation package Sphinx. The docstrings should follow the recommendations in the Google Python Style Guide. Additional details on the docstring formatting can be found in the Napoleon documentation as well.
-
tflearn – a more general and less abstract deep learning toolkit built over TensorFlow
-
nlpnet – deep learning tools for tagging and parsing
-
NNBlocks – a library build over Theano containing NLP specific models
-
Nematus - A tool for training and running Neural Machine Translation models
-
seq2seq - a general-purpose encoder-decoder framework for Tensorflow
-
OpenNMT - open sourcce NMT in Torch
If you use the tool for academic purporses, please consider citing the following paper:
@article{NeuralMonkey:2017,
author = {Jind{\v{r}}ich Helcl and Jind{\v{r}}ich Libovick{\'{y}}},
title = {{Neural Monkey: An Open-source Tool for Sequence Learning}},
journal = {The Prague Bulletin of Mathematical Linguistics},
year = {2017},
address = {Prague, Czech Republic},
number = {107},
pages = {5--17},
issn = {0032-6585},
doi = {10.1515/pralin-2017-0001},
url = {http://ufal.mff.cuni.cz/pbml/107/art-helcl-libovicky.pdf}
}The software is distributed under the BSD License.
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent